How to test if a print was made at 16 bit?

[Mark D Segal]

Sorry, which check box? Can you post a screen capture?

Got it - here you go.

[elliot_n]

Is this where I found out I'm doing something wrong?!

[Mark D Segal]

I don't understand how apps relate to the Epson print driver.

Is the print dialog in ACPU a barebones version of the Epson print driver?

If it is, then when we use it to print Doug's ball target (or a slither of it, to save ink), at 8 bit and at 16 bit, we thereby demonstrate (to the OP) that there is a visible difference in Epson's 8 bit and 16 bit printing pipeline.

We can't do this with Photoshop*, because when printing from Photoshop, the Epson driver utilises Photoshop's dithering algorithm, which makes an 8 bit print of a 16 bit file looks identical to a 16 bit print. (*Unless we switch off dithering in Color Settings.)

Right?

(All my test prints from Photoshop have been done using 'Photoshop Manages Colors'. Whether 'Printer Manages Colors' handles 16 bit gradients in the same way is another can of worms.)

Let's be clear: ACPU is used only for printing profiling targets - or at least that's the only use for it I'm aware of. Do not use it for normal photographic printing. It does not open Photoshop. No Photoshop settings are relevant. It handles 16-bit data between itself and the drivers of supported printers. It moves the file onto paper somehow neutralizing colour management so that only the native behaviour of the printer and the target media are being characterized. This is what is needed for building a printer profile.

[Mark D Segal]

Is this where I found out I'm doing something wrong?!

I see nothing wrong there - looks like best practice choice of options.

[elliot_n]

I think I'm clear about ACPU's use. Before today, I've only ever used it to print profile targets.

What I'm curious about is that if you use it to print a 16 bit gradient, and don't check '16 bit', then the resulting print displays banding. This doesn't happen if you do the same thing when printing from Photoshop.

[Mark D Segal]

But it's not meant for printing gradients, why even try. I'd just use it for its intended purpose and forget about playing with it for unintended uses; I value my time - a lot.

[elliot_n]

But it's not meant for printing gradients, why even try.

Doug Gray suggested it.

I'm trying it because I think it demonstrates to the OP that there is a genuine 16 bit Epson printing pipeline (rather than a cosmetic 16 bit checkbox that does nothing).

The reason that the 16 bit checkbox appears to have no beneficial affect when printing from Photoshop, is that Photoshop's 8 bit pipeline utilises a very sophisticated dithering algorithm that successfully masks any banding.

[digitaldog]

Is this where I found out I'm doing something wrong?!

That is 100% Epson. I have it on my 3880 but not on my newer P800 and as mentioned, the 3880 only on Mac OS had it hence my question to Mark about the newer printer he owns. But he’s on a Mac too. So I would love to know if someone on Windows has such an option.

[Mark D Segal]

Doug Gray suggested it.

I'm trying it because I think it demonstrates to the OP that there is a genuine 16 bit Epson printing pipeline (rather than a cosmetic 16 bit checkbox that does nothing).

The reason that the 16 bit checkbox appears to have no beneficial affect when printing from Photoshop, is that Photoshop's 8 bit pipeline utilises a very sophisticated dithering algorithm that successfully masks any banding.

That makes sense.

[Doug Gray]

My copy prints 16-bit images, no problem. What makes you say this and why would Adobe not allow 16-bit images to be printed?
Just output my Gamut Test File in high bit to the 3880 using ACPU.

Yeah. Mine prints 16 bit images too. Nothing I said suggests it doesn't print 16 bit images. It just converts them to 8 bits first and produces exactly the same banding as printing BillsBlackBall in Photoshop w/o dither and no color management to my 8 bit driver.

I just did it on my Windows 10 x64.

And selecting dither fixes it.

Keep in mind this is hard to see. You have to print the Ball out, trim the white edges, and view it under a strong light - a flashlight works pretty well. But only illuminate the print. if you have anything nearby that the strong light hits it will keep your eyes from adapting.

Recall that creating an image that can actually detect 16 bit v 8 bit printer paths is not easy. In fact the bands are only .4 dE00 apart. This is why they aren't normally visible. However, if you follow the instructions to remove the white borders and look at the print under a strong light, this has the effect of multiplying the dE00 separations and makes the banding visible. This is also why to check a monitor one has to get rid of the white background and ideally use a black one.

The dE00 metric assumes one is viewing two patches, close but not touching, against a neutral L*=50 background. The relative brightness overwhelms the banding. But viewing against a black background brings it out because the eye adapts to the lower illumination and smaller dE00s become magnified and visible.

But this is with Windows and it's possible the Mac ACPU differs.

[digitaldog]

Yeah. Mine prints 16 bit images too. Nothing I said suggests it doesn't print 16 bit images.

I must have misunderstood this:
Adobe Print Utility only works with 8 bit images.

[Doug Gray]

I must have misunderstood this:
Adobe Print Utility only works with 8 bit images.

The full quote was this:
Adobe Print Utility only works with 8 bit images. I'm just guessing but there is a simplified TIF reading C utility shown in libtiff. It can read virtually any tiff format file but has the downside that it converts anything not in 8 bit to 8 bit.

My bad. Poor phrasing. It only prints images using 8 bit paths. It converts 16 bit ones to 8 bits then prints it. This is very annoying if you have a 16 bit image and are trying to print fractional RGB values. OTOH, I1Profiler converts fractionals to ints before printing or saving a tiff so it's also limited to 8 bit precision. Such is life in our little narrow color management niche.

