PNG vs TIFF — The Format That Won’t Hurt Your Scanned Photos
When you begin to scan your photo collection of paper prints, slides, and negatives, one of the biggest questions you’re going to have to answer for yourself is which file format will you want to save all of your scanned photos with.
For those interested in saving the highest quality images, two of the most popular formats that often stand out are the PNG and the TIFF. But, which one would be best for archiving scanned photographs?
The decision-making process starts out seeming a bit tricky because not only do both the PNG and TIFF produce high-quality images, save with transparencies, and save with 16-bits per channel, but they also both maintain their original image quality even if they have been edited and re-saved over and over.
But, the choice start getting a little clearer once we compare the characteristics that actual separate the two.
PNG
Portable Network Graphic ● Pronounced PEE-en-JEE ● Extension: .png
- Created in 1996 by the PNG working group to be an improved replacement for the GIF file
- The format was designed for transferring images on the Internet
- Reduced file sizes without any loss of image quality (using lossless compression)
- Can only hold full-color non-palette-based RGB or RGBA (with alpha channel)
- “Web-friendly” support for display on the internet by web browsers
- Limited (not standardized) support for storing metadata with XMP (when supported by the application)
TIFF
Tagged Image File Format ● Pronounced TIF like “Tiffany” ● Extension: .tif or .tiff
- Created in 1986 by the Aldus Corporation (later acquired by Adobe Systems in 1994) for use in desktop publishing
- The format was designed as an attempt to get desktop scanner vendors to agree on a common scanned image file format across multiple computer platforms
- Files are usually stored uncompressed, but the option to reduce the file size is also available (using lossless compression [ZIP or LZW] and sometimes even lossy compression when supported by the application)
- Can store “layers” within (useful in high-end editing applications like Adobe Photoshop)
- Can hold all color, color depths and color groups (e.g. RGB as well as CMYK, LAB, etc.)
- Full built-in support for storing standardized EXIF and IPTC metadata fields inside the file (e.g. “date taken” and “captions”)
- Can store multiple images, such as the pages of a scanned or faxed document (similar to a PDF file)
Make Sure the PNG or TIFF Is Even the Right Choice For You
I think it's important to bring up how critical it is to know whether the PNG or the TIFF is even the right choice for you. You may have been led to believe these were your two best options, and this is why you arrived at this post.
Even though these are two of the most popular file formats for images in general, it doesn't necessarily mean either one of these will be the best for your own photo scanning needs that also matches your level of goals for your photo collections.
Choosing the right file format to archive your scanned items is kind of like picking a tool for each situation in a household project. Sometimes what you are working on requires a flat head screwdriver. Other times you need a Phillips head screwdriver. Yet other times, many are surprised what they really need is a nail and hammer instead.
Each item in your list to be scanned might need a different file type to be used — documents, film, and prints, etc., “One size” likely won't “fit all” here.
Below is a special video that's a section taken from one of my lessons called “Digital Masters and File Types” in my video training course on scanning and organizing photos. Until now you needed to be a Member, but for a limited time, I'm also making this portion available for you to watch too.
Think of this as a Mini-Lesson on learning the best file types for scanning your photo collections!
The Main Reasons I Still Don't Recommend the PNG File Format Over the TIFF for Archiving Your Scanned Photo Collections
I think the PNG file format is fantastic and really has many benefits. In fact, I use it all the time on this website for certain images where I know they will produce a more detailed image than a JPEG would and with a smaller file size.
It's not that the PNG can't or should never be used with your scanning project, it's just that there are other formats more suited for the purpose of archiving scanned photos.
If you're looking for a bit more information, let me pick 4 of my favorite points and go into some additional detail.
There Are Formats Designed For Scanning
The TIFF file format was created in the mid-1980's for use in desktop publishing and was then designed to become the go-to standard format for saving scanned photographs and has been for over 3 decades.
And the PNG file literally tells us in its name that it wasn't designed for such a purpose. PNG is short for Portable Network Graphic — key word there is “portable.” And this makes sense because files meant to be transferred around the internet do need to be as portable as possible — especially on devices like smartphones and tablets trying to access cellular networks in areas with poor service.
But, when I think about saving and archiving my scanned photograph collection, portability isn't something I'm concerned with at all. Archiving isn't about movement. Frankly, it's the opposite. I'm thinking about “longevity”, “compatibility” and “no compromises.”
If the TIFF file is the standard for scanned photos, for me, there would have to be a hugely compelling reason to choose another format; especially when most of us are only storing away a single master image for every single photograph from our original (analog) paper print, slide, and negative collection.
Whichever file format you choose to archive your scanned items in needs to give you confidence that they will be your one-and-only authoritative version of the (analog) originals you scanned them from.
Lack of Scanner Software Support
Even though there are a lot of file formats and often variations of each of them to choose from, when you're actually in your scanning software ready to save a photograph you just scanned, you will quickly find out that even the most popular scanning software only gives you a small handful of options to choose from — and the PNG format likely isn't one of them.
If the PNG format was really a strong contender for saving scanned items with, isn't it strange — strike that, incredibly strange — that after all of these years it still has not been added to the list of file type options to choose from in major scanning software?
It's not an option to choose from. It's not there.
Wouldn't scanner manufacturers and scanning software developers want you to easily have access to saving in this format it was meant to be used for saving scanned items?
So, this means if your goal is to archive your scanned images in the PNG format while scanning with software that can't natively save in the PNG format, you're going to have to go through an additional step later to convert the file format from one to the other.
First, you will save your scanned images in an alternate file format that is available as an option in your scanning software, and then next you will “batch” re-save each one of them using the PNG format using another application that can handle this conversion process.
If you're good with computers and comfortable working in your computer's file system, this procedure likely won't give you pause. But, if you are a beginner or just not at all confident with your file system, this conversion process could be overwhelming for you and therefore likely a step you typically would choose to avoid at the risk of doing it wrong.
Something important to keep in mind when doing the conversion is you will want to make sure to first save your scans from your scanning software in a lossless format (and not a lossy one like the JPEG format). This way you aren't losing any image detail before re-saving it in your other application in the lossless PNG format (in an attempt to preserve what has already been lost). And most often, unless you are using a high-end scanning application that gives you a RAW file option, the only lossless format to choose from in most scanning software is the TIFF.
Non-Standardized Metadata Implementation
What's really neat about working with your scanned photo collection is the ability to tell stories with all of these old photos. And the best way to do this is to add important descriptive information that is stored inside of each that can be accessed by someone using an application that has the ability to read and display this information.
This added information is referred to as “metadata” and a few popular examples would be the “shoot date” of when the photo was actually taken, a “caption” of description of what we're seeing in the photograph, the name of the person who took the photo, and the location of where the photo was taken.
How Photo Metadata Works
The best method of doing this has been using an application that will allow you to click in these metadata fields and manually add or change this descriptive information. For many years now, this data is has been stored inside the file using one of two standardized sets of “metadata” fields named EXIF and IPTC (With RAW file formats, the metadata is sometimes stored in sidecar files).
In the simplest of explanations, the EXIF (Exchangeable Image File Format) holds camera information such as the model of the camera, the date the photo was taken, what type of lens was used, was the flash attachment used, etc. And the IPTC (International Press Telecommunications Council) holds information that helps to inform what's in the photograph, who took the picture, and who is now the caretaker for it.
Even though most other file formats adopted this standardized way of writing metadata into the EXIF and IPTC tagsets, the PNG format has never been built or even updated to correctly take advantage of these pre-established standards. Stated very well by the PNG's Wikipedia page:
The PNG specification does not include a standard for embedding EXIF image data from sources such as digital cameras. Instead, PNG has various dedicated ancillary chunks for storing the metadata that other file formats (such as JPEG) would typically store in EXIF format.
PNG's Current Metadata Support
At the time I published the original version of this article you are reading, I don't believe there was any support for the popular metadata fields with the PNG file. This was the biggest reason I included the word “hurt” in the title of this article — because withholding the ability to safely add descriptive metadata to a scanned photo would definitely be detrimental to your photo collections.
However, through the years since, there has been an increasing amount of “limited” support. And I say limited because it's currently hard for me to quantify for sure how much backing there really is industry-wide (with all the photo software out there) with the PNG's non-standardized metadata implementation.
It's my understanding that instead of writing metadata in the standardized methods inside your files in the way other file formats all follow, when your metadata is written to a PNG file, it's saved in these unique “ancillary” chunks within the file.
An analogy for this might be instead of leaving your car registration and owners manual in the glove box like everyone else, you decide instead to hide it in the trunk under the spare tire. They're there, and technically accessible, but who would even know to look there?
Now, also to my understanding, when you load a PNG file, it's up to the application then to find and then decipher what this “ancillary” chunk of information is and then decide what to do with it. Which means, some applications will support reading and writing known metadata information to PNG files, and some won't.
