Anyone who's followed the Mark Segal articles knows that he has yet to address Dan's salient point.
George,
In truth, as the author of the two-part article, I think I know a bit more about the feedback than you do. Your use of the word "Anyone" makes your statement factually incorrect. There are some people who don't think I've addressed Dan's salient point. There are far more either who do think I have, or who have accessed the piece and not weighed-in one way or another. All to be expected, and in this complex domain, par for the course.
I think it would be best for this thread to die a quiet death. Just about everything that can be said about the topic has been, and obviously neither 8-bit or 16-bit proponents are going to come to an agreement here.
Agreed?
Sure, although I find it interesting that the original post asks about proof or illustration of 16-bit benefits with reference to Dan. I simply provided files that to my mind illustrate there's an issue in 8-bit not present in high bit. Did anyone try this? Am I wrong? Or should I say, did I at least use an image supplied to anyone, with a set of suggestions in how I rendered the image and what I saw to suggest I'm on or off track? Has anyone done something similar? If so, I'd like to examine those files and procedures. The procedures Mark produced in his papers CAN be tested by anyone who wants to spend the time. YES, its faster to simply believe what Mark or Dan says. Its easier too. gmachen says anyone will know the salient points are not addressed but will not tell us what he means. We should just believe this is the case because he says so. As I said in my first post, I'm not interested in religious debates. Its a waste of time and only upsets people. Now if you want to discuss this WITH salient points, I'm all ears. Thus far, I haven't heard any. Even if everyone knows, I'd like to see some proof of concept. I don't accept either gmachen or Dan JUST because they say it (or, god forbid, actually believe it).
But that's not the point. The point is, the math based on high bit editing can't be argued. But what about actual images? Well in two controversial cases, two people have tried to present their ideas to peer review while one person never has. OK, lets forget about Dan, I've got him pegged. Lets get back to the notion that there is or isn't an advantage to using a high bit workflow. And of course, the quality of Camera Raw and Lightroom are also useful discussions.
As a high bit proponent, give me a reason why I should discount what I'm seeing. Or provide equally good evidence I should only work in 8-bit. This doesn't need to die because we have the tools to resolve the facts, we just have to want to look at them and share them with other peers. That would do the imaging community some good!
Blah blah blah. This thread has overtaken "ludicrous" and is quickly gaining
on "soapbox for the disgruntled". Will you please stop repeating yourself?
It's annoying to see the "New Posts" icon in the "Finishing School" subforum,
only to find that it's another of your rambling "Dan is a poo-poo head"
diatribes. WE GET THE BLOODY MESSAGE! Sheesh.
Blah blah blah. This thread has overtaken "ludicrous" and is quickly gaining
on "soapbox for the disgruntled". Will you please stop repeating yourself?
It's annoying to see the "New Posts" icon in the "Finishing School" subforum,
only to find that it's another of your rambling "Dan is a poo-poo head"
diatribes. WE GET THE BLOODY MESSAGE! Sheesh.
Hey, Bernard. Watch your bedside manner, please.
This is a subject that interests many (8 bit vs. 16 bit). If you have a point to make about the thread, let's keep the sarcasm and the hard edge out of it.
If you have a point to make about the thread, let's keep the sarcasm and the hard edge out of it.
The point is, this thread has become a Dan bashing thread, not an 8 bit
vs. 16 bit thread. There is NO NEW INFORMATION in any of the last 20 or
so posts. There's only Dan bashing by one person who seems bitter about
something. I'm sorry about the sarcasm, but it's all a bit much. And I did made
my feelings about it clear earlier without resorting to sarcasm. Others have
shared my feelings and expressed them as well, yet the bashing and soapbox
politics continue. It's insane and I don't understand why this thread hasn't
been closed yet. Many other threads less offensive and pointless than this
one have been closed in the past. I don't get why this one is being treated
differently. In any case, I'm just going to ignore it, from now on, no matter how
annoying it is to see the "New Post" icon only to find that, no, there's nothing
new here.
By the way - and it's not important to your dead-on observation, and no criticism of you - the Imre Lakatos work to which you refer is "Proofs and Refutations"; the great philosopher of science, Karl Popper, whom I believe also would turn in his grave in the face of some of the pseudo-scientific drivel promulgated by you-know-who, was the author of the not unsimilar "Conjectures and Refutations."
