8 bit vs 16 bit images

[tecboy]

I heard 8 bit image in photoshop and tif could not tell the difference from 16 bit. Is 8 bit better for printing image? 8 bit uses less file space than 16 bit.

 

[ristretto]

Looks the same. 16's better for making edits to.

 

[BigNegs]

The difference only becomes visible in a final print and then only if using a high end printer. The visible difference occurs in the color gamut of the printer and is very much dependent on the quality of the exposure (or scan of film) on a calibrated system. You will not see much of a difference on a computer display. Google search "soft proofing" and you might see what is occurring. If that does not help, try finding a book on fine art printing by Martin Evening or Jeff Schewe in your local library. Their explanations are highly technical and revolve around Adobe Photoshop but are the best that I have read.

 

[Tailgunner]

The difference only becomes visible in a final print and then only if using a high end printer. The visible difference occurs in the color gamut of the printer and is very much dependent on the quality of the exposure (or scan of film) on a calibrated system. You will not see much of a difference on a computer display. Google search "soft proofing" and you might see what is occurring. If that does not help, try finding a book on fine art printing by Martin Evening or Jeff Schewe in your local library. Their explanations are highly technical and revolve around Adobe Photoshop but are the best that I have read.

So does it hurt the image for printing if you convert the processed 16-bit file to Jpeg or do you send you printer company a processed 16-bit Tiff file instead?

 

[Ysarex]

The difference only becomes visible in a final print and then only if using a high end printer. The visible difference occurs in the color gamut of the printer and is very much dependent on the quality of the exposure (or scan of film) on a calibrated system. You will not see much of a difference on a computer display. Google search "soft proofing" and you might see what is occurring. If that does not help, try finding a book on fine art printing by Martin Evening or Jeff Schewe in your local library. Their explanations are highly technical and revolve around Adobe Photoshop but are the best that I have read.

So does it hurt the image for printing if you convert the processed 16-bit file to Jpeg or do you send you printer company a processed 16-bit Tiff file instead?

If you're using a commercial printer and you aren't paying $150.00 for an 8x10 then there's no reason to send them a 16 bit TIFF. Most commercial printers are set up to print JPEGs and that's what you should send them.

If you're making the print yourself with paper than costs $5.00 a sheet and using a state-of-the-art 8 or 12 ink printer and you've built and tweaked careful calibration profiles for the printer and you have the right software and know what you're doing you'll print that 16 bit TIFF.

Joe

 

[Tailgunner]

The difference only becomes visible in a final print and then only if using a high end printer. The visible difference occurs in the color gamut of the printer and is very much dependent on the quality of the exposure (or scan of film) on a calibrated system. You will not see much of a difference on a computer display. Google search "soft proofing" and you might see what is occurring. If that does not help, try finding a book on fine art printing by Martin Evening or Jeff Schewe in your local library. Their explanations are highly technical and revolve around Adobe Photoshop but are the best that I have read.

So does it hurt the image for printing if you convert the processed 16-bit file to Jpeg or do you send you printer company a processed 16-bit Tiff file instead?

If you're using a commercial printer and you aren't paying $150.00 for an 8x10 then there's no reason to send them a 16 bit TIFF. Most commercial printers are set up to print JPEGs and that's what you should send them.

If you're making the print yourself with paper than costs $5.00 a sheet and using a state-of-the-art 8 or 12 ink printer and you've built and tweaked careful calibration profiles for the printer and you have the right software and know what you're doing you'll print that 16 bit TIFF.

Joe

Thanks,

I'm using a commercial printer who caters to Professional photographers...not some place like Walgreens etc. They except all major formats including Tiff files and offer a large assortment of quality paper.

How about metal and metallic printing? Jpegs or Tiff?

Thanks again

 

[KmH]

Pro printer labs like Miller's and WHCC want JPEGs - not TIFFs.

At their end the issue is storage and file size as much as it is the capabilities of the print machines and their RIP software. (RIP = Raster Image Processor)

 

[sscarmack]

I use 8 bit, for some reason in 16bit I get weird colors with gradients and what not and I never looked into fixing it. 8bit works perfect for what I need.