In the worst case, if you load a bunch of your precious metadata-rich PNG photos in an application that doesn't handle this non-standardized metadata properly, and it not only doesn't read it, but it inadvertently strips them all out or corrupts them. Apparently, unlike the other chunk type called “critical”, data written in “ancillary” chunks are supposed to be safely ignored if the application isn't familiar with how to interpret it. However, it's up to the application's developer to follow and honor this protocol.
So, this is just all to say, be very careful when working with new and unfamiliar applications and workflows if you're saving metadata in your PNG files. I would hate for you to lose all of that information you've painstakingly added.
Adobe XMP Implementation
Adobe is working hard to support their Extensible Metadata Platform (XMP) which is a file labeling technology that lets you embed metadata into files. Applications that support XMP, can also apparently make it possible to exchange metadata from files with metadata in the standardized EXIF and IPTC tagsets.
What this means is if you stick with applications written by Adobe (Adobe Bridge, Adobe Lightroom Classic, etc.), you will likely find a reliable workflow when working with PNG metadata. Also, you will likely find success with other photo applications who are actively trying to be as useful as possible to their customers by also adopting XMP or even any other metadata implementations that have become popular.
If you're moving forward with saving your scanned photos in the PNG format, and you wish to utilize the EXIF and IPTC metadata entry fields found in the best (non-destructive) photo managers, just take this as a stern warning that PNG's are breaking new ground here.
Experiment with your workflows to make sure your metadata is being carried over from one step to the next before moving on. And be sure to have backups of your entire collection just in case something ever goes wrong.
Unexpected File Size Savings
For someone who's seriously considering saving their scanned photos in the PNG format, the main reason they are choosing to deviate from the standardized TIFF format is likely because they've read or heard from someone how much smaller PNG file sizes are in comparison — and therefore will take up much less space on their storage drive.
It is true, the PNG file format can be used to produce really small-sized images when it's used with the right type of image. But, most people will be surprised to discover that when you're saving scanned family photographs, the amount of file storage savings you will gain isn't nearly as much as you might expect.
The “JPEG Comparison”
Most of us are so used to viewing and sending lots of high-quality photographs (images) over the internet that we have probably lost the perspective of what has been given up in order to make this happen so easily.
Likely, what most of us are sharing are small (low resolution) and highly-compressed (file size reduced) images that are saved in a ubiquitous file format called the JPEG (sometimes written as JPG).
It's a miraculous format because the images can appear to the human eye as having quite a lot of detail and yet the file sizes can be tremendously small, making them transfer from one device to another over the internet with ease.
I explained how JPEG's compression is different and the drawbacks you need to be aware of in the video above. So, you're now mostly up to speed with how it works.
Now, take this photograph from my family photo collection as an example. I have it saved and archived as an uncompressed TIFF file and it takes up an unapologetic 25.5 megabytes (MB's) of space on my storage drive.
If I convert this file to the highly-compressed and portable JPEG format, even with the compression setting set to a modest 80% quality (out of 100%), check out how small the file size was able to be reduced down to:
| Image | TIFF | JPG | Storage Savings |
|---|---|---|---|
 | 25.5 MB | 885 KB (.885 MB) | 96.53 % |
| File Settings/Type | Uncompressed | Lossy Compression – 80% Quality |
This photograph was shrunk down to less than 1 megabyte (MB) of file storage space — 96.5% smaller! — yet still maintains a high-enough level of image quality that it's suitable to be passed around via email or social networks and be enjoyed.
The big compromise with JPEG images is that in order for the file sizes to be so incredibly small and therefore “portable,” they have to go through a process called “lossy compression” that uses irreversible “inexact approximations and partial data discarding to represent the content.” However, if done with the appropriate settings, like in the example above, the quality of the image will still be acceptable for the purpose it was created for.
How PNG File Size “Compression” Works
The PNG format is different from the JPEG format in that it attempts to replicate the original exactly. It instead uses “lossless compression” which means your images will be saved without having any of the data (details) of your photo “tossed out.” So, you won't lose any image quality and your images will take up less space on your hard drive than if they were saved with an uncompressed format (like the “uncompressed” version of the TIFF file).
But, it is this lack of image detail loss that keeps the PNG format from producing file sizes as small as the JPEG format can achieve from highly detailed and low-contrast scanned photographs.
As an experiment for this article, I scanned two additional paper prints, each slightly different in dimension and colors. In my scanning software, I also saved them as uncompressed TIFF files. Then I loaded all three of them up in the latest version of Adobe Lightroom Classic and created a second version of each by converting them to the PNG format (with settings maintaining all of the color and bit depth as possible from the TIFF).
Now, take a look at this chart I created that will give you an idea of the file storage space savings of each of the photographs when compared to the “highly-compressed” JPEG version we just talked about.
| Image | TIFF | PNG | Storage Savings |
|---|---|---|---|
 | 34.3 MB | 28.1 MB | 18.08 % |
 | 33.2 MB | 26.8 MB | 19.28 % |
| 24.5 MB | 20.8 MB | 15.10 % |
| File Settings/Type | Uncompressed | Lossless Compression |
You'll quickly notice that unlike the JPEG version that was able to be compressed over 96%, the average PNG compression (storage savings) I was able to achieve was about 17.5% — an average of the three. And specifically, the same image I used in the JPEG example had a compression savings of only 15.10%.
Now, 17.5% certainly isn't a tiny number and shouldn't be discarded as insignificant. But, when we go back to our goal for our archived master images to be the authoritative version of the (analog) originals you scanned them from, is 17.5% really enough storage savings to deviate from the using the go-to standardized file format for saving and archiving scanned photographs?
To me, it's not enough to make it worth it. Especially when considering most of us will be saving our massive scanned photo collections on an external storage drive dedicated to just this one project. And you can get a storage drive with more space than you will ever need for the average photo collection — with even a lot of space left over — for between $50 and $70. A negligible price to archive your irreplaceable one-of-a-kind memories.
TIFF with Lossless Compression Alternative
An alternative option for those wishing to take advantage of the feature-rich TIFF format, but still reap the modest file size reduction found in the PNG, is to save your scanned photos using a variation of the TIFF format that utilizes the same type of “lossless compression.”
To add to this experiment, I saved all of the same scanned images we discussed above in the TIFF format using ZIP lossless compression, and look how the results turned out when compared to the file sizes of the PNG.
They are nearly identical.
| Image | TIFF (Original) | PNG | TIFF |
|---|---|---|---|
| 34.3 MB | 28.1 MB | 28.3 MB |
| 33.2 MB | 26.8 MB | 27.1 MB |
| 24.5 MB | 20.8 MB | 20.7 MB |
| File Settings/Type | Uncompressed | Lossless Compression | Using ZIP Lossless Compression |
I want to thank Peter Fuller all the way over there in beautiful Wellington, New Zealand for providing me with this question. A very interesting one indeed.
And if you have a question of your own you would like me to cover in an article like this one, ask away using my contact form.
Cheers!
From your site and the other information I've found on the net, I think we should scan our photo collections in TIFF, at 600 dpi, using your naming convention/workflow. You don't cover TIFF versus other formats in your articles, but I see you are using that format and there seems to be general acceptance that it is the best format for archiving.
What do you think of the PNG format?
Peter Fuller
Wellington, New Zealand
What about saving jpg/jpeg photos that have been shared TO you and are likely not available as a tiff file? Would you go the extra mile to convert those to tiffs before adding to your collection?
I think it depends, Cathy. The main reason you would ever want to do this is if you go to do a visual edit to the photo, such as color restoration. By saving your edited file as a TIFF file, you will then ensure you are preserving as much data in this newly saved version as possible without losing any additional data to “lossy” compression (as a JPG). If the file is just sitting there in your collection, just being viewed, then there’s less to no reason to convert it.
Thanks, Curtis. I guess that same advice applies to all our more recent digital photos that download as jpgs? Unless you’re editing, leave it be?
To complete my last question, the format I’m using is TIFF and the images consistently turn out mostly black.
When I try to scan old photos with my HP8630 printer/scanner, the scanned image comes out almost completely black. What’s going on here and how do I correct it?
Hello my question is about are there any remarkable the difference between tiff format and bmp format.
Both have nearly 17-20mb space with 600 dpi but I could not decide which one is better for scanning album photos ?
I would go with TIFF because it’s the standard for photo scanning, like I mentioned in this article. Developed by Microsoft for Windows, BMP (Bitmap Image File) isn’t as universally compatible as the TIFF across multiple platforms and operating systems. And I’m not even sure if a BMP holds the standard metadata fields used to store scanned photo information (something you would need to look in to before committing long-term to this file format). Also, almost most importantly, I’m not aware of scanning software that even lets you save your scanned image as a BMP file. So, this would require a lot of file conversion from TIFF to BMP after the scan.
Thank you for your detailed and kindly response, my scaning software (hp printer program that includes scanning) lets both bmp and tiff thats why I could not distinguish for the most proper so the best quality one due to nearly same file space per photo.
If I am not wrong metadata that you mention is not related to visual quality but like exif information. And yes bmp file does not have such data, eventually if there is no serious visual data loss between them, tiff can be preferred for its being prevalent but not for its superior less visual loss ?