It's been 25 years, but I'm a bit ashamed of having either confused the two, or mistyped the name. Of course, it is Proofs and Refutations that I was talking about. I recommend it to anyone who is interested in the growth of knowledge. It takes Popper's ideas about scientific method and applies them to mathematical discovery. While this sounds dry as dust, Lakatos writes the book as a dialogue between teachers and students. The particular subject is Euler's Theory that in a polyhedron Vertices + Faces = Edges + 2. I know its hard to believe that this would be particularly enlightening or fun to read, but it really is.
It has ceased to be a thread on that subject many posts ago.
Your single opinion duly noted.
The point is, this thread has become a Dan bashing thread, not an 8 bit vs. 16 bit thread.
To be accurate, its been both.
There is NO NEW INFORMATION in any of the last 20 or
so posts.
An opinion and not a factual statement. OK, you're seriously exaggerating.
It's insane and I don't understand why this thread hasn't
been closed yet.
Probably because the people who have control don't believe in censorship for one. No one has yelled fire in a crowded theater. There are files and opinions about 16-bit processing that can still be discussed assuming anyone actually wants to take the time to try. Or fully read Mark's article. There's all kinds of good data in the articles.
In any case, I'm just going to ignore it, from now on, no matter how annoying it is to see the "New Post" icon only to find that, no, there's nothing new here.
David, Andy, I am definitely moving on, as promised. No more posts from me
in this thread. Especially now that Andy's disturtbing avatar is winking at me.
*shudder*
I had not seen your annex 3 before. Again, I think its fine work, and you do show the excellent capabilities of ACR 4.1 for pictures that I would have thought present more difficulties.
If you look at the histogram on the picture of the oranges, it appears that the range of the reds and yellows is higher than the greens, and the blues is in a low range. As I understood it, when the contrast lies in different ranges in the various channels, a channel by channel adjustment should (at least in theory) be able to squeeze out some more detail than the global adjustments that Camera Raw provides. So the question, on the picture with the oranges, is whether a channel by channel treatment of the image, either through separate curves or through some plate blending, would yield even better results than what you have done.
I don't know the answer to this question. I think it would largely be a matter of skill. Before reading your article, I would not have been able to get results as good as you did using ACR. I intend to study the article a bit more to get a better idea of what the tools actually do. I also have my doubts about whether I could get as good a result using Photoshop. Could a highly skilled user of Photoshop get better results than you did using channel based adjustments?
Here's an interesting comparison on this: When dealing with 16 bit v. 8 bit, the 16 bit proponents will often bring up the clarity of the math. The math shows that 16 bit editing is superior, and since we don't know where the break down is going to occur, then it must be better. Here, we get the same people on the opposite sides of the fence. Dan has basically argued that where the channel structure is such that the detail has its contrast in different ranges along the different channels, you can squeeze the most contrast out of the image by doing channel by channel adjustment. As a matter of math and theory, this is clearly correct. Since the ACR tools don't work on a channel by channel basis, they therefore _must_ be inferior as a matter of logic. Whether they, in fact, are inferior is still open to question.
BTW, I think there has been enough signal among the noise of this thread to allow it to continue.
I use 8 bit only becausemy computer is old and limited in CPU cycles, memory and disk space. 16 bit processing takes much longer for me, and after lots of experiments with my own images, I have never been able to see a difference.
I will be buying a new computer soon, and I fully expect the performance improvement to be dramatic. Assuming there is no significant difference between 8 and 16 bit performance, I will start using 16 bit because I will have nothing to lose. Milliseconds don't matter to me, but minutes do.
Eventually, old, creaky machines like mine will vanish altogether, and we can all look forward to working in 16 bit mode as the default--and arguing about the advantages of 32 bits.
If you look at the histogram on the picture of the oranges, it appears that the range of the reds and yellows is higher than the greens, and the blues is in a low range. As I understood it, when the contrast lies in different ranges in the various channels, a channel by channel adjustment should (at least in theory) be able to squeeze out some more detail than the global adjustments that Camera Raw provides.
I assume you're referring to the Histograms in CR/LR, not the later ones in Photoshop. Be aware that this isn't exactly what's going on under the hood. The data is in a linear encoded color space (using ProPhoto primaries). The histogram isn't what you're seeing here but instead based on the sRGB tone response curve on top of these ProPhoto primaries. The Histogram is mildly useful to see both highlight and saturation clipping (better, the visual overlay modes), all the other stuff ain't too useful at all.
