Posted by Paul DeCesare on March 15, 2012 at 3:43 PM
There's still some confusion about if 16 bit printing is available and how it can be leveraged. For subtle gradations, I definitely see smoother tones coming off of my Canon PixmaPro 9000 when I use the 16-bit XPS drivers for Windows 7 that Canon provides (keep that XPS acronym handy for a moment, we'll get back to it). But there's a bit more to it, and after some research I think I've found the real story.
We all know Mac OSX supports 16-bit printing, IF your printer and its drivers support it. If they do, USE THEM. This is mainly for folks running Windows and wanting to know what's going on with 16-bit. Windows to this day does not support native 16-bit printing, and there wasn't any easy way to make it happen, that is, until Windows 7 came along. With Windows 7, Microsoft provided an API called XPS Print Path which allows printer manufacturers to write 16-bit drivers for the printers that can support 16-bit printing. So while Win7 natively does not support 16-bit printing, it does allow third-party 16-bit drivers to pass the data through to the printer. If your printer manufacturer site states that their XPS drivers work on Vista and XP, that simply means they will work, that does not mean that you're getting 16-bit throughput, only Windows 7 can leverage that. Anyway, I did download the XPS Drivers a while back figuring they were better (and now I know what on earth XPS means), so I’ve been printing 16 bit on my PixmaPro 9000 all along. That explains why the prints that come off that thing are so much richer in detail, I thought it was just good engineering on Canon’s part... which it IS, partly through the 16 bit driver they’ve written.
But there's a bit more info that's important so you truly ARE sending more that 8 bits of data to your printer: the 16-bit driver and printer must be used with a "16 bit" image file, meaning, there's no advantage to the drivers or a 16-bit printer for people shooting JPEGs, which are ALL 8-bit, whether it be DSLR or point n shoot. If you print a JPEG or any format saved to 8-bit, that 8-bit data will travel to the printer in a big roomy 16 bit container but do nothing to increase image quality. It's like riding in a bus with many empty seats, more comfortable but you still arrive at the same time. Getting more image data to your printer via 16-bit printing only works when you shoot RAW and then open the RAW file in a 16-bit Photoshop or TIFF file (Photoshop files are actually 15-bit with an added channel, not true 16-bit, but still bigger than the RAW source file of 14 bit). So when I tell my 15+ bit Photoshop file to send the 14-bit image that’s housed within it to the printer via a 16 bit driver, THEN I'm getting as much data from my image as possible to the printer. As you can see it's not a true 16-bit printing workflow, especially being that most DSLRs create a 14-bit image, but even 14 bits has a boatload more info than a puny 8-bit image file. So in the end your 14-bit image gets stuffed into a 15+ bit Photoshop file and gets placed on a 16-bit tunnel straight to the printer. All that adds up to better image quality hitting the paper. BTW, I believe with LR it's the same situation, because our cameras at 14-bit are the biggest limitation. In fact the classic 5D is 12-bit, but that's still way better than 8.
Some argue that they've printed both and see no difference, but before that flag gets rammed into the dirt, the test environment needs to be truly verified that all steps were taken to get a near-16 bit data file to the printer. Then an area of the print like a graduated sky needs to be inspected, I don't think a high frequency image (such as an image full of trees/foliage) would reveal the advantages. It's the same as what we see on the screen... you can wail away with sharpening on pine trees, but try boosting saturation on a sky and watch that banding appear. 16 bit printing will gain you a little more wiggle room for things like a smooth sky. Also, using photo paper like Canon's Pro Platinum Ultra Gloss will really reveal the difference because of how sharp that paper prints. The ink barely spreads so the detail is HIGH. If you print to something like Somerset Velvet which is a rag paper that soaks up ink and spreads a bit, you're going to have a smoother softer image just due to that, and maybe an 8-bit workflow and a 16-bit workflow would reveal no difference in that case.
There it is, 16-bit printing 101 or at least as much as I understand it for now, hope it makes sense! If anyone has found info contrary to what I presented here, by all means, pipe up and let us know.
Posted by Mark Stahlke on March 15, 2012 at 4:59 PM
Good post!
To add a data point, my Canon iPF6100 supports 16 bit printing but only through the Photoshop plugin provided by Canon. Last time I checked Canon's GARO driver only supported 8 bit printing. I can see a difference between 8 bit and 16 bit on matte papers like Moab Entrada. For the record, I'm running OS X Snow Leopard (10.6) and PSCS3.
It is my fervent hope that one day I will have a clue what you guys are talking about..
Sandy
Posted by Paul DeCesare on March 15, 2012 at 7:23 PM
LOL! Sandy, like everything else computer-related, all of this gibberish and effort gains you maybe 10% improvement when you iron it all out. Folks on here have been cranking out gorgeous images for years printing at 8-bit, so don't lose sleep over this one.