Best regards.
i really enjoyed your article, I now understand the differences between PNG & TIFF much better, when I scan an image as a TIFF file it has compression options, I assume these would be lossy ??? So do I need to use a ZIP program to get lossless compression ? Be better if they were still TIFF files but saved with lossless compression, maybe TiffZ as a format……otherwise all my files will be .zip and would have to be extracted first to view them..thanks
I capture an image printed on white paper. Then in a software, place that photo and similar others and see a grey border around them (though the original paper was also white, white balance was also performed). I don’t want to edit the picture at all. How can I avoid this grey border; actually I can cut out the grey border, but inside the photo, the white area are grey; those I cannot edit. I have tried using various formats for addressing my problem but it does not go away.
Useful Information! As a Photo Editor work TIFF image files occupy a lot of space on the storage device and Png due to lossless compression, its file size gets large. I avoid using PNG format for printing photos in any size. I used Stellar Repair for Photo – an efficient tool to repair corrupt photos of RAW, JPEG, and TIFF file formats shot from any camera
Hi! I am currently scanning some old Kodachrome from my grandfather. I choose TIFF so I could edit them on LR … And some of them had no problem when I importe them but today I scanned a few others and when I tried to import them to LR it said that there was a “Preview”problem. What am I doing wrong? hahaha Please help !
Great channel btw. 🙂
Thanks, this is a useful thing I have seen, maybe I will help people with the tool how to convert jpg to png fastest at: http://jpg4png.com
Thank you for this! I was thinking of saving the photos, especially the very old ones in both TIFF and JPEG. TIFF for archiving and repairing, and JPEG for sharing, but this will require an extra step. Is it necessary?
You’re welcome Sophie! It’s not technically necessary to archive your scanned photos as TIFF files, but it’s better to do it this way. The more advanced your goals are, the better off you will be having an uncompressed TIFF master image of your photos to always go back to. And especially if you would like to do repair work later in “post processing”, having the uncompressed data is even more important.
Converting to .JPG for sharing, yes, is an extra step. But, with all of the easy to use photo managing applications and image editors out there that easily allow you to “batch” export out sharable copies, it’s possible it can add up to only a couple extra minutes of your time. Well worth it to archive the best master image you can. 😉
If PNG is lossless..isn’t it er, LOSSLESS? How can I misunderstand this! Why doesn’t VueScan allow PNG files to be saved? I’m confused..or..am I not?
That is a question for Hamrick Software.
Excellent article. Only thing I’d add is to be conscious that Windows file names are limited to 255 characters when you include directories and subdirectories as part of the full name or file designation.
What does the author mean when he says “every time you save a jpeg, you lose some quality.” When would we re-save them?
Since JPG (JPEG) is a lossy format, every time such a file is saved, it loses some data and quality to keep a small file size. Quality loss from repeated saves will often be most noticeable in large areas of relatively uniform color, such as a clear blue sky. Instead of having a relatively uniform blue, the sky will show distinct bands of slightly different values of blue where some of the color data has been discarded. The same applies to the entire image, but is most visible in areas of uniform color such as the sky.
Any time you make a change (edit), such as adjusting the brightness, contrast, color balance, cropping, etc. and save your change, you’re re-saving your file. Since it’s usually recommended to save each change before making another edit, just in case your computer crashes or some other problem occurs, if you decide to make multiple edits to a single image, you could easily re-save that image 4 or more times. Since a JPG loses some data and quality each time it gets saved, if you’re making extensive edits to one, you can easily end up destroying its quality beyond any hope of recovery. If you have only a JPG file to work with, open it in your editing program, SAVE AS a TIF or PNG, before making any edits. Saving in either of those formats won’t recover any data lost when the file was originally saved as a JPG, but it will prevent further data loss with repeated saves. When you’ve finished editing, you can save a copy of your edited file as a JPG if you really want that format.
Its not quite as bleak as you claim, as technically whilever that JPEG is open in your editor then all edits are being done on the uncompressed copy in RAM. No matter how many times you save it, you are only performing one re-compression from that copy in RAM.
It only becomes problematic if you are using batch scripted command-line application to make the edits as those DON’T keep it in RAM between actions, but then the simple answer to that is to make a copy of the original as PNG and do all work on that, only saving to JPEG once finished.
The same applies to an editor, any work-in-progress save should be made as a PNG or other lossless format, only convert to JPEG once you are finished. Or even better, keep it as a PNG so no further quality is lost.
“Its not quite as bleak as you claim, as technically whilever that JPEG is open in your editor then all edits are being done on the uncompressed copy in RAM. No matter how many times you save it, you are only performing one re-compression from that copy in RAM.”
Technically spoken thats nonsense.
Without any knowledge which editor is used there is no possibility to predict if the behavior you assume is a matter of fact. There are in fact applications which will work with the newly saved data after saving the image.
With one of these your quality will subsequently get worse and worse with every saving.
Also there is no “uncompressed copy” in RAM after loading a jpeg.
The damage which was done while saving, cannot be undone while loading from such a file.
Thats the LOSS in lossy compression algorithms.
You will have an uncompressed copy of all the remaining information in RAM and if you save this again after loading from such a file, there will be further loss of information.
Now, wait a minute. Every time a JPG files is saved it loses some data? How can that be? Now if you load it into an application and change a few pixels around and save it I could understand that. But just saving it?
Yes, each time you save – you recompress. Each time you recompress you loose image quality.
Curtis,
Thank you so much. I am an expert in image processing but not in photography and I have learned a lot from your blog. I plan to follow a lot of your advise now that I am about to start scanning my photo collection of more than 30 years, however, I don’t plan to use TIFF (I guess I fall into the “basic” category).
I plan to save my collection “masters” in high quality JPEG, why? Because I know from experience that if I use the correct dpi, the quality is 99.x% of the time excellent (in my eyes anyway). In addition, it takes so much less disk space to store in JPEG which not only affects the amount of storage, but also the time it takes to load a picture to display or to do any post processing. I also plan to keep my “masters” untouched and ready to be used for any post processing. If I don’t like the results of post processing or if later on I want a different kind of post processing, I will start with the “original” JPEG masters to do the work, not the post processed ones. This way JPEG recompression artifacts will not pass beyond one generation and I am sure the results will be very good to excellent. What is wrong with this way of thinking?
This is a misunderstanding! It ONLY loses quality if you re-save as JPEG from an image editor.
Only copying from one place to another, of course, does not change the file! (The file system does not even care what kind of file it is)
:/ nah.. Too much text…
Still would use JPG!
This is one of the most informative articles and blogs I have ever read on this subject. I was pretty well convinced that I will use Tiff as my output but there are a couple of you who suggested that if I have VueScan Pro, I should use dng + jpeg. No one really commented further about that so I am a bit confused. I do have that program and I use Lightroom. I have all my Nikon digital pictures in Lightroom as NEF (Nikon raw)files. LR does allow you to import your files as dng but I did not do that. My intention is to import all of my scanned files into Lightroom as Tiff’s since I have not yet used the dng capability but the contributors to this blog aroused my curiosity. Can anyone comment further, either positively or negatively, regarding the dng + jpeg method?
My apologies for being late responding, but if you haven’t yet realized it, VueScan Pro lets you save as DNG, TIF, and JPG with one scan pass. LR should let you edit your DNGs, which records the maximum data your scanner hardware is able to capture. Since DNG, like your NEF, is read-only, you need to save your edited version in some other format, such as TIF, PNG (if you want to preserve transparency where you may have removed a background), or as a JPG. Note that if you saved a JPG in VSP, it will be overwritten unless you give the edited version a slightly different file name. My usual practice is to save as DNG and JPG in VSP, then save the edited DNG as TIF or PNG.
Hi,Great Article. Thank you so much for sharing!
Cheers!
My choice for archival would be PNG because the format is widely supported and decompression is fast. For work in progress I would use either TIFF or BMP without compression, or PSD.
One reason why PNG is not a good choice for direct output from scanning software is that compression to this format is rather slow. This becomes significant when working with large files. So it makes no sense to spend time compressing a scan that still needs editing (inluding simple edits such as cropping or rotation), especially if compression only reduces the size to around 50% of the original. If the scanner does encoding on the fly without a temporary uncompressed file, and the CPU can’t keep up, then the scanning head would be requird to stop, and possibly create a tear in the image.
If you expect data coruption, then an uncompressed format has an advantage. As long as the file size doesn’t change and unreadable sectors are replaced with other data instead of omitted, then corruption will only affect the pixels that could not be read. A compressed file will be corrupted by a 1 byte error until the end. Recovery might be possible for some formats, but it is not trivial.
JPEG-LS is an interesting format with faster compression and better ratio than PNG, comparable decompression speed, but unfortunately it doesn’t support metadata.
What about tags in the files? Is anybody using this?
In windows explorer there is only one field for PNG and that is ‘date taken’. This cannot contain ‘x’.
For TIFF there are a whole lot of fields to use. Any suggestions on which fields to use for extended filenames?