Please listen to the Lightroom Podcast #8 with Bruce Fraser, Thomas Knoll and Mark Hamburg and how Bruce feels about the feedback from the Histogram. Should be able to get it here:
Years ago, we struggled to explain what a Histogram was, now people are placing far too much into what they tell us. Often, looking at the image tells us far more!
They also talk about 16-bit and floating point math used in LR.
Thanks for the link to those podcasts. It looks like there is lots of interesting stuff there, even for someone who doesn't have lightroom.
I typically don't have the histogram palette up when I work, and would use it, like the gurus say, just to find the clipping points. The reason I mentioned it is because it showed graphically something that I already knew would be true about a picture of oranges, with fairly saturated bright orange colors: the subject would fall in the upper end of the red channel, somewhere towards the middle in the green, and in the low end in the blue channel. If the scales change, the way those ranges get squashed will change too. But I would think that the relative relationships between the channels would still be there.
It looks like there is lots of interesting stuff there, even for someone who doesn't have lightroom.
If you're using Photoshop CS3 and Camera Raw 4.1, the same applies as LR. they share the same processing pipeline and the controls are virtually identical.
I had not seen your annex 3 before. Again, I think its fine work, and you do show the excellent capabilities of ACR 4.1 for pictures that I would have thought present more difficulties.
If you look at the histogram on the picture of the oranges, it appears that the range of the reds and yellows is higher than the greens, and the blues is in a low range. As I understood it, when the contrast lies in different ranges in the various channels, a channel by channel adjustment should (at least in theory) be able to squeeze out some more detail than the global adjustments that Camera Raw provides. So the question, on the picture with the oranges, is whether a channel by channel treatment of the image, either through separate curves or through some plate blending, would yield even better results than what you have done.
I don't know the answer to this question. I think it would largely be a matter of skill. Before reading your article, I would not have been able to get results as good as you did using ACR. I intend to study the article a bit more to get a better idea of what the tools actually do. I also have my doubts about whether I could get as good a result using Photoshop. Could a highly skilled user of Photoshop get better results than you did using channel based adjustments?
Here's an interesting comparison on this: When dealing with 16 bit v. 8 bit, the 16 bit proponents will often bring up the clarity of the math. The math shows that 16 bit editing is superior, and since we don't know where the break down is going to occur, then it must be better. Here, we get the same people on the opposite sides of the fence. Dan has basically argued that where the channel structure is such that the detail has its contrast in different ranges along the different channels, you can squeeze the most contrast out of the image by doing channel by channel adjustment. As a matter of math and theory, this is clearly correct. Since the ACR tools don't work on a channel by channel basis, they therefore _must_ be inferior as a matter of logic. Whether they, in fact, are inferior is still open to question.
BTW, I think there has been enough signal among the noise of this thread to allow it to continue.
Duffy
Thanks Duffy, glad you enjoyed the Annex. Sorry to be replying after a week's delay but this web forum is not sending me posting alerts even though I activated them. This is my first visit in a week to dGrin.
Channel by channel contrast adjustments must be performed very carefully not to upset colour balance when working in Normal Blend Mode; but there is much more liberty adjusting individual channels in Luminosity blend mode (speaking of RGB images here). That said, not all images - by a longshot - would show much value-added doing this. I learned the channel-by-channel approach years ago, have experimented with it on various images, and frankly find it more trouble than it's worth in most instances. I've looked at the histograms of the Annex 3 citrus image; while their positioning of course differs some, I didn't find using this approach adds value to the result. If you would like to verify this yourself by playing around with this image starting from my raw file, please send me an email address and you may have it - provided your ISP allows you to receive 11 MB of data in one email. One other person asked me for the file some time ago for similar research. I sent it but have not heard back about any outcomes.
...snip...
In both areas, it looks like he might be holding onto views that he developed with technologies that have now been surpassed. 16 bit and Wide Gamut is the wave of the future at the high end of print photography. In a few years we will probably be having the same wars over 16 bit v. 32 bit image editing. The mathematical argument will be the same. The question is at what point do the extra bits cease to matter.
...snip...
Duffy
I work in the film visual effects industry and over the last 6 years I've seen the shift from scanning, rendering and compositing in 12 bit linear, to 10bit log to a little bit of 16bit linear (10bit log gives you a very similar look for a lot less data) and now to the Open EXR format which is a full 32bit floating point image space.
