jpeg versus tif files

[David T T]

I have a question about file formats after a very confusing and frustrating conversation with the owner of my excellent local lab.

This lab defaults to jpeg output from their Frontier film scanner, to the point that even though I request tif files all the time and have stated outright that I never want anything but highest-resolution tif files from my negatives, they will give me jpegs every time I forget to specify tif on the on the order form.

So last evening I got a notice that the color portion of my order has been uploaded to Dropbox. And it's f*****g jpegs. I went to the lab to turn in the rolls from that day's shoot and to request a re-scan of my processed film. He acted like I was being ridiculous for wanting tif files and stated that no one can tell the difference. I told him that it is for the purpose of editing that I want the tif format, and that editing jpegs is like editing an mp3... you always want to work on an uncompressed file. He then told me to convert the jpegs to tifs, and then work on those, saying that this is just as good.

So am I crazy? I always thought that one should always get tif format if editing is going to take place, was I wrong? Any answers would be appreciated, especially if they contain background information as he may argue with me.

 

[Poisson Du Jour]

Ask for, and expect to get, and you don't. That's my experience with high street labs (so bad that I stopped using them decades ago).
If you are not getting the file formats you asked for, I would pack up and take the job to somebody who actually listens and cares. On your part, make sure you specify exactly what you want on the jobsheet!
Tifs can be quite large files and awkward, slow and tedious to edit; about the limit of my work is dust removal and even that is now being replaced by auto-serial droplets.
JPEGs are better in regard to editing, but their lossy nature means the more you edit, the more you save and re-save, the more information is lost from the file. One might complete editing a tif file and then export it as a smaller (print-ready) file (again as tif) and another file as JPEG.

I wouldn't convert JPG to tif. There will be a casualty.

 

[RalphLambrecht]

I have a question about file formats after a very confusing and frustrating conversation with the owner of my excellent local lab.

This lab defaults to jpeg output from their Frontier film scanner, to the point that even though I request tif files all the time and have stated outright that I never want anything but highest-resolution tif files from my negatives, they will give me jpegs every time I forget to specify tif on the on the order form.

So last evening I got a notice that the color portion of my order has been uploaded to Dropbox. And it's f*****g jpegs. I went to the lab to turn in the rolls from that day's shoot and to request a re-scan of my processed film. He acted like I was being ridiculous for wanting tif files and stated that no one can tell the difference. I told him that it is for the purpose of editing that I want the tif format, and that editing jpegs is like editing an mp3... you always want to work on an uncompressed file. He then told me to convert the jpegs to tifs, and then work on those, saying that this is just as good.

So am I crazy? I always thought that one should always get tif format if editing is going to take place, was I wrong? Any answers would be appreciated, especially if they contain background information as he may argue with me.

I think, You are right in expecting to get what you've asked for.a tif is the next best thing torah and can be converted jpg if required but converting jpg to tif is, although possible,kidding oneself. His software may not be able to create tif, hence the resistance. Look for someone else or get your own scanner. He is being a jerk.

 

[Eric Rose]

The Frontiers I've been told can only handle jpg's. Even if he gives you tiff files he has converted a jpg to tiff. A total waste of time and money. For quick proofing I scan in jpg and when I find an image I really want to work on I rescan it as a tiff.

IMHO jpg's are the VW Beatles of the digital world.

If you can't do your own scanning then get some one who will provide a professional scanning service to give you good tiff files for your best images.

 

[David T T]

I've looked into it and it seems that the Frontier sp-3000 can indeed export as TIF files. I wonder if the resistance to providing this file format is that they have to reconfigure the scanner to output TIFs for that job.

It also seems that JPEGs are 8 bit color, while TIFs are 16 bit. So a JPEG converted to a TIF would lose 8 bits of color information. Is this true, and if so wouldn't that be a significant loss of quality?

 

[Les Sarile]

When I have a problematic frame, I do scan in TIF so I can try to recover as much as I can - a last ditch effort. Otherwise, I can work with JPEGs that have very little compression.

 

[bdial]

Getting the lab to do what you ask aside, if you take your first generation jpgs open them in your favorite editor and save them as tiffs or perhaps PS format, it's unlikely you'd be able to tell the difference between doing that and getting your scans in tiff to begin with. That is, assuming the lab is saving in the highest jpg quality (minimum compression).
If you're trying to do huge prints you'll need better scans than what a minilab scanner can do anyway.

 

[removed account4]

Just convert to tif I've done that and it worked fine even with 16/20 prints.
you could always do a controlled test have a file scanned and delivered as a tiff
work on it have it printed and do the same thing with the same file
But given to you by your lab and look at them. don't get too hung up on the
Small stuff don't be a photographer using a loupe and examining micrograin
it doesn't matter
if you CAN tell the difference then buy a nice scanner ( they aren't too expensive sometimes )
and just have the person process them

 

[Billy Axeman]

Jpegs are using lossy compression but when it is saved uncompressed you don't lose any info isn't it?

 

[Bob Carnie]

I've looked into it and it seems that the Frontier sp-3000 can indeed export as TIF files. I wonder if the resistance to providing this file format is that they have to reconfigure the scanner to output TIFs for that job.

It also seems that JPEGs are 8 bit color, while TIFs are 16 bit. So a JPEG converted to a TIF would lose 8 bits of color information. Is this true, and if so wouldn't that be a significant loss of quality?

The frontier that I used to own outputted 40mb largest size which was good quality for up to 16 x20
I believe the frontier does have a raw option where the file is not jpeg.

you are right to expect the highest quality if you pay for it, but remember that the frontier was set up as a mini lab replacement setup and not a full on high quality scanner for large prints.