[date] – [event] – [location] – [people] – [scanner settings] – [numbering system].tiff
The field ‘tags’ seems useful for searchable tags for example all of the persons in the photo, location etc. This is searchable in windows explorer. Can these fields be read by a photo managing software?
I’ve bought a Canon CanoScan 9000F Mark II for my retired mother to scan our photo collection. She hasn’t started yet as I’m still doing research.
I’ve really learned a lot from this site. Thank you!
Glad I found your website. I am so uninformed on this I only understood about a 10th of the discussion. But, I feel as if I have found a source that with study will help me to become more informed before I begin to scan my 30 years of photos of Switzerland. I am eager to start the scanning, but first, the learning. Thanks!
No worries at all Barbara — knowing 10% of the conversation is a really good starting point! This stuff is hard at first. And you certainly don’t need to understand a 100% to get started on your collection.
But, I’m certainly with you if you decide you don’t want to start scanning and organizing until you’ve got a pretty firm grasp on a lot of it. It’s not a good feeling worrying that you may have to redo everything you’ve already done.
So, here’s to the learning… and we’re (me and everyone who visits this website regularly) behind you when you’re ready to finally jump in and get started. :coffee:
Cheers!
Nice answers in return of this query with real arguments and describing everything regarding that.
Hi Curtis,
Thank you so much for all your work on this blog. It’s been incredibly helpful; I am just beginning to archive my family’s photo collections, and have found myself referring back to your site multiple times a day for weeks!
I finally decided to invest in the Epson V550, which is now set up. I’ve decided to use the .tif format (though I am curious about the .dng format and would love to learn more about the pros and cons). From what I’ve read, it seems you use the Epson Scan software, which is what I’m planning on using as well. Right now I’m in the File Save Settings within the program, but I see there are two ‘.tif’ options:
1. TIFF (*.tif)
2. PRINT Image Matching II (TIFF) (*.tif)
I looked up the difference in the user manual, and (if I am understanding it correctly) it says the latter will provide a higher quality scan for printing. First and foremost, my goal is to archive, but I’m sure at some point down the line I will want to print hard copies of some of the photos. That being the case, should I use the latter? Is there any reason to choose the the first option over the second? Any advice and information you can provide is appreciated!
And again – thank you so much for sharing your experience; it’s been invaluable!
I have started scanning my photos using tiff and the ‘extended’ filenames and I’m happy with how it’s going. I do have a question regarding my digital photos though – it seems they are automatically jpg files of about 1-2 mb, obviously I can’t scan them as tiffs because they are digital, not hard copies (unless I have the last 15 years of photos printed!) so I guess they are already compressed and there is not much I can do with them? I have mostly used a point and shoot camera.
Hi Debbie. There’s certainly nothing wrong with .JPG masters taken with your point and shoot camera. Most people shoot and save their masters in this format with this type of camera, so you are not alone. If there’s a setting deep within your camera settings to adjust the quality that it saves them as, then this would well be worth taking advantage of. Many cameras ship with the setting set to the middle — like 5 quality out of 10 — in order to preserver space on the internal memory card. But, bumping that up to or near 10 will improve the quality often by a lot!
So, this being said, if you keep your masters saved in the .JPG format, just be careful when using an image manager with them. Opening/viewing them will not hurt them, but opening and then re-saving them as .JPG files over time will degrade the image quality (even more) each time it re-comresses the .JPG file. So, since quality is important to you, if you aren’t using a non-destructive image manager, and you re-save your edits as a new .JPG file (or to itself, replacing your original image — which I wouldn’t recommend if you want to preserve your original as it was shot), then make sure your image manager is set to save at the highest quality .JPG that you can to reduce the amount of quality loss.
Curtis, the Typing Gremlin got you in the second line of your last paragraph in response to Debbie. It Should read: “Opening/viewing them will NOT hurt them, but opening …” The Gremlin left out the word “not” for you. The rest of your reply is correct. An option would be to save the original, master JPGs as PNGs. That won’t increase any quality, but edits made to the PNGs won’t lose quality with each subsequent Save like they would to edited JPGs. If a particular shot needs extensive edits, adjust exposure, adjust contrast, adjust color balance, cloning or other retouching of several spots, with a Save or Save As after each edit, (a good practice in case the computer crashes and all existing edits since the last save are lost), it’s possible to have 6 or more saves for one file in just one editing session. That many save operations are likely to become noticeable with reduced image quality. That is not a problem with PNGs, which can be saved as JPG if desired when all edits are done.
Art, thank you for the catch! :coffee: The Typing Gremlin most definitely took advantage of my tired state last night, after working for a long time on my taxes! (I just corrected my comment above)
If someone were to convert their .JPG’s to .PNG’s, .PNG files still don’t have a clean workflow with IPTC metadata though correct? Last I had read, they still don’t support this and several other types of metadata to be embedded.
Yes, you are correct about PNG and IPTC metadata, although I believe they can have XMP sidecar files created by programs like Lightroom, PS, PS Elements, ACDSee, or other programs. However, it should be possible to copy the IPTC, EXIF, or other metadata from the original, master JPG into the JPG saved/converted from the edited PNG, so long as the metadata has been preserved in the original JPG. Some of the less expensive or free image editing programs available may not read/write or even preserve such metadata anyway, except possibly within a proprietary database instead of within the individual, actual file. That’s one reason to keep a copy of any and all metadata in a separate RTF or TXT file with a matching file name for each and every image file. RTF (Rich Text Format) and TXT (plain text) files are likely to be readable 1 month, 10 years, 100 years from now in whatever software is prevalent then. JPG, PNG, TIF image formats will also likely be readable with future software, although just as now, not all software will necessarily recognize any existing metadata embedded in the image files, let alone any metadata buried in today’s proprietary databases for programs. To be even more sure of preserving important images and their metadata, as well as archiving such files to various media as media change with technology (e.g. floppy disk, ZIP/JAZ disk, IDE to ATA to USB hard disk, CD, DVD, BD,???) and migrating to each new media generation, use archival paper and inks to print hard copies of both images and relevant metadata and preserve these hard copies in archivally-safe storage containers. That way, so long as humans can see to read, they’ll be able to see any particular image with its metadata, regardless of how much technology changes.
I just want to say a few quick things about PNG…
It actually does support 16 bit per channel, and from my experience any PNG viewer can handle 16bpc PNGs even if it can only display 8bpc.
Epson is a Japanese company, and from my knowledge Japanese businesses are typically quite conservative when it comes to adapting new technologies (especially in computer software*). Therefore, it’s very possible that PNG isn’t supported as a scanned “save as” format simply because it is a newer format (“only” 15 years) compared to JPEG, TIFF, BMP, etc.
Consider that Windows 7’s built-in TWAIN scanning software lets you save as PNG even if you’re doing so with an Epson scanner! (which is what I’ve been doing)
*Example: the 2011 PC version of the Japanese visual novel “Fate/stay night: Realta Nua” uses MPEG-1 (seriously?!) videos at 640×480 with “burned-in” letter-boxing (actual video content only uses 640×360 area).
I have recently purchased the Epson V370 scanner with the intention of scanning my paper photos (again, as I want to scan at a higher (600) resolution). I haven’t started yet, aside from a couple of practice runs to check out the scan software (Epson Scan), mainly because I am not sure which format to scan in. I was going to use Tiff, for the reasons Curtis mentions but after reading this article and all of the comments I am confused, I didn’t even know about PNG, only JPG.
I would consider myself to have advanced goals rather than basic because archiving is important to me but I am relying on the scanning software and the expertise of other user’s knowledge and experience.
I do not have any numbers after my name and a lot of the comments went way over my head so to speak, so back to basics – Curtis do you still prefer to use Tiff 4 years from originally writing this article? I don’t mind spending time scanning and I have a separate hard drive to back up in.
Over the last two days I have read quite a few of your articles and really appreciate the effort you put into answering questions and comments as well as inspiring others to ‘get started’.
I should mention that another reason for deciding on tiff is that I am scanning multiple photos and then need to crop them into single photos. I am assuming that the scan I make is scanned as one photo therefore when I crop them I am changing the dimensions of each photo which is why I think tiff will work better for me as the original sizes will remain the same. Or is there a better way of separating each photo when doing a multiple scan? As I mentioned in my previous post, I have the Epson Perfection V370 Photo scanner.
I also want to add the file names suggested in another article and add comments to each photo. Yep, long days ahead! 🙂
Debbie, with Epson Scan you likely have a choice of saving as TIF, JPG, or PDF. You can choose any one format per scan. Please use the PDF option only for text documents, NOT photos. It compresses your files and makes it very difficult to edit photos. JPG files are always compressed so you will get better quality using TIF. If you invest in VueScan Pro (about $90.00), you can save each scan as a DNG (absolutely all the data your hardware can capture),TIF (loses some data through software processing, even without compression), JPG (always compressed), and as PDF. You have a choice of any single format or any desired combination of formats by checking the options on the OUTPUT tab in Professional mode. I recommend DNG for maximum quality plus JPG for immediate use without further editing other than resizing for web or email.
