16bit printing...your thoughts?

[natas]

Hey guys,

I am in need of an upgrade, my current mac is just not cutting it. Editing 40mp files is killing me but I am having a hard time deciding on what to do. The mac pro's are way behind and from everything I have read out there and update is not coming anytime soon. So I was thinking of switching back to a PC....but now I am having second thoughts because I will no longer have 16bit printing support for my Epson 7900. Should I allow this to be a show stopper? Does 16bit printing really make a huge difference.

I know you guys are going to come back and say to just do a test but I am color blind (red+green) so I may not be able to see the difference.

[digitaldog]

Does 16bit printing really make a huge difference.

At this time, no. At least sending good high bit data to the printer (of which only a few will pass that high bit data). You do want to work in high bit! But sending the data to the few drivers that can take all that data, its difficult to see a difference in all but the most demanding images (stuff with very smooth gradients). And then, you need to have your nose touching the prints.

[natas]

Thanks DigitalDog. I always work in 16bit when editing

[Mark D Segal]

Excuse me - MacPro way behind? Way behind what, how? Please tell me. I'm using a mid-2010 MacPro with 24GB of RAM, 24 effective core processing capacity, 2.63 GHz clock-speed and this thing is blazing fast and rock solid even processing 1 gigabyte files. There are PCs that can do this too, if you want all the performance-degrading baggage that comes with virus, spyware and malware protection you need for those systems (but by the way we may soon be needing for Mac too, as the cyber-criminals have now developed a toolkit for hacking into OSX).

No - from what I've read and observed of my own work sending images to the 4900 in both the 8-bit and 16-bit print options, one doesn't really need a 16-bit print path for a conventional RGB to inkjet workflow, because the heavy-lifting where 16-bit is important happens at the image preparation phase.

[neile]

Ditto to what everyone else has said. Here's my blog entry where I did side-by-side comparisons: http://www.danecreekfolios.com/blog/2010/7/6/is-there-a-difference-printing-16-bit-instead-of-8-bit.html.

Neil

[natas]

Excuse me - MacPro way behind? Way behind what, how? Please tell me. I'm using a mid-2010 MacPro with 24GB of RAM, 24 effective core processing capacity, 2.63 GHz clock-speed and this thing is blazing fast and rock solid even processing 1 gigabyte files. There are PCs that can do this too, if you want all the performance-degrading baggage that comes with virus, spyware and malware protection you need for those systems (but by the way we may soon be needing for Mac too, as the cyber-criminals have now developed a toolkit for hacking into OSX).

Well the new i7 Sandybridge processors are fast...really fast. Test show that the i7-2600 beats the 6core Xeon in PS CS5 test. The new motherboards also support 32Gigs of ram. . Plus you can get this system (with 24Gigs of ram) for less than $1600 compared to $3699 base (3gigs of ram).

Yes a dual 6core is going to beat the i7 in CPU test but I seem to recall that in CS5 test it will not. That's all I plan to use this machine for. CS5, Lightroom and printing.

I am willing to pay the extra for a Mac System (I have 3 Mac's in the house) but Apple needs to do something.

And I guess I went a little to far by saying way behind ...but its really hard right now to shell out that much money for a system that is actually slower at double the price.

Yes the Spyware and all that jazz is a major issue for me...this is why I have all macs in the house. I am just really annoyed that I cannot get the latest system from Apple without having to get an Imac.

[Mark D Segal]

Not the place or time to get into a Mac-PC thing, but when I did a rigorous feature-for-feature comparison last year between buying a Dell and a Mac with as closely matched a feature set I could assemble, the price difference was minimal. Yes, build-it-yourself solutions are bound to be cheaper, but then you need to build it yourself and trouble-shoot it accordingly. (HP/Compaq and Gateway were not compared.)

[natas]

Not the place or time to get into a Mac-PC thing, but when I did a rigorous feature-for-feature comparison last year between buying a Dell and a Mac with as closely matched a feature set I could assemble, the price difference was minimal. Yes, build-it-yourself solutions are bound to be cheaper, but then you need to build it yourself and trouble-shoot it accordingly. (HP/Compaq and Gateway were not compared.)

That may of been the deal last year, but as of right now it's not the case....well unless you build a system with Xeon's

It would be awesome if Apple came out with a new line. I understand the need for Xeon's in servers but for guys like us (unless you get the 2 cpu systems) I really think its pointless. If Apple came out with a desktop machine with i7 CPU's in a tower I think they would sell much better than the Mac Pro's. Price them in the $1000-2500 range. I bet gamers would even go for them.

Anyway I guess I really need to think this over. I am really having a hard time deciding on what to do. Part of me wants to just suffer 6 months and wait for the next Mac pro and the other part of me wants to just go to Dell's website and get a i7-2600 for $1600

[Scott Martin]

There's some debate about what printers can actually process high bit data. Processing data at levels higher than 8 bits is computationally and financially expensive. Do $300 printers really have the hardware needed to process 16 bit data that the driver is now capable of sending to it? Highly unlikely.

Canon is the only manufacturer that makes a big deal about their onboard beefy L-COA processor being able to handle high bit depth data. They are quite open about this conversation and admit that their testing showed a "sweet spot" at 12 bits and have chosen to implement 12 bit processing on their iPF printers for that reason. When I asked Epson's and HP's product managers about this directly they were quick to say "No comment" although their evangelists are quick to say "Yes all Epson printers, even small ones can process full 16 bit data."