32bit float space for rendering and compositing is a total boon, the level of correction applied to filmed images and rendered assets to fit them together would astound you. In EXR Gammut size isn't an issue as you can have values over 1 for your whites and below 0 to perserve all the shadow detail you could ever want.
In visual effects land the argument a couple of years ago wasn't 16 bit or 32 bit, it was WHICH 32bit float format would win. Open EXR did. Grab the photoshop plugin and some of the sample images for a play. http://www.openexr.com/photoshop_plugin.html
I work in the film visual effects industry and over the last 6 years I've seen the shift from scanning, rendering and compositing in 12 bit linear, to 10bit log to a little bit of 16bit linear
Interesting! How does this affect what you're doing as more directors and DPs start to shoot totally on digital? I've been following the back and forth between Quentin Tarantino & Robert Rodriguez as to what to shoot, (Tarantino might bite someday). I would suspect like Raw data, these cameras capture far more than 8-bit per color.
Photoshop now has more robust 32 bit processing (I think even more in the Extended version which I use but haven't looked at). For HDR, its probably a must.
So yet another industry moves into a higher bit workflow.
Interesting! How does this affect what you're doing as more directors and DPs start to shoot totally on digital? ...snip..
The simple answer is not a lot. It only affects IO (input output) people. mainly because you are coming off tape and not scanning film anymore. They have a long experience in LUTs (Look Up Tabe, kinda like colour profiles that photographers are used to but we'll often have many output destinations to check against) and it's just a matter of creating a LUT for the camera instead of for a specific film stock.
Comments
George,
In truth, as the author of the two-part article, I think I know a bit more about the feedback than you do. Your use of the word "Anyone" makes your statement factually incorrect. There are some people who don't think I've addressed Dan's salient point. There are far more either who do think I have, or who have accessed the piece and not weighed-in one way or another. All to be expected, and in this complex domain, par for the course.
Mark
Sure, although I find it interesting that the original post asks about proof or illustration of 16-bit benefits with reference to Dan. I simply provided files that to my mind illustrate there's an issue in 8-bit not present in high bit. Did anyone try this? Am I wrong? Or should I say, did I at least use an image supplied to anyone, with a set of suggestions in how I rendered the image and what I saw to suggest I'm on or off track? Has anyone done something similar? If so, I'd like to examine those files and procedures. The procedures Mark produced in his papers CAN be tested by anyone who wants to spend the time. YES, its faster to simply believe what Mark or Dan says. Its easier too. gmachen says anyone will know the salient points are not addressed but will not tell us what he means. We should just believe this is the case because he says so. As I said in my first post, I'm not interested in religious debates. Its a waste of time and only upsets people. Now if you want to discuss this WITH salient points, I'm all ears. Thus far, I haven't heard any. Even if everyone knows, I'd like to see some proof of concept. I don't accept either gmachen or Dan JUST because they say it (or, god forbid, actually believe it).
But that's not the point. The point is, the math based on high bit editing can't be argued. But what about actual images? Well in two controversial cases, two people have tried to present their ideas to peer review while one person never has. OK, lets forget about Dan, I've got him pegged. Lets get back to the notion that there is or isn't an advantage to using a high bit workflow. And of course, the quality of Camera Raw and Lightroom are also useful discussions.
As a high bit proponent, give me a reason why I should discount what I'm seeing. Or provide equally good evidence I should only work in 8-bit. This doesn't need to die because we have the tools to resolve the facts, we just have to want to look at them and share them with other peers. That would do the imaging community some good!
Author "Color Management for Photographers"
http://www.digitaldog.net/
on "soapbox for the disgruntled". Will you please stop repeating yourself?
It's annoying to see the "New Posts" icon in the "Finishing School" subforum,
only to find that it's another of your rambling "Dan is a poo-poo head"
diatribes. WE GET THE BLOODY MESSAGE! Sheesh.
http://bertold.zenfolio.com
Hey, Bernard. Watch your bedside manner, please.
This is a subject that interests many (8 bit vs. 16 bit). If you have a point to make about the thread, let's keep the sarcasm and the hard edge out of it.
Thanks!
Dgrin FAQ | Me | Workshops
It has ceased to be a thread on that subject many posts ago.