Thank you Art, you are right about the choices I have to scan, PNG is not an option, and I have started using Tiff. I can crop each photo while it is in preview and then the photos are scanned separately so it works well and I am happy with it. It took about 2 minutes to multi scan 6 photos. I will scan all my photos as Tiff even though I don’t intend to crop or edit a lot of them – YET. I’m sure if I scan certain photos as JPG, they will be the ones that I want to crop in the future, Murphy’s law and all that.
I have taken the wise advice in naming each photo with year/month/day and using xx when required. I am adding name/event/location then settings then giving them a number using the 00001 system, also using _ and -.
At the moment I am just adding these when I save the photo, I had a look at Picasa but it is being retired in favour of Google photos, see here http://googlephotos.blogspot.com.au/
I am about to check Google photos out as I would like to be able to leave a comment on each photo rather than in the name.
Debbie, I agree with everything Art told you in his reply. I just wanted to jump in as well to answer you specifically in my case, yes, I still do prefer to use .TIFF (uncompressed version) over any format such as .JPG or .PNG.
The only exception would be, as Art pointed out, .DNG … if you have the luxury and desire to work with them.
If you choose to use JPEG compression, be sure to disable Chroma Subsampling. Some software will not allow it to be configured, and will either always use it, which is very bad, or tie it to an overall quality slider. Chroma subsampling will halve color resolution, usually in both dimensions, which might not be noticeable on most images, but will be visible on red, saturated parts, such as fire or flowers. You might discover that those sections are blocky when you zoom in to them while editing, when it is already too late.
A scanner is almost equally capable of resolving any color, and delivering a ‘sparkling’ detail, if it is present in the copy.
My software has a dumb TIFF/JPEG writer which would aways use subsampling and produce significantly poorer quality at any setting, compared to plain JPG. Photoshop would open such files with nasty 2×2 blocks.
Quality 98 or Photoshop’s 12 is ‘nearly’ lossless, and an acceptable choice.
Just a note about the file formats you mentioned. This is very good information but I want to add just a little more to it.
I am a retired mechanical engineer and now I do freelance work at home. I use AutoCad 2010 and often go to a customer site to photograph a part of machinery or electronic equipment. I have learned that Autocad allows importing a photo onto a layer so I can then draw over it on another layer. But the Tiff and jpg images can’t be manipulated as much as the png image. Apparently AutoCad has accepted this file format to handle the best in a drawing.
I realize this will apply to very few of your readers if any at all. I just wanted to mention it to you.
Best Regards, Franklin Reid
basic or advance type of person? Curtis…. if we are reading your “sort” articles and posts… or those from Art… (no offense to any of you) we all have to be advance users 🙂
Good site!!!
Keep in mind that TIFF turns 30 in 2016. You should be able to easily open a 1986 TIFF file in todays graphics programs, and possibly even open one of todays images in a 1986 graphics program (that predates Photoshop by about 4 years). So for longevity, its hard to beat that.
So does this mean I have to re-scan all my paper photos and slides that I’ve foolishly saved at .jpg?
Hi Wendy. I don’t think it means you have to rescan all of your photos that you scanned and saved as a .JPG file. As a purest who is teaching people how to secure as much quality as possible from our photos as each of our abilities and hardware allows us, I suggest people to save as an uncompressed .TIFF file when given the choice because it rules out any quality being tossed out from the process of compression.
And please, don’t think of yourself as a fool for saving as a .JPG. This wasn’t a foolish move by any means. Not knowing the difference or even choosing to scan as a .JPG doesn’t make you a fool. They are both very viable formats to save your photos. They just compliment two types of separate goals — one more for those with basic goals, and one for people with more advanced goals.
If it makes you feel any better, know that most scanning services out there where you mail in your photos, they scan them and then mail them back to you, offer .JPG files with all of their basic packages. For most of their customers, .JPG files offer the easiest file format for them to do whatever they want with it without any other knowledge of their computer or software on it.
For example, if all scanning services scanned in 48-bit uncompressed .TIFF file, and sent their customers 80 MB images for all of their analog photos, it’s very possible their customer support numbers might start to light up with people wondering why they can’t post these massive files onto Facebook!
But, for those inclined to learn a couple more steps in the archival workflow, the uncompresseed .TIFF offers a whole lot more room to store and edit your master images with the highest amount of quality and detail as possible.
If you still feel like you missed out on your chance to save as .TIFF files, my personal suggestion and opinion would be for you to scan the rest of your collection from here on out and save as uncompressed .TIFF files. (I keep saying uncompressed because there are versions of .TIFF that save as compressed — so if you have the option, you want to make sure you have it set as uncompressed). Then, when you are finished scanning the rest of your collection and you’re done, THEN if you still want to rescan the first photos you saved as .JPG, then rescan them then.
The most important thing is to get digital copies archived as soon as possible and backed up preferably in more than one location. So, by finishing the scanning first, you’re putting the longevity of your photo collection first, before your personal goals of having all of them saved at the highest quality.
Hello everyone, I have another question for you experts: when I scan booklets of my cd collection and save to PNG and open them in photo managers, the scans appear not in the position you usually read them but in 90° angle to the right. Is there in loss when I rotate the scan 90° to the left with for example microsot office 2010? I’ve noticed the size of the file increasis a bit after a rotate and save and the increase varies depending on what photo manger I use for the task. Should I use any particular software for rotating them to the original “reading” position?
Thanks
Elias, since no-one else has replied, i’ll offer my own not-too expert opinion:
No, rotating any exact multiple of 90 degrees will result in no loss to image quality because images are rectangular arrays of square pixels. (Rotating by any other angle, on the other hand, involves “re-sampling” and thus potential degradation of quality.)
Any slight changes in file size, if it occurs, is likely a result of other factors discussed at length above.
A couple of people with advanced goals suggest PNG is great.
Library of Congress accepts both uncompressed TIFF and PNG as for archival.
http://www.digitalpreservation.gov/formats/content/still_preferences.shtml
National Archives of Australia has chosen PNG as the “preservation format” for bit-mapped images and converts images in other bit-map formats to PNG during its ingest process.
http://www.naa.gov.au/Images/Digital-Preservation-Software-Platform-v1_tcm16-47139.pdf
Library and Archives Canada has adopted PNG as a recommended format for still images.
http://www.collectionscanada.gc.ca/obj/012018/f2/012018-2200-e.pdf
Mariano, thanks for posting these links for us! It’s very interesting to see how new technology and in this case formats are being used and adopted.
I originally wrote this post on “PNG vs. Tiff” some time ago. And the PNG format, as well as others, has grown. So, there is always a need to go back and re-evaluate whether a set of conclusions needs to be modified. Maybe the PNG format has evolved: is now accepted by more scanning software companies as a “save as” option, allows you to save IPTC metadata inside, and you should therefore really consider it as an option in your own personal photo collection. Or maybe it hasn’t.
One thing that jumped out to me after looking briefly at the pages these links took me to, was the question whether .PNG was primarily added mainly as a format “option” because it solves the problem of transport. What I mean is, how would people get their digital photos TO these organizations. Asking people to burn a DVD and mail it through the postal service is asking a lot of people now. But, having someone email or even upload a photo to a website is less of a chore and may generate a greater response. In this case, PNG becomes a great option because it reduces file size making it easier to send from people’s home broadband lines, but still retains the “lossless” quality.
While I agree that asking individuals to snail-mail digital image files to any archive, particularly national ones such as those referred to above, is asking a lot, don’t forget that people likely wishing to donate files to such an archive are probably going to be planning to contribute collections with numerous images, probably numbering in the hundreds at least. Burning an extra copy of such images to sufficient DVDs/BDs (Blu-ray Disks) for the archives is not going to be a major time factor when personal archive disks are being created. Considering the amount of time involved in emailing such a number of files, even as relatively compact .png format, it’s not a project I’d relish undertaking. I’d much rather package a batch of DVDs and take them to the post office to mail them to the recipient. The other factor to consider is the time required at the archive to download emailed files and then transfer them to archival media, be it DVDs/BDs or external hard disk drives. For budgetary reasons, most archives operate with minimal paid staff and hopefully, lots of volunteers, but they’d likely be reluctant to use staff time to sit at a computer downloading a large batch of emailed image files.
Until such time as the .png format is revised to include EXIF information with the image information in a single file, each image will require both its .png image file AND its appropriate side-car file with the EXIF information. It seems to be all too easy to accidentally miss including all the side-car files in a batch of image files if they’re to be emailed, either at the initiating end or the receiving end. If they’re all kept in the same folder on the originator’s computer, they’re much more likely to be missed when burned to DVD/BD.
Certainly, most archives will place a higher value on files with EXIF data, either contained in the image file as with .tif or in a side-car file as with .png, than on files without that data.
These are just a few recent thoughts about the .tif/.png discussion.
Some of the files could not be read at all. This happened a few years ago. Threw the disk away. The cd drive would read and read. And then say disk could not be read. Happened on any drive. Some images I got off perfectly. And then others had damage. Where you couldn’t see some of the image. Or the file wouldn’t load at all.