 

[BrianVS]

JPEG has a lossless mode where all of the components are stored in the image, you get ~70% size savings. There is a 12-bit JPEG standard, but no one uses it. The difference between lossless 8-bit JPEG and 8-Bit TIFF for image quality- none. The difference between 8-bit and 16-bit images, huge. If the TIFF format option is 16-bit, demand it or find another lab. If the TIFF option is only 8-bit, find another lab or buy a good scanner.

 

[Billy Axeman]

JPEG has a lossless mode where all of the components are stored in the image, you get ~70% size savings. There is a 12-bit JPEG standard, but no one uses it. The difference between lossless 8-bit JPEG and 8-Bit TIFF for image quality- none. The difference between 8-bit and 16-bit images, huge. If the TIFF format option is 16-bit, demand it or find another lab. If the TIFF option is only 8-bit, find another lab or buy a good scanner.

Is this also true for B&W saved as sRGB?

 

[Trail Images]

OP, you're correct in requesting other then .jpeg files or converted from .jpeg. I would seek out another vendor with some research and maybe follow up with input from customers of the vendor too. I scan everything in either .tiff or .psd. The very last thing I do for web usage is convert to .jpeg - sRGB.

 

[Ko.Fe.]

You explained them loud and clearly why, but they acted like idiots. You have two options - deal with idiots or find not idiots.
Sorry, I have nothing to say about TIFF vs JPEG1. It is so obvious, you and I are not idiots

 

[BrianVS]

Is this also true for B&W saved as sRGB?

It is true for Monochrome saved as RGB, the JPEG lossless mode will not drop resolution due to compression. The real loss with B&W is 8-bits versus 16-bits to digitize each pixel. Use of 8-bit pixels will result in contouring of the image. With B&W having deeper pixel-depth is more important than with color film. Find a company that provides 16-bit TIFF.

 

[Billy Axeman]

It is true for Monochrome saved as RGB, the JPEG lossless mode will not drop resolution due to compression. The real loss with B&W is 8-bits versus 16-bits to digitize each pixel. Use of 8-bit pixels will result in contouring of the image. With B&W having deeper pixel-depth is more important than with color film. Find a company that provides 16-bit TIFF.

Thanks Brian, I must spend some time studying this more seriously.

 

[BrianVS]

I learned this stuff when my wife had me write custom software for processing her microscope images... almost 25 years ago when these standards were new. The DNG format and most "raw" formats used by digital cameras are based on old TIFF 6.0.

 

[MattKing]

It really is asking a lot of a lab to provide you with TIFF files when you haven't expressly (on the order form) requested TIFF files.
Other than that, it depends a lot on the jpeg files they give you.
I request TIFFs when they are available. I convert from JPEG to TIFF when I get JPEGs from the others, because I prefer to do my editing in a lossless environment.

 

[David T T]

I looked at a TIFF file from the lab. It is apparently 24 bit, 8 bit padding, true color, 16.7 million colors for color and for B&W. Nothing refers to 16 or 8 bit except the "padding."

 

[David T T]

It really is asking a lot of a lab to provide you with TIFF files when you haven't expressly (on the order form) requested TIFF files.
Other than that, it depends a lot on the jpeg files they give you.
I request TIFFs when they are available. I convert from JPEG to TIFF when I get JPEGs from the others, because I prefer to do my editing in a lossless environment.

I'm not assuming they will give me TIFFs without my asking. It's just that another lab, which I'm not using for other reasons, "learned" me after awhile. This new lab just hates providing TIFFs, and they are very good, so that caused me to question myself. However, everything I'm learning supports my decision to in fact request the highest quality lossless format.

 

[Chan Tran]

Get a scanner and spend time scanning. I wouldn't want anyone besides me to scan my negatives.

 

[MattKing]

It's not so much "learning" you.
It is having a system that doesn't consistently follow the written instructions of the customer, because someone in the lab "knows" you.
That is either a really small system, or a system that may result in a lot of redos.

 

[BrianVS]

I looked at a TIFF file from the lab. It is apparently 24 bit, 8 bit padding, true color, 16.7 million colors for color and for B&W. Nothing refers to 16 or 8 bit except the "padding."

24-bits means 8-bits per color, the padding probably means each 24-bit word is padded to 32-bits for longword alignment for reading. So 8-bits Red, 8-bits Green, 8-bits Blue. For monochrome images, the RGB values should all be the same, meaning you are getting 256 (8-bits) levels of grey. By comparison, my M Monochrom digitizes and stores 14-bits per pixel with 2-bits of padding to make each pixel into a Word in the DNG file. I wrote my own DNG processor that reads the original, applies a Gamma Curve, and stores the result as a 16-bit image. When scanning negatives, I select 48-bit color or 16-bit monochrome. It makes a difference, you lose a lot of resolution when digitizing to 8-bits as "all that detail" that would be subtle levels of 16-bit grey get lumped together.

 

[David T T]

24-bits means 8-bits per color, the padding probably means each 24-bit word is padded to 32-bits for longword alignment for reading. So 8-bits Red, 8-bits Green, 8-bits Blue. For monochrome images, the RGB values should all be the same, meaning you are getting 256 (8-bits) levels of grey. By comparison, my M Monochrom digitizes and stores 14-bits per pixel with 2-bits of padding to make each pixel into a Word in the DNG file. I wrote my own DNG processor that reads the original, applies a Gamma Curve, and stores the result as a 16-bit image. When scanning negatives, I select 48-bit color or 16-bit monochrome. It makes a difference, you lose a lot of resolution when digitizing to 8-bits as "all that detail" that would be subtle levels of 16-bit grey get lumped together.

Thank you!

 

[pdeeh]

If BrianVS is who I think he is, I would take his information above almost anybody else on these kinds of matters. (He knows his way around lenses too )