The point is, this thread has become a Dan bashing thread, not an 8 bit
vs. 16 bit thread. There is NO NEW INFORMATION in any of the last 20 or
so posts. There's only Dan bashing by one person who seems bitter about
something. I'm sorry about the sarcasm, but it's all a bit much. And I did made
my feelings about it clear earlier without resorting to sarcasm. Others have
shared my feelings and expressed them as well, yet the bashing and soapbox
politics continue. It's insane and I don't understand why this thread hasn't
been closed yet. Many other threads less offensive and pointless than this
one have been closed in the past. I don't get why this one is being treated
differently. In any case, I'm just going to ignore it, from now on, no matter how
annoying it is to see the "New Post" icon only to find that, no, there's nothing
new here.
http://bertold.zenfolio.com
It's been 25 years, but I'm a bit ashamed of having either confused the two, or mistyped the name. Of course, it is Proofs and Refutations that I was talking about. I recommend it to anyone who is interested in the growth of knowledge. It takes Popper's ideas about scientific method and applies them to mathematical discovery. While this sounds dry as dust, Lakatos writes the book as a dialogue between teachers and students. The particular subject is Euler's Theory that in a polyhedron Vertices + Faces = Edges + 2. I know its hard to believe that this would be particularly enlightening or fun to read, but it really is.
Duffy
Your single opinion duly noted.
To be accurate, its been both.
An opinion and not a factual statement. OK, you're seriously exaggerating.
Probably because the people who have control don't believe in censorship for one. No one has yelled fire in a crowded theater. There are files and opinions about 16-bit processing that can still be discussed assuming anyone actually wants to take the time to try. Or fully read Mark's article. There's all kinds of good data in the articles.
That's very kind of you, thanks.
Author "Color Management for Photographers"
http://www.digitaldog.net/
Andrew, that goes for you, too.
Please, either talk about 8 vs 16, or move on.
Dgrin FAQ | Me | Workshops
Thanks, pyrtek. We truly value your opinion. We just need to keep it civil around here. Sarcasm doesn't last long here.
Dgrin FAQ | Me | Workshops
Civil, adult debate and discourse ONLY here. Ya'll are very smart peeps - focus on making your points, and not on anything else.
I hope I'm clear.
Portfolio • Workshops • Facebook • Twitter
in this thread. Especially now that Andy's disturtbing avatar is winking at me.
*shudder*
http://bertold.zenfolio.com
I had not seen your annex 3 before. Again, I think its fine work, and you do show the excellent capabilities of ACR 4.1 for pictures that I would have thought present more difficulties.
If you look at the histogram on the picture of the oranges, it appears that the range of the reds and yellows is higher than the greens, and the blues is in a low range. As I understood it, when the contrast lies in different ranges in the various channels, a channel by channel adjustment should (at least in theory) be able to squeeze out some more detail than the global adjustments that Camera Raw provides. So the question, on the picture with the oranges, is whether a channel by channel treatment of the image, either through separate curves or through some plate blending, would yield even better results than what you have done.
I don't know the answer to this question. I think it would largely be a matter of skill. Before reading your article, I would not have been able to get results as good as you did using ACR. I intend to study the article a bit more to get a better idea of what the tools actually do. I also have my doubts about whether I could get as good a result using Photoshop. Could a highly skilled user of Photoshop get better results than you did using channel based adjustments?
Here's an interesting comparison on this: When dealing with 16 bit v. 8 bit, the 16 bit proponents will often bring up the clarity of the math. The math shows that 16 bit editing is superior, and since we don't know where the break down is going to occur, then it must be better. Here, we get the same people on the opposite sides of the fence. Dan has basically argued that where the channel structure is such that the detail has its contrast in different ranges along the different channels, you can squeeze the most contrast out of the image by doing channel by channel adjustment. As a matter of math and theory, this is clearly correct. Since the ACR tools don't work on a channel by channel basis, they therefore _must_ be inferior as a matter of logic. Whether they, in fact, are inferior is still open to question.
BTW, I think there has been enough signal among the noise of this thread to allow it to continue.
Duffy
I use 8 bit only because my computer is old and limited in CPU cycles, memory and disk space. 16 bit processing takes much longer for me, and after lots of experiments with my own images, I have never been able to see a difference.
I will be buying a new computer soon, and I fully expect the performance improvement to be dramatic. Assuming there is no significant difference between 8 and 16 bit performance, I will start using 16 bit because I will have nothing to lose. Milliseconds don't matter to me, but minutes do.