Hi Graig Smith,
Can you send Curtis screen shots of any of the files from your dvd disks that have gone bad to show what the files look like when you try to view them? I’m not questioning the truth of your comments, just interested in seeing what disc rot actually looks like. Possibly other readers would also be interested in seeing what it looks like so Curtis could post any shots you might be able to provide him as a learning tool for all.
As Malcolm and I both mentioned late last month in this forum, hard disks can become unreadable, even if they’re in mint condition. Over Christmas, I visited a friend who had received a brand new Dell computer but he’s unable to connect the hard disk from his old computer to retrieve his data. The new computer has a totally different, although now standard, interface to control the hard disk(s) and it’s incompatible with the older standard. Fortunately, he believes he has most, if not all, of his images and other data transferred to CD since he’s unable to access the old hard disk. A few years ago, adapters were available to connect old disks to new computer mother boards so they could be used but when I checked online this week, such adapters were not to be found. Again, as Malcolm asks about SCSI drives, will today’s hard disks be accessible in a few years? This is a point worth considering. As with optical media or tape media, serious archival and backup strategies should include plans to migrate all data to new media every few years to retain accessibility.
Art Taylor
“Optical media is preferable as long as the dyes in the media is archival grade and they are store properly…”
i disagree, as i said before, i had dvd disks go bad after only a couple of years. disc rot does happen. even if you have it stored in a dust free scratch free environment which is how i stored my disks. honestly. i have had better luck with hard drives. the only thing is hard drives are prone to catastrophical failures, where you can loose a whole disk at once. which is why i am using online backup. online backup. sends stuff out for backup hourly or so every time you add some images, it sends them off and they are already backed up. time machine isn’t the only backup i use. it’s just an extra. my main backup is crashplan’s service.
“The best media for backup is tape (still).”
i agree. all sorts of large data companies use tape backup. it’s reliable. i bet crashplan uses tape backup on their servers.
An excellent point Malcolm makes is the need for TWO tape drives. This is an issue seldom likely to occur to most of us until it’s too late to do anything about it. I also appreciate his comments about TIFF vs PNG formats. Some valid points are raised.
Thanks, Malcolm for some excellent information about RAID. Any information about a RAID controller failure rather than a RAID drive failure? I seem to recall reading somewhere that a controller failure is at least as likely, if not more so, than a RAID drive failure. Do you know if one could leave the RAID drives powered on, although disconnected from the controller, while a faulty controller is replaced?
Malcolm touches on the issue of hard disk drive controllers when he mentions SCSI (Small Computer System Interface), which is used to be (but no longer are) standard on Macs and was available with an optional SCSI controller add-in card for PCs. The contemporary controller on PCs was IDE (Integrated Drive Electronics), later referred to as ATA, in turn superceded by EIDE (Enhanced Integrated Drive Electronics), then by SATA/PATA, and eSATA (for connecting external drives to the computer’s motherboard. See http://en.wikipedia.org/wiki/Hard_disk_drive#Access_and_interfaces for more information. Today’s PCs often don’t have IDE controllers so older hard disks may pose problems being connected to modern computers, let alone any internal mechanical problems such as Malcolm mentions, such as the lubrication drying out. Today, SSD (Solid State Device/Drives) are becoming common since they use no moving parts. A friend of mine has several older IDE/EIDE drives which he can no longer connect to his current computer since the computer lacks a suitable drive controller. To access any data on these older drives, he has to hook them up to an older computer in his computer room. Since he likes to tinker with computer hardware, he often has such a unit available but not always. For those who have only one computer available, this should be an item to seriously consider before deciding to archive exclusively to hard disks.
While I’ve had no personal experience with an SSD, one acquaintance of mine, who uses a very recent model Mac, has reported no problems with his SSD, although I believe he uses it for data storage rather than as the system disk with the Operating System on it. A friend, who built his own PC from parts a little over a year ago and who uses an SSD for his system disk, reports that while initially it was extremely fast in comparison with SATA drives he’d used previously, over time he has noticed its performance has slowed noticeably so that it seems to be no faster now than a SATA drive, installed at the same time, would have been. This MIGHT be just because he’s added more programs over time and no longer has as much free space available as he had at first. Maybe it’s a (common) feature of SSDs. Since I’ve heard of only these two specific drives, used differently by different individuals, I don’t want to draw any conclusions about the pros and cons of SSDs, just to mention there may be unforeseen differences which might cause one to consider if an SSD is worth the extra money for a comparable size disk instead of a SATA/USB/Firewire drive at present. Perhaps the SSD technology should be allowed to mature a little more before we all rush out and replace all of our present drives with SSDs. At best, we should be prepared to migrate our archived data to newer HDDs over time, as technology progresses, just as we needed to migrate it from Zip/Jaz to CD to DVD, and now likely to Blu-Ray, if we choose to use optical disks as part of a back-up and archiving strategy.
Art
Wow, this article has got a lot of traction, and (unfortunately) a little misinformation (not a lot, but…)
First, on backup: a hard drive (in “time machine” or anything else) is NOT a backup. It is a copy, an offline copy (probably), and that is A Good Thing, but if one is concerned about optical media as a backup format, one should run screaming from any hard disk. The reason for this is that for backup you have be concerned about all sorts of failure modes, including the often overlooked “how will you read the thing in 10 years?” one (suppose you’d used SCSI disks as a backup media a decade ago… now what?). For an HDD (hard disk drive), you need to worry not only about media faults (not being able to read a particular chunk of media) but also motor/bearing/interface failures, which render the whole thing about as much use as a brick. But even worse, because HDDs have critical moving parts, they also need lubrication in order to work…. and that lubrication will evaporate over time. So eventually a disk sitting on a shelf will become a brick, and the only way you’ll get your data back is by spending a lot of money with a disk recovery business, and crossing your fingers.
Optical media is preferable as long as the dyes in the media is archival grade and they are store properly… but the main reason to prefer it is that the media is cheaper (although the time to burn it isn’t, so that must be considered). Because the media is cheaper, the thinking goes, you’re more likely to have multiple copies, and because it is light and relatively robust, you’re more likely to have remote copies, too.
The best media for backup is tape (still). Unfortunately, a sound backup strategy needs *2* tape drives (one to write, the other to verify), otherwise you risk creating backups that only one drive in the world can read… and that drive will be the one that just went up in smoke. The good news is that “cloud backup” services usually include tape-based backup of the storage, so this is (yet another) reason to go that route.
By the way, if you do have a RAID drive fail, the best thing to do is to keep the whole thing running and use it as little as possible; don’t power it, or the attached computer, off until the new drive is installed (or ready to be installed if you don’t have hot plug), because a huge area of complexity for RAID controllers is figuring out which drives to use if, for instance, the formerly dead disk “gets better”. They should get it right, but it’s your data, not theirs, that will be corrupted if they don’t…
Even more by the way: RAID1 is always a better choice than RAID5, except for the pesky issue of cost and connectivity. So six disks arranged as 3 RAID1 pairs is better than four disks as one RAID5 set, given that we’re all made of money! (Oh, wait…).
Second, others have pointed out the fallacy of worrying about lossless compression. Certainly, if you save anything to a cloud, your data is going to be (losslessly) compressed anyway, so this is a non-issue.
Third, the comments about vector formats being supported by TIFF are (at least partially) valid, but entirely irrelevant to a discussion about scanning. Sure, if you want a single file format for multiple types of image, then this is a very reasonable concern, but for scanning, you’re not going to be getting vector formats.
Fourth, Adobe’s DNG is equally irrelevant to scanning, because DNG is a format designed to be a superset of camera sensor raw formats, and a lot of value of raw sensor data is that things like white balance and exposure can be tweaked. A good scanner will control those (the lamp produces the same color light, for example), so DNG doesn’t really have a role to play here.
Fifth, as has been suggested above, the Achilles Heal of TIFF is that it is not fully standardized (the spec is actually copyrighted by Adobe), and (to make it worse) it is a “Container” type format (like AVI or MOV for video). So a TIFF file can have the image data in one of many formats, including lossy compression (JPEG, no less)! Of course, at first this is only a problem if you make it one (for yourself), but as a data interchange format, TIFF is a long way from being ideal. PNG is also a container-type format, but the structure of mandatory-vs-optional elements is much more rigorously defined.
Sixth: PNG’s big strength is in the “P”: portable. the same file can be displayed on a phone, an iPad, any notebook or desktop machine regardless of whether it is a Mac, PC, or Linux (generally), because the rendering logic is built into all modern browsers (in fact, there are three little PNGs just below the text box in which I’m writing this).
Seventh: the fact that any particular scanner does or does not support a given format is not intrinsically evidence of that format’s suitability for the task. Few scanner vendors care much about their software, and with higher-end applications like Silverfast, other considerations may apply (e.g. PNG requires compression, which requires CPU time, so saving files as uncompressed TIFFs is generally faster during the scan phase, although you may lose that time copying around larger files, and so on). Oddly enough, no-one’s mentioned JPEG2000, which is probably A Good Thing as it would just confuse the issue more (it’s another lossless format, but less well supported than PNG, so why bother….). Anyway, the bottom line is that whatever the source (camera, scanner, software) produces may well not be the best format for your personal workflow, and that format may not be the best for archival.