 

[Ysarex]

So does it hurt the image for printing if you convert the processed 16-bit file to Jpeg or do you send you printer company a processed 16-bit Tiff file instead?

If you're using a commercial printer and you aren't paying $150.00 for an 8x10 then there's no reason to send them a 16 bit TIFF. Most commercial printers are set up to print JPEGs and that's what you should send them.

If you're making the print yourself with paper than costs $5.00 a sheet and using a state-of-the-art 8 or 12 ink printer and you've built and tweaked careful calibration profiles for the printer and you have the right software and know what you're doing you'll print that 16 bit TIFF.

Joe

Thanks,

I'm using a commercial printer who caters to Professional photographers...not some place like Walgreens etc. They except all major formats including Tiff files and offer a large assortment of quality paper.

How about metal and metallic printing? Jpegs or Tiff?

Thanks again

Metal prints: JPEG.

You want to use a 16 bit TIFF file to take advantage of the extended color gamut that the additional storage space in a 16 bit file provides. This only matters if the printers are likewise capable of that extended color gamut. Both printers: the hardware device itself is printer #1 and the person operating it is printer #2. Using a 16 bit TIFF file you can use the Adobe 1998 color space which offers a larger color gamut especially in the greens. If the inks or dyes that are being used by the printing hardware can't create those extended gamut colors or the printer (technician) doesn't know how or isn't set up to use that capability then it's not doing you any good to have them in your file; in fact it's a source of potential trouble.

But talk to your printer. For starters just do a gamut check with your printer's supplied ICC profiles. If you don't have a set of ICC profiles from your printer and haven't actually spoken to someone there then: 8 bit JPEG. If you're going to properly take advantage of the capability in 16 bit wide-gamut printing then you're going to have a conversation with the tech people doing your printing and they have emailed you ICC profiles that you have installed on your system. I use MPix when I need a print and I'm not going to do it myself. I have their ICC profiles installed on my computer. Below is a photo I took last week. I processed the raw file to 16 bit TIFF and set the color space to Adobe 1998. Then I did a gamut warning using MPix's metallic print ICC profile. Where you see grey those colors simply are not within the physical capability of their print hardware. They can't print those colors. If I convert the photo to 8 bit sRGB I don't get any grey when I do a gamut warning.

Joe

 

[unpopular]

The difference only becomes visible in a final print and then only if using a high end printer.

If you're using a commercial printer and you aren't paying $150.00 for an 8x10 then there's no reason to send them a 16 bit

Absolutely none of this discourse is true.

Even if a printer accepts a 16-bit file, there is no reason to print at that depth as no subtractive color space will be capable of reproducing the full 16-bit gamut, and I have a hard time belieiving the eye is capable of distinguishing between 1/65536 of a shade, let alone a resulting color shift, though I could be wrong.

It may be possible (and I repeat possible) that some printers use 16-bit files in their colorspace conversions, though, as far as I understand it, this is done by the color management system (CMS), which is perfectly capable of accepting 16-bit files. So unless there are printers that use some proprietary CMS scheme, then the *printer* has nothing to do with it. Though, I have no idea why a printer would make their own color management system when the ICC already has developed an industry standard.

And as I said, again, the CMS is capable of handling 16-bit afaik (under the same rationale that photo printers "print in RGB" [they don't]), so yes there is a theoretical advantage I suppose - but only if the profiles mismatch tremendously.

The one thing that was correct here is that printing likely isn't going to matter. The advantage of 16 bit is and always has been editing. Take for example adding contrast to the shadows of an image. With an 8-bit image there are 128 shades per channel below middle grey, and only 64 shades below half way between middle grey and absolute black. So that only leaves 32 shades on either side, one to push, the other to pull on the curve.

As you do these kinds of compressions shades are naturally truncated, giving the impression of increased contrast. Remember, all edits are destructive. So you start out with only 32 shades, which *is* certainly distinguishable by the eye, when you increase the contrast over this region you are truncating tones on an already limited pallet.

