QTR, 3800, 8 bit/16 bit question

Hi all...

Normally as one of the last steps in making my digital negatives, I convert to 8 bit before printing the neg in QTR.

Has anyone printed a 16 bit neg in QTR, instead of converting to 8 bit in PS?

NOw, of course, I know that the result will still be 8 bit, but is it mandatory to convert in PS or can I feed a 16 bit file to the printer?

Any downside to that, like longer printing time, longer file prep time, not as clean as doing it in PS, etc...?

Thanks!

PVia said:

NOw, of course, I know that the result will still be 8 bit, but is it mandatory to convert in PS or can I feed a 16 bit file to the printer?

Any downside to that, like longer printing time, longer file prep time, not as clean as doing it in PS, etc...?

Thanks!

Click to expand...

I am not familiar with QT rip, but what makes you think bit depth has anything to do with printing? Printers use ink dots of pure colors of the subtractive color system (the ink cartridges of your printer - Cyan, Magenta, Yellow, and any other intermediates your printer may have), layed down in a specific pattern, to create the illusion of a range of colors, and color gradients...

But in reality, non of these intermediate colors exist! Only the pure colors as are in your printers cartridges, are the ones being layed down onto the paper.

This means the concept of bit depth, as a means to encode the precision with which the original images RGB (Red, Green and Blue) channels are to be stored in the file, DOES NOT apply to printed images.

There is no such thing as the "bit-depth of a printed image", only a color gamut, which represents the total range of colors a printer and inkset can simulate for the average human eye on a specific paper sort, by using the printer drivers software routines designed for that task.

However, for any of the tasks before the actual printing, bit-depth does play a role, and the bigger, the better. 16bit should potentially give you smoother gradients, and at least allows you to do bigger adjustments in PS with less risk of damaging the image, and causing non-smooth gradients in it.

In my experience, 16 bit images do not print excessively long compared to 8 bit, even for very large images, so that should not be a reason to withhold you from using them.

Marco

I've never converted to 8 bit with QTR, so all my output has been 16 bit TIFFs. You've made me curious though, I'll have to try some 8 bits files to see if there's any benefits w/ printing times- I spool fully to the printer before it starts printing, so might shave some time off that.

I believe that Marco is mostly right in that the difference between 8 and 16bit is that a 16 bit image is the same gamut sliced more thinly. An 8bit image may be evident in a print with a long gradient of mostly pure colour as banding may become visible. Most images are best created in 16bit and converted to 8bit by the rip or prior to printing as all printers to my knowledge are 8bit.

Well, damnation, I'm cautiously elated. Just for laughs I reprinted some problem digial negs in 8bit and I'm getting much less dither artifacts than with the 16bit files.. Funny, when I used RNP color negs I always converted to 8 bit, but just spaced that step when switching over to QTR. Many thanks for bringing this up Paul, you might have solved a long standing irritation with highlight artifacts.

*Forgot to mention- the print spool times were in fact reduced quite a bit.

Colin Graham said:

Well, damnation, I'm cautiously elated. Just for laughs I reprinted some problem digial negs in 8bit and I'm getting much less dither artifacts than with the 16bit files.. Funny, when I used RNP color negs I always converted to 8 bit, but just spaced that step when switching over to QTR. Many thanks for bringing this up Paul, you might have solved a long standing irritation with highlight artifacts.

*Forgot to mention- the print spool times were in fact reduced quite a bit.

Click to expand...

As a normal part of my routine I always convert 16 bit files to 8 bit before printing. This is the very last thing I do before printing, after making all other corrections. I save a 16 bit master in the event that I need to make more corrections later.

My understanding is that if you not convert the file in Photoshop from 16 to 8 bit it will be done by the printer interface anyway so I think it makes more sense to do it before sending it to the printer as printing will be faster.

I compared results a few times by sending the file to the printer as 8 and 16 bit and frankly did not see any difference.

Sandy

Well, much might be at work. OS, profile, working space, QTR version etc. It's plain to see in my situ that the PS-converted 8bit files have a much cleaner dither than the 16 bit sent straight to QTR and I'm thrilled, whatever the reason.

Colin...

Thrilled to hear about that...!

What dither algorithm are you using? I use Adaptive Hybrid...

I need to work out a QTR curve for gum printing. My first few experiments have been very good, but I'd like to customize one. My palladium curves are working out great.

Paul

Thanks Paul. Right now I'm using ordered, but I only recently switched from adaptive hybrid, just to try something new. I couldn't honestly tell much difference when I first switched but should try adaptive again with the new bit depth too.

Gum is on my list too- the tricolor work posted here has been out of this world! Look forward to hearing of your progress there.

Colin...