Posted by Paul DeCesare on March 15, 2012 at 7:28 PM
my Canon iPF6100 supports 16 bit printing but only through the Photoshop plugin provided by Canon
Extremely good point, Mark. The one area I did not cover was the specifics of drivers or the plug-ins that are available, that's a whole separate post, one that you can expound on if you'd like so we understand that part too. My post was to hammer out the data path through Windows, with the assumption that you have your ducks in a row driver-wise or as you pointed out, plugin-wise. Microsoft provides the accesibility through XPS Print Path, but it's up to Canon and HP and Epson and the rest of the photo printer makers to actually DO something with it. I know Canon does obviously, but for the rest I'm in the dark there.
Discussion
4 repliesWe all know Mac OSX supports 16-bit printing, IF your printer and its drivers support it. If they do, USE THEM. This is mainly for folks running Windows and wanting to know what's going on with 16-bit. Windows to this day does not support native 16-bit printing, and there wasn't any easy way to make it happen, that is, until Windows 7 came along. With Windows 7, Microsoft provided an API called XPS Print Path which allows printer manufacturers to write 16-bit drivers for the printers that can support 16-bit printing. So while Win7 natively does not support 16-bit printing, it does allow third-party 16-bit drivers to pass the data through to the printer. If your printer manufacturer site states that their XPS drivers work on Vista and XP, that simply means they will work, that does not mean that you're getting 16-bit throughput, only Windows 7 can leverage that. Anyway, I did download the XPS Drivers a while back figuring they were better (and now I know what on earth XPS means), so I’ve been printing 16 bit on my PixmaPro 9000 all along. That explains why the prints that come off that thing are so much richer in detail, I thought it was just good engineering on Canon’s part... which it IS, partly through the 16 bit driver they’ve written.
But there's a bit more info that's important so you truly ARE sending more that 8 bits of data to your printer: the 16-bit driver and printer must be used with a "16 bit" image file, meaning, there's no advantage to the drivers or a 16-bit printer for people shooting JPEGs, which are ALL 8-bit, whether it be DSLR or point n shoot. If you print a JPEG or any format saved to 8-bit, that 8-bit data will travel to the printer in a big roomy 16 bit container but do nothing to increase image quality. It's like riding in a bus with many empty seats, more comfortable but you still arrive at the same time. Getting more image data to your printer via 16-bit printing only works when you shoot RAW and then open the RAW file in a 16-bit Photoshop or TIFF file (Photoshop files are actually 15-bit with an added channel, not true 16-bit, but still bigger than the RAW source file of 14 bit). So when I tell my 15+ bit Photoshop file to send the 14-bit image that’s housed within it to the printer via a 16 bit driver, THEN I'm getting as much data from my image as possible to the printer. As you can see it's not a true 16-bit printing workflow, especially being that most DSLRs create a 14-bit image, but even 14 bits has a boatload more info than a puny 8-bit image file. So in the end your 14-bit image gets stuffed into a 15+ bit Photoshop file and gets placed on a 16-bit tunnel straight to the printer. All that adds up to better image quality hitting the paper. BTW, I believe with LR it's the same situation, because our cameras at 14-bit are the biggest limitation. In fact the classic 5D is 12-bit, but that's still way better than 8.
Some argue that they've printed both and see no difference, but before that flag gets rammed into the dirt, the test environment needs to be truly verified that all steps were taken to get a near-16 bit data file to the printer. Then an area of the print like a graduated sky needs to be inspected, I don't think a high frequency image (such as an image full of trees/foliage) would reveal the advantages. It's the same as what we see on the screen... you can wail away with sharpening on pine trees, but try boosting saturation on a sky and watch that banding appear. 16 bit printing will gain you a little more wiggle room for things like a smooth sky. Also, using photo paper like Canon's Pro Platinum Ultra Gloss will really reveal the difference because of how sharp that paper prints. The ink barely spreads so the detail is HIGH. If you print to something like Somerset Velvet which is a rag paper that soaks up ink and spreads a bit, you're going to have a smoother softer image just due to that, and maybe an 8-bit workflow and a 16-bit workflow would reveal no difference in that case.
There it is, 16-bit printing 101 or at least as much as I understand it for now, hope it makes sense! If anyone has found info contrary to what I presented here, by all means, pipe up and let us know.
To add a data point, my Canon iPF6100 supports 16 bit printing but only through the Photoshop plugin provided by Canon. Last time I checked Canon's GARO driver only supported 8 bit printing. I can see a difference between 8 bit and 16 bit on matte papers like Moab Entrada. For the record, I'm running OS X Snow Leopard (10.6) and PSCS3.
Sandy
Extremely good point, Mark. The one area I did not cover was the specifics of drivers or the plug-ins that are available, that's a whole separate post, one that you can expound on if you'd like so we understand that part too. My post was to hammer out the data path through Windows, with the assumption that you have your ducks in a row driver-wise or as you pointed out, plugin-wise. Microsoft provides the accesibility through XPS Print Path, but it's up to Canon and HP and Epson and the rest of the photo printer makers to actually DO something with it. I know Canon does obviously, but for the rest I'm in the dark there.
Good reply, thanks!