As always, real world testing trumps theory and the typical marketing oriented talk that permeates our industry. From my own and my client's testing, we are only seeing 16 bit printing improvements on Canon iPF printers and then only slight as seen in gradation smoothness. Interestingly enough, XRite's new i1Profiler has proven that profile quality has more effect on output smoothness than bit depth. With an excellent i1P profile you'd be hard pressed to say that an 8 bit granger rainbow isn't totally smooth on an 9900, for example.

As others have pointed out, high bit depth development is the key area where working at high bit depths really matters. Capture would be a close second and printing appears dead last as far as importance goes. Best to prioritize having an superb output profile there.

[natas]

Thanks for the reply Onsight. That was very helpful!

[Scott Martin]

It would be awesome if Apple came out with a new line

Thunderbolt/Sandybridge iMacs released this morning, new Mac Pros to follow. Thunderbolt is going to make real differences for us. Even if we can't get the drives till next month it's really nice to know that Apple's already up to speed. Fingers crossed for a matte surface, Thunderbolt 27" Cinema Display....

Thanks for the reply Onsight. That was very helpful!

As my taco joint waiter often says to gringos like me "grassy-ass!"

[Sven W]

Does 16bit printing really make a huge difference?

Not at all on RGB. But when printing true (one channel) B&W in 8 bit, you play
with only 256 shades of grey. Todays top printers with their three or four black inks can produce more than 256
shades and you'll see a small improvement in 16 bit (65536 possible shades).
And for the flexibility to always edit in 16 and just print it.
/Sven

[hjulenissen]

I believe that paper can (at best) do a 100:1 contrast, and that 8 bit using ideal gamma encoding is in principle enough for that range to avoid visible banding.

However, there are practical deviations from my idealised case. If the printer does heavy internal processing and/or processing inside the 8-bit domain is done without proper regard to dithering, there may be reasons to do >8bit for specific combinations of gear.

-h

http://www.poynton.com/PDFs/GammaFAQ.pdf

At a particular level of adaptation, human vision responds to about a
hundred-to-one contrast ratio of intensity from white to black. Call these
intensities 100 and 1. Within this range, vision can detect that two intensi-
ties are different if the ratio between them exceeds about 1.01, corre-
sponding to a contrast sensitivity of one percent.
To shade smoothly over this range, so as to produce no perceptible steps,
at the black end of the scale it is necessary to have coding that represents
different intensity levels 1.00, 1.01, 1.02, and so on. If linear light coding is
used, the “delta” of 0.01 must be maintained all the way up the scale to
white. This requires about 9,900 codes, or about fourteen bits per compo-
nent.
If you use nonlinear coding, then the 1.01 “delta” required at the black
end of the scale applies as a ratio, not an absolute increment, and
progresses like compound interest up to white. This results in about 460
codes, or about nine bits per component. Eight bits, nonlinearly coded
according to Rec. 709, is sufficient for broadcast-quality digital television
at a contrast ratio of about 50:1.
If poor viewing conditions or poor display quality restrict the cont

[BartvanderWolf]

I believe that paper can (at best) do a 100:1 contrast, and that 8 bit using ideal gamma encoding is in principle enough for that range to avoid visible banding.[

However, there are practical deviations from my idealised case.

Indeed. Do not forget that the image data has yet to undergo a conversion from the document's colorspace to the output modality's colorspace. That can be a perfect oppostunity to introduce posterization in smooth gradients.

Cheers,
Bart

[alain]

Indeed. Do not forget that the image data has yet to undergo a conversion from the document's colorspace to the output modality's colorspace. That can be a perfect oppostunity to introduce posterization in smooth gradients.

Cheers,
Bart

Hi Bart

Doesn't that normally happen inside print software going van 16 bit (if the image is 16bit) and only at the last moment converting to 8-bit and then sending that to the printer?

[deanwork]

So the Canon 16 bit plug in is pretty much a wasted effort in most cases?

john

[Scott Martin]

So the Canon 16 bit plug in is pretty much a wasted effort in most cases?

No, the printing plug-in has tons of other benefits. If you're printing from Photoshop anyway, the plug-in and it's streamlined workflow, is the way to go.

[deanwork]

Scott,

I know the plug in is great and avoids a lot of problems, I was referring to the 16 bit capability.

j

[Scott Martin]

I know the plug in is great and avoids a lot of problems, I was referring to the 16 bit capability.

Well like I said above you can see slight differences but they are just that - very slight. Modern drivers also support a 16 bit path so the Canon PS plug-in isn't the only 16bit party in town anymore.

[BartvanderWolf]

Hi Bart

Doesn't that normally happen inside print software going van 16 bit (if the image is 16bit) and only at the last moment converting to 8-bit and then sending that to the printer?

Hi Alain,

Presumably, yes. One can only hope for the sensible thing being done, first a colorspace conversion then followed by a rounding down to 8-b/ch data in case of an 8-b/ch pipeline. However, when starting with an 8-bit/channel source that is an opportunity lost. The only thing left to somewhat mask the potential posterization is clever dithering or adding noise.

Cheers,
Bart