This is my first tricolor gum: http://www.flickr.com/photos/viapiano/3682203223/

Oh, man- that's incredible! Love the subtle colors- exactly why I can't wait to try the process. Thanks for sharing, that's inspirational.

Colin Graham said:

Thanks Paul. Right now I'm using ordered, but I only recently switched from adaptive hybrid, just to try something new. I couldn't honestly tell much difference when I first switched but should try adaptive again with the new bit depth too.

Click to expand...

I recently switched from adaptive hybrid to ordered and am not able to see any difference at all in result in printing with digital negatives. I use un-directional printing at 2880 dpi, with PK.

Sandy King

sanking said:

I recently switched from adaptive hybrid to ordered and am not able to see any difference at all in result in printing with digital negatives. I use un-directional printing at 2880 dpi, with PK.

Sandy King

Click to expand...

Last night I wrote a long post in response to this topic quoting several messages written by Roy Harrington. Unfortunately I dropped the reply when I posted it.

Here is part of the post again:

Question: "Do you plan to support 16 bit printing with QTR? Just so I understand what's going on better, are you dithering when you go from 16bit to 8bit?"

Harrington: "On the Mac its done by Photoshop (or the print system), on the PC yes I dither in the conversion of 16bit tiff files."

Harrington: "I will probably do 16-bit as some point. But I really don't think it will make much difference. All the ICC internals are already 16-bit and if you
have a 16-bit input file all the profile processing is already done in 16-bit.
The downsizing to 8-bit is done as the last step and its basically just
dithering the output in two stages -- 16 to 8 bit and then 8 to printer data."

Harrington: "QTR resamples all input files to 720x720 pixels. Then dithering is done to whatever dpi you've selected -- mostly 2880x1440 or 1440x720. A couple printers also have 1440x1440.

The resampling to 720x720 is done by the OS on the Mac so I'm not sure
of the algorithm. On the PC it's a bi-linear algorithm -- basically two linear
interpolations -- horizontal then vertical. As far as I know all the Epson drivers do either 360x360 or 720x720 resampling. In general the smaller printers tend to do 720x720 and the larger ones do 360x360. If the driver has a Finest Detail switch that gives the user a choice between these resamplings."

Harrington: "If your RGB file is monochrome there's no difference between RGB and grayscale as far as QTR is concerned. A color RGB file will be converted to grayscale by QTR but its a default conversion so I'd recommend converting yourself for the best results.

If you use Photoshop, edit all in 16-bit. As long as you convert to 8-bit using the intelligent dithering that PS does it makes no difference whether you send 8-bit or 16-bit. It boils down to gray 8 bit is much smaller file that RGB 16-bit for the same results."

Question: "What are the differences between the choices [dithering]? I've been using Ordered but just noticed there are others. What effect will Adaptive Hybrid or Hybrid Floyd-Steinberg have? I'm on a PC."

Harrington: "The different dither algorithms are a holdover from gimp-print. They are available so I've allowed people to use them.

The only ones I've used and I'd recommend trying are Ordered and Adaptive Hybrid. Ordered is much faster and for everything I've seen produces great results. Adaptive works very differently so the dots will be in different places. Some people have noted better results in rare cases -- probably a specific printer banding issue where is just good to do something different."

A lot more of this kind of information can be found on the Yahoo QTR group. I would encourage anyone using QTR to join the group and spend the time doing some searching.

Don Bryant

Thanks, Don!

PVia said:

Any downside to that, like longer printing time

Click to expand...

One last tip:

If printing speed is an issue, one thing you might consider, is a dual harddisk setup. Generally, printing involves both reading and writing to harddisk. By putting the Windows spool directory (the directory where the temporary file is created for printing) on a second hardrive, separated from the drive where you store your images and that may be used for swapping purposes by your operating system as well, you can get a considerable increase in pre-printing processing time.

I have my Windows XP installation write the spool files to a second backup disk in my computer, and it makes a clear difference in speed.

Here is a Microsoft article that explains how to change the Spool directory:
http://support.microsoft.com/kb/308666

Don't know how it's done on Mac...

Marco

PVia said:

Colin...

Thrilled to hear about that...!

What dither algorithm are you using? I use Adaptive Hybrid..

Paul

Click to expand...

This is an information regarding QTR dither algorithm I gathered on the net a while ago.
I don't recall the original poster, tribute to him/her, here the quote:

The dithering algorithm recommended in the documentation is Ordered as this should provide the best balance between speed and precision of detail rendering. In general I have found this to be true, but I have found that for certain images with a lot of fine detail the Adaptive Hybrid dithering algorithm renders the details (and overall contrast) much better. Why dont I use adaptive hybrid all the time? Because in images with less detail, or more subtle gray transitions, adaptive hybrid dithering can give the print a rather over sharpened look to my eye anyway. In these cases the Ordered algorithm seems to work well.

Sidney