Conclusion: (sorry it’s gone on this long…) there are plenty of subjective reasons to pick a file format, but when it comes to scanning there are no _objective_ reasons to pick TIFF over PNG or vice versa. Personally, I like to use PNG for scans, mainly because it’s also my final output (what gets sent to the printers), with all intermediate files (including the final “master”) are in PSD. If someone batch scans stuff for me, it all gets converted to PNGs on arrival — it only takes CPU time, and I’ve got plenty of that (as long as I’m not rendering video). This works for me, but probably wouldn’t for Curtis, but for my money the portability of PNG wins the day!
(Why, yes, I have been bitten by unreadable file formats and incompatible TIFFs! How _did_ you guess?)
I recommend SSD drives.
Looks like I might have beat Curtis to this response, Nancy. As you’ve discovered, most popular print places and photo sharing sites will convert your uploaded photos to jpg if you haven’t already done so before uploading. No, you’re not really gaining anything in print quality by scanning to .tiff IF you get prints done by an on-line company. You could likely get the benefit of your .tiff files if you put them on CD/DVD or USB stick and go to a place like Costco or other bricks-and-mortar store that offers digital printing.
While your .tiff files will take up more space for archiving, you retain more detail and quality for your archives with .tiffs. EVERY time you make a change to a .jpg and save it, you discard more detail and quality. You’d be better to scan routinely to .tiff, do any desired editing to the .tiff, then save an edited copy to .jpg for on-line printing or photo sharing.
One reason for the on-line companies converting everything to .jpg is simply that .jpg files are much smaller than .tiff files so they upload and download much more quickly and require less server space on their computers.
If your scanning software (EpsonScan and VueScan both do) and image-editing software (Photoshop Elements, Photoshop CSx, and some other programs do) offer the option to scan at 48-bit color instead of 24-bit color, using this option will give better color quality in your scans. While your monitor and any printer won’t be able to reproduce such a large range of colors, they do give better results when you do any editing or color-correction to your scanned images. To benefit from this color range, you must scan as .tiff since the .jpg format is limited to 24-bit color. When you’re done editing, save a COPY of your edited image as .jpg for viewing, sharing, and printing but keep your original .tiff with all the colors in case you later decide to re-edit for any reason. Feel free to contact me via Curtis’s blog if you have any questions about this.
Art Taylor
so…here’s my question…
I’ve scanned both to .tiff and .jpg.
using .tiff on the prints that I want the best scan.
BUT when I upload them to popular places to print…they all end up converting them to .jpg anyway…so I’m not really gaining anything on the print…am I ??
If I scan to .jpg …I archive that file…and only make a duplicate for editing. am I still losing detail?
this is part of the “8 years between scans question for me!”
thanks.
Hey Nancy! Yes, Art did beat me to a response for you (see below) but that’s okay! You’re in good hands with him and his response. I couldn’t agree more with what he just told you.
If the places you would like to upload your photos to for printing only allow .JPG files, then it says they think their customers are okay with compromising some image quality and detail for the sake of price and convenience.
Expounding on what Art said, not only is it more convenient for you as a consumer to upload smaller .JPG files than larger uncompressed .TIFF files, it’s also cheaper for the printing company on their end because they aren’t saving and processing these larger file sizes. Seconds of processing time, Gigabytes of hard drive space, all add up to more money.
And they save on tech support staff as well who would have to be trained to help their customers understand how to even save as .TIFF files and manage their larger sizes on their possibly outdated computer hardware. Overall, for these companies we are speaking of, .JPG is a good level of compromise for them and their customers.
Especially considering even if you showed many of them a printout from both a .JPG or a .TIFF source, many of them may not even care if some of the detail is “soft” looking or “pixelated” in the .JPG version. I mean, It’s a picture of their granddaughter! “She’s so beautiful!” They’re happy. 🙂
And to make sure Art’s point is made even clearer (if that’s possible), let me add that if you do a nice scan on one of your photos, and then save it as a .JPG file, know that at that moment, you are already “throwing out” some quality and detail. Then if you later re-open that .JPG file and make some edits to it and re-save it again as a .JPG, now you have just thrown out even more quality and detail. So it’s not just when you make edits to a .JPG where there is a loss, it’s from the second you settle on .JPG as your file type to begin with.
So, I would say if the highest quality is really important to you, and you would still like to work in the .JPG format for some of your photographs, try and use editing software that will let you save at the “highest quality” possible at each save. This way you will lose the least amount of detail as possible each time.
Often times, this menu item as you are saving is like a slide bar from 1-10 — 10 being highest quality. If it just says “Save as .JPG” for example, how do we know if it’s saving it as a quality “10” or a quality “5”.
Cheers!
Curtis
Art Taylor, I highly highly recommend against using optical media (DVD/Bluray) as a backup medium. DVDs and Blurays are very susceptible to “disc rot,” which is when the disc appears to be rotting from the inside, which results in data degradation. I believe it is due to rusting or some other chemical change happening within the disc. This is an extremely common phenomenon that can affect discs in as little as 2 years (I have seen it myself happening on large numbers of DVD-R/DVD+Rs, etc). Entire boxes of DVD-Rs can rot at the same time.
Please avoid using optical media as a backup medium.
Hi John,
You are the first, and only, person I’ve learned of who has actually experienced the problem of ‘disc rot’. Although I’ve heard and read about ‘disc rot’, I’ve not recognized it IF I’ve had it happen to me. I’ve checked with others and they have also not recognized the problem, if they have it with their discs.
It would be beneficial to other readers of this blog, and to me, if you would send Curtis some photos showing any images damaged by ‘disc rot’ and any photos showing visual indications on either the label side or recorded side of discs exhibiting the problem. If we could see examples of the symptoms of ‘disc rot’, it might help us to learn if we have media with the problem. It would also be helpful to know specific brand names of discs you’ve noticed particularly susceptible to ‘disc rot’. If you’ve noticed any particular brand(s) showing problem after a specific period of time, it would be helpful to know the names involved. Are they name-brand media or generic, house-brands? I’m certain Curtis would be willing to post example photos, especially if you included a description of what they show, so please send him some examples to share with his readers.
I don’t doubt you’ve experienced ‘disc rot’ but without an accurate description of the symptoms, specific brand names, and photos showing examples of the problem, your warning is of limited value to anyone reading this blog.
Art
I said in an earlier post what happened to me. i had burned cd disks and DVD disks. that even though they didn’t have a scratch on them, no dust on them. stored in a protective case for a couple of years they STILL went bad.. i lost some of the data on just a couple of my dvd disks. i permanently lost scans that i couldn’t get back. it was a bummer. and I also do not recommend storing anything in burned disk format. this comes from personal experience. it’s not a safe format to store things in.
“Maybe someday, the ISO folks will provide a revised version that overcomes these shortcomings of the PNG format.”
I doubt it. the intended purpose of PNG was as a general purpose web graphic. though you never know.
raid is a good way to store images but not as your only storage medium. Drives fail. And if only one drive fails like it did, then at least it did it’s job in that situation. But that’s not the only thing that could go wrong with a raid setup. say you get hit by lightning and all your drives hardware gets fried and none of it works. It can happen. or maybe you just get a brownout and power surges, and multiple drives fail. don’t let raid be your only backup, i know you said you had the images on other drives too, which is definitely a good thing. one way to improve your backup even more is to have offsite backup.
One thing i use is a service called Crashplan plus. And it’s not really that expensive when you consider how much it costs to buy a hard drive. crash plan automatically backs up anything on your computer you tell it to backup, and saves it on their remote servers. so even if your house burns to the ground and all your computers and backups are destroyed. later when you get a new computer you can still restore your data with your crash plan account. and they offer unlimited data for your backups. you can buy it for a year, or more the more you buy the cheaper it gets. i bought a 5 year family plan, and even used it to back my mom’s computer up. recently her computer hard drive died and i used crash plan to restore all her data. crash plan is the best of those online backup companies, i can’t recommend it enough. crash plan works on windows mac and linux too.
What Greg says about TIFF vs PNG is correct. As others have already mentioned, the other disadvantages of PNG, aside from needing an XMP ‘sidecar’ file for EXIF and IPTC data, are its lack of support for layers (useful for work-in-progress) copies of Photoshop/Photoshop Elements PSD files; and its inability to preserve color profiles. Possibly these limitations are why Epson and other scanner manufacturers don’t include an option to save scans as PNG files. Maybe someday, the ISO folks will provide a revised version that overcomes these shortcomings of the PNG format.
I didn’t realize the chances of multiple RAID5 drives failing were similar to or equal to those of a single drive failing. Like Curtis, I wouldn’t rely exclusively on RAID5 or any other version of RAID for backups and archives, although I see no problem using RAID5 as Part of a Backup and Archive Strategy.