[Doug Gray]

Here's a scan of the 16 bit circular gradient BillsBlackBall printed w/o color management using 3 different techniques to an 8 bit driver.

Top one: This is using Photoshop. No dither selected. No color management.
Middle one: Same as above with dither selected.
Lower one: This is with ACPU.

Two images are provided showing the 3 prints. One is the B&W scans unadjusted. The second has contrast boosted very strongly using curves.

This also illustrates why you need to view the trimmed prints under a strong light w/o any nearby bright objects to see the banding visually.

Again, it's worth repeating that if one sees banding from an 8 bit driver with any normal image one enters unicorn territory.

[elliot_n]

The full quote was this:
Adobe Print Utility only works with 8 bit images. I'm just guessing but there is a simplified TIF reading C utility shown in libtiff. It can read virtually any tiff format file but has the downside that it converts anything not in 8 bit to 8 bit.

My bad. Poor phrasing. It only prints images using 8 bit paths. It converts 16 bit ones to 8 bits then prints it. This is very annoying if you have a 16 bit image and are trying to print fractional RGB values. OTOH, I1Profiler converts fractionals to ints before printing or saving a tiff so it's also limited to 8 bit precision. Such is life in our little narrow color management niche.

This is not how it works on a mac.

With ACPU and an Epson 3880 there is a 16 bit checkbox. One can thereby choose between an 8 bit or a 16 bit print pipeline.

With the 8 bit pipeline, your ball image prints with visible banding (subtle, but clearly visible under a bright light - no need to trim off the prints white edges). With the 16 bit pipeline the image prints smoothly.

 

[Alan Goldhammer]

I must have misunderstood this:
Adobe Print Utility only works with 8 bit images.

There is so much confusion on this thread that it is difficult to figure out what to comment on.  ACPU will work with 16 bit images.  I can prepare profile targets using ArgyllCMS with either 8 or 16 bit imaging but I suspect that for my Epson 3880 with the Epson driver on Win 8.1, that the final print is done using an 8 bit pipeline. 

As I noted earlier in this thread, there is an ongoing discussion on the ArgyllCMS email group.  Graeme Gill posted this comment yesterday which might be of interest (perhaps he will weigh in on this thread if he has time):

Yves Gauvreau wrote:

> At first, I kind of wondered if in a forensic kind of way it was
> possible to distinguish between a 8 and a 16 bit print. It seem, it
> could be a waste of time to even try. But I’ll still get myself an I1
> Studio and try my best to get the most out of my work at least color wise.

should be possible, but it depends on how much trouble you are prepared to go to. i.e., rough outline:

If you assume that the gamut of a printer roughly spans 150 delta E units, then 8 bit precision amounts to 0.6 delta E if the space was perceptually uniform. So you might aim for a measurement precision of 0.2 delta E or better, so as to be able to distinguish 0.6 delta E steps.

Say you pick a test color which you guess will be sensitive to quantization, and print and measure it 10 times, and then calculate the standard deviation as 0.9 delta E. To get a standard error of 0.2 you will need to measure and average (0.9/0.2)^2 = 20 or more samples.

So then you would print a series of test patches,
(say) 10 that span two 8 bit precision values, print and measure each 20 times (total 200 patches to print and measure on 10 sheets) and plot the average of each 20. If the graph is continuous, you can assume that the printing system has better than 8 bit resolution. If the graph has one distinct step in it, then the printing system has 8 bit resolution.

Some of the quandaries are that the space may not be very uniform, and you really want to test a color that is most likely to be quantized than not.
You also can't really be certain if a quantization step you have found is in the the input to the printing system or is generated within the printing system itself in the process of screening etc.

Cheers,

Graeme Gill.

[Alan Goldhammer]

It's Epson's - at least with the SC-P5000 printer. I don't know for other printers but I would assume likewise as long as the printer supports it.

For the Epson 3880 printing under WinOS with the most current driver, only 8 bit printing is supported.  There is no checkbox for 16 bit printing.  I cannot remember whether for THIS printer 16 bit printing is/was supported under MacOS.  It may be that the current Epson printers support 16 bit printing under WinOS but I don't have one of those.

[elliot_n]

I cannot remember whether for THIS printer 16 bit printing is/was supported under MacOS.

Er, read my previous post.

[faberryman]

For the Epson 3880 printing under WinOS with the most current driver, only 8 bit printing is supported.  There is no checkbox for 16 bit printing

Why do you say that only 8-bit printing is supported? Maybe without a checkbox, 16-bit is now always selected. That certainly makes more sense. There never was any good reason to uncheck the box to begin with.

[Alan Goldhammer]

Er, read my previous post.

Sorry, I posted before reading your post.   I just checked the Epson 3880 manual and 16 bit printing is supported for MacOS but not WinOS.  Here is the statement about MacOS and 16 bit printing,

If you are printing 16-bit color miages, select 16 Bit Channel for most accurate transitions and gradations.

Of course the software you are using to prepare the image must save it as a 16 bit image.

[elliot_n]

Why do you say that only 8-bit printing is supported. Maybe without a checkbox, 16-bit is now always selected. There never was any good reason to uncheck the box to begin with.

I think Doug Gray has demonstrated that that is not the case. His print has banding, whilst mine does not (when '16 bit' is checked).