By comparison, 16-bit images have 16,384 tones in the same two regions, which gives much more room to work with before posterizing, which is a result of truncation to extreme.

But this isn't even 100% accurate since most cameras record information at 10, 12 or 14bits. But when you choose to edit in 8-bit, you are deleting information from your file that can be used to make edits, and the difference between 10bits and 8 bits may not sound like much, but in fact it's 25% - which, depending on how you work and look at things, can be pretty significant.

 

[Ysarex]

Even if a printer accepts a 16-bit file, there is no reason to print at that depth as no subtractive color space will be capable of reproducing the full 16-bit gamut, and I have a hard time belieiving the eye is capable of distinguishing between 1/65536 of a shade, let alone a resulting color shift, though I could be wrong.

Start your letter writing campaign now and tell Canon to get that additive ink the h*ll out of their printers.



Canon (and Epson) have been supplying 16 bit printer drivers now for years; at least through half of the 21st century.

Joe

 

[Tailgunner]

Thanks guys, this helps a lot!

I always create a Jpeg file but I also save a Tiff or PSD file just in case I want to change something up later. I just wasn't sure which one to send. Thanks again.

 

[unpopular]

Even if a printer accepts a 16-bit file, there is no reason to print at that depth as no subtractive color space will be capable of reproducing the full 16-bit gamut, and I have a hard time belieiving the eye is capable of distinguishing between 1/65536 of a shade, let alone a resulting color shift, though I could be wrong.

Start your letter writing campaign now and tell Canon to get that additive ink the h*ll out of their printers.

Oh dear.... fortunately I know you well enough (i hope).

anyway, yeah, it doesn't really surprise me that the drivers can accept 16-bit files, but I doubt very much that they are actually PRINTING in 16-bit space, but rather converting to the device space in 16-bit. I thought that this would be done by the system CMS, but now that I think about it, that doesn't quite make sense ... I am not really sure how these pipelines work.

 

[Ysarex]

Even if a printer accepts a 16-bit file, there is no reason to print at that depth as no subtractive color space will be capable of reproducing the full 16-bit gamut, and I have a hard time belieiving the eye is capable of distinguishing between 1/65536 of a shade, let alone a resulting color shift, though I could be wrong.

Start your letter writing campaign now and tell Canon to get that additive ink the h*ll out of their printers.

Oh dear.... fortunately I know you well enough (i hope).

Yep, it's me, and I know you well enough too -- you're one of my favorite people here -- seriously.

anyway, yeah, it doesn't really surprise me that the drivers can accept 16-bit files, but I doubt very much that they are actually PRINTING in 16-bit space, but rather converting to the device space in 16-bit. I thought that this would be done by the system CMS, but now that I think about it, that doesn't quite make sense ... I am not really sure how these pipelines work.

I think our concerns around this topic are somewhat different. First I completely agree that the biggest reason for 16 bit is editing overhead. Otherwise when I think about 16 bit files relative to printing my concern is the gamut of the color space.

It's not a requirement that 8 bit files be sRGB and 16 bit files be Adobe RGB or Pro-Photo but it's at least convention and for some good reason. Assigning the larger color spaces to the smaller data set can cause problems. So then in terms of printing I'm accustomed to think of 8 bit as sRGB and 16 bit as Adobe RGB, but yes, that's not an actual rule.

The newer/newest generation of inkjet printers can now exceed the sRGB gamut in certain regions. I believe it still gets messy because I don't think any of the printers yet completely include the sRGB gamut, but I could be wrong about that as I keep getting older and slower. Anyway, if you've got a new Canon or Epson 8/12 ink printer you have the hardware potential at least to generate some of the color that's only available in the Adobe RGB space. That's basically what I'm talking about. If you have access to one of those printers and are set up to take full advantage of it then you would want to print an Adobe RGB color space file and I certainly think of that as being a 16 bit file.

On the other hand if you're sending your work out to a lab with a LightJet printer and they're all set up to print sRGB JPEGs it's best to send them what they know and expect.

Joe