‘Cloud’ storage may be relatively safe and secure, if not too expensive for large file volumes, but I’m concerned about the provider being there for months or years, then suddenly gone overnight, taking your files with it. As an option for one part of a backup and archive strategy, it’s probably worth considering but I’d hate to make it my only choice. Probably every ‘cloud’ storage provider uses multiple RAID drives, be they RAID5 or other, but what happens to your files if they happen to all be on the single RAID box that fails? What happens to your data if the provider suffers a catastrophic fire, flood, tornado, hurricane, or other natural disaster? Does your choice of storage provider offer geographically diverse storage facilities with redundant storage space in several locations for every customer?
I know both Apple and Microsoft offer ‘cloud’ storage space for free or very low prices and neither of these companies seems likely to vanish overnight. However, for many decades the Minolta camera company produced quality cameras, lenses made using glass ground in-house to their own specs, and introduced many innovative features in their cameras. They also produced several highly-rated scanners, both film and flatbed, over the years. Then one day in about 2005, they sold the camera production facilities to Sony, discontinued the production and support of their scanners, and completely exited the photographic camera, lens, and scanner markets. Fortunately for Minolta camera and lens owners/users, Sony has continued to provide at least some support for the Minolta cameras and the Sony Alpha D-SLRs use a lens mount compatible with the original Minolta auto-focus mount of 1985 so the older lenses can be used on today’s camera bodies. What do you do if you’ve made ‘cloud’ storage your only backup and archive medium and your provider suddenly vanishes, as Minolta did from the camera field? I’d strongly advise against using only ‘cloud’ storage with any provider.
For what it’s worth, here are my comments on file formats for archival images. As Curtis mentions, TIFF files are compatible across a wide range of scanners and computer platforms, (Mac, Windows, Linux) and virtually all professional, and many consumer-level, image editing software programs can read and write both compressed and uncompressed TIFFs. Having been developed before the advent of computers and desktop publishing, they were designed to include both bitmap graphics, like scanned images, and vector graphics created using mathematical formulae in programs like Adobe Illustrator and CorelDRAW. Today, Photoshop, Photoshop Elements, and probably other programs can create vector elements in bitmap files, using the Pen Tool and the Text Tool. TIFFs can include these vector lines and retain their editability, but PNGs need to have any vector lines rasterized or converted to potentially jaggy-edged bitmap elements. PNGs are also limited to 24-bit color depth, unlike TIFFs. While PNG might be OK for archiving original scans, if you are scanning for maximum quality at 48-bit (or 64-bit with the Infra-red dust and scratch channel included), TIFF is the better format. Also, once you have saved a COPY of your original scan and want to perform any editing on the copy, TIFFs will let you save and re-open any adjustment layers, text layers suitable for future editing of the text, any alpha channels or layer masks, and layers in general. While PNGs can preserve transparency, they do not retain layers as TIFFs do.
Back in the mid-’90s, the company that had developed the GIF format started to try to charge licencing fees to any company that wished to include GIF read-write capability in their software. Not surprisingly, this idea met with some opposition. The PNG format was developed to be an open-source format, that any company could freely use. It was intended to become a replacement for GIF, retaining the ability to include transparency as GIF can, while permitting the full 24-bit color depth of JPG. As Curtis has mentioned, JPG cannot include transparency. While it took several years for PNGs to become popular, their lossless compression, higher color depth, and transparency capability are now being recognized and more widely used. Another limitation of PNGs compared to TIFFs is that PNGs cannot incorporate EXIF or IPTC (the What, Where, When, Why, How) data in the image file. Such information needs to go into a separate side-car file which must be kept in the same folder as the PNG or it becomes separated and lost. EXIF and IPTC data can be included in TIFFs. Also, Adobe’s Lightroom, will not readily work with PNGs but has no problems with TIFFs.
Another format I don’t recall being mentioned yet is Adobe’s DNG (Digital NeGative). Since each digital camera manufacturer uses a proprietary RAW file format which is incompatible with any other manufacturer’s format (i.e. Canon/Nikon/Sony, etc. are incompatible and not interchangeable), and various versions within a particular brand may not be compatible with earlier versions from the same company (i.e. version 2 of brand X is not compatible with version 1 of brand X and cannot be processed using the different version’s conversion software), Adobe developed and released the open-source DNG format and encourages all the camera manufacturers to include DNG as a possible format in their respective cameras. The intent is to attain maximum future file compatibility for archives and libraries, much as the Adobe PDF format has achieved and retained near-universal compatibility for documents. Recent versions of Adobe software (Photoshop, Photoshop Elements, Lightroom) and other programs, such as Zoner Photo Studio Pro, VueScan Pro, can open DNG format files. VueScan Pro can also create DNG files as well as TIFF and JPG files as it scans.
Since DNG files, like RAW files, include all of the data captured by the scanner (or camera for RAW files) and do not do any image processing as for JPGs, maximum scanned data and image detail are available for processing at any time in the future. Color balance, color cast correction, exposure adjustment, etc. can be changed at any time using appropriate software. If one is shooting with a digital camera that provides a RAW option, the RAW file can be embedded in the DNG file so the original camera’s captured information is still available and can be edited in the appropriate camera RAW converter software.
My suggestion would be, and my practice is, scan using VueScan Pro, save as DNG and JPG, for ‘quick-and-dirty’ review and editing. Final editing is done using the DNG file. Note that the lower priced VueScan will not work with DNG format, the Pro version is required.
For backing up my photo archives (10s of 1,000s of slides plus hundreds of negatives and prints from over 40 years of personal photography), I burn at least two copies to DVD-R, using the slowest available burning speed. After burning the first copy, I visually check it in an image viewer/editor, usually Zoner Photo Studio Pro or ACDSee Photo Pro, to be sure it burned correctly, before burning the second copy. The second copy is also visually checked after burning. I do NOT use a back-up program to create archive backups. I drag-and-drop or copy-and-paste the original files to my burning software. This avoids any possible compression with backup software and any possible future incompatibilities with future versions of such software.
As newer media, such as Blu-Ray, come to dominate the market, my archives on DVD will be migrated to such media. When Blu-Ray starts to be replaced with something newer, the archives will again be migrated. Anybody have archived images or other files on floppy disks, ZIP or JAZ disks, or Syquest drives or tape drives that is no longer retrievable because the hardware is no longer available?
Art Taylor
yeah, i don’t use either. i mostly take digital photographs, so really i am pretty done with scanning in general. i just use apple aperture, and throw all my raw images into that. it’s a really nice program for the price you pay. And hey, pentax uses lossless compression on their raw files 🙂
i can see your point about an archive though, if you want a museum like level of preservation. but even a lot of backup software will compress the data too.
“So how much risk are we talking here? Honestly, I am not qualified to answer that. But, also honestly, I’ve never had a .PNG file not open correctly — at least not that I was aware of. But why take any chances at all, I mean not even a .00005% chance if you aren’t gaining anything from it.”
i really don’t think theres any risk to using compression. The biggest risks to archival longevity are using incorrect backup strategies. I have lost data before. here’s how it went down. i had a DVD or a CD backup of my data. and my hard drive crashed.. problem 1 i only had one backup. 2 the DVD or cd with some of my images went bad, some of my images were gone. and some others were damaged. in this case maybe uncompressed tiffs would have made a difference to some of my damaged images, but then maybe not. The storage medium was damaged. hard drives crash… and the cd or dvd i used was kept in a dust free case. but it STILL failed.
So now i don’t trust any CD or DVR to backup… i use a hard drive backup, using time machine. and i also use online backup called crashplan. this way if my hard drive crashes, and my time machine drive burns out at the same time, i still have my online backup.
Got the answer I wanted on whether to scan my old family photos in TIFF or PNG format. I think I fall into the advanced side and am willing to spend more time and money (hard drive space) to preserve these photos. TIFF it will be. Now it’s just a matter of what DPI to set it at.
Hey Yolanda. I’m pleased to see that you believe you fall into the “advanced side.” I don’t think you will regret it at all. It may take you a little longer, but in the end your photo collection will be so much better off. I’m scanning about 30 photos every morning now. It’s just part of my daily routine and I love it!
As far as the DPI, you will just need to find the sweet spot you feel comfortable with. For slides, for those with “advanced goals” like us, I feel between 3000-4000 dpi is the best. I am currently scanning mine at 3200 (a pulldown setting in my EpsonScan software) which in 48-bit mode (which I chose) makes them about 80 MB’s a piece. I wrote a post about it called “Q&A: What’s the Best DPI or Resolution to Scan Your Film Negatives?” that might help you with more information.
For prints, for those with “advanced goals,” I found that it really depends on the size of the print as to what the best dpi is. You want the resolution to come out big enough that you can do enlargements later if you want — or extreme crops. I also wrote a post about this that goes into a lot more detail that you might get a lot from. At the end, I present a chart with my DPI suggestions for each sized print. The article is: “The DPI You Should Be Scanning Your Paper Photographs.”
I hope this helps you out! Please let me know how your scanning/organizing goes. Feel free to contact me through my “contact” me page. 🙂
Best ~ Curtis