Eventually, old, creaky machines like mine will vanish altogether, and we can all look forward to working in 16 bit mode as the default--and arguing about the advantages of 32 bits.
Regards,
I assume you're referring to the Histograms in CR/LR, not the later ones in Photoshop. Be aware that this isn't exactly what's going on under the hood. The data is in a linear encoded color space (using ProPhoto primaries). The histogram isn't what you're seeing here but instead based on the sRGB tone response curve on top of these ProPhoto primaries. The Histogram is mildly useful to see both highlight and saturation clipping (better, the visual overlay modes), all the other stuff ain't too useful at all.
Please listen to the Lightroom Podcast #8 with Bruce Fraser, Thomas Knoll and Mark Hamburg and how Bruce feels about the feedback from the Histogram. Should be able to get it here:
http://idisk.mac.com/george_jardine-Public?view=web
Years ago, we struggled to explain what a Histogram was, now people are placing far too much into what they tell us. Often, looking at the image tells us far more!
They also talk about 16-bit and floating point math used in LR.
Edited update: Podcast is here:
http://photoshopnews.com/2006/07/07/lightroom-podcast-episode-8-posted/
Author "Color Management for Photographers"
http://www.digitaldog.net/
Thanks for the link to those podcasts. It looks like there is lots of interesting stuff there, even for someone who doesn't have lightroom.
I typically don't have the histogram palette up when I work, and would use it, like the gurus say, just to find the clipping points. The reason I mentioned it is because it showed graphically something that I already knew would be true about a picture of oranges, with fairly saturated bright orange colors: the subject would fall in the upper end of the red channel, somewhere towards the middle in the green, and in the low end in the blue channel. If the scales change, the way those ranges get squashed will change too. But I would think that the relative relationships between the channels would still be there.
Duffy
If you're using Photoshop CS3 and Camera Raw 4.1, the same applies as LR. they share the same processing pipeline and the controls are virtually identical.
Author "Color Management for Photographers"
http://www.digitaldog.net/
Thanks Duffy, glad you enjoyed the Annex. Sorry to be replying after a week's delay but this web forum is not sending me posting alerts even though I activated them. This is my first visit in a week to dGrin.
Channel by channel contrast adjustments must be performed very carefully not to upset colour balance when working in Normal Blend Mode; but there is much more liberty adjusting individual channels in Luminosity blend mode (speaking of RGB images here). That said, not all images - by a longshot - would show much value-added doing this. I learned the channel-by-channel approach years ago, have experimented with it on various images, and frankly find it more trouble than it's worth in most instances. I've looked at the histograms of the Annex 3 citrus image; while their positioning of course differs some, I didn't find using this approach adds value to the result. If you would like to verify this yourself by playing around with this image starting from my raw file, please send me an email address and you may have it - provided your ISP allows you to receive 11 MB of data in one email. One other person asked me for the file some time ago for similar research. I sent it but have not heard back about any outcomes.
Mark
I work in the film visual effects industry and over the last 6 years I've seen the shift from scanning, rendering and compositing in 12 bit linear, to 10bit log to a little bit of 16bit linear (10bit log gives you a very similar look for a lot less data) and now to the Open EXR format which is a full 32bit floating point image space.
http://www.openexr.com/
http://en.wikipedia.org/wiki/OpenEXR
32bit float space for rendering and compositing is a total boon, the level of correction applied to filmed images and rendered assets to fit them together would astound you. In EXR Gammut size isn't an issue as you can have values over 1 for your whites and below 0 to perserve all the shadow detail you could ever want.
In visual effects land the argument a couple of years ago wasn't 16 bit or 32 bit, it was WHICH 32bit float format would win. Open EXR did. Grab the photoshop plugin and some of the sample images for a play. http://www.openexr.com/photoshop_plugin.html
Interesting! How does this affect what you're doing as more directors and DPs start to shoot totally on digital? I've been following the back and forth between Quentin Tarantino & Robert Rodriguez as to what to shoot, (Tarantino might bite someday). I would suspect like Raw data, these cameras capture far more than 8-bit per color.
Photoshop now has more robust 32 bit processing (I think even more in the Extended version which I use but haven't looked at). For HDR, its probably a must.
So yet another industry moves into a higher bit workflow.
Author "Color Management for Photographers"
http://www.digitaldog.net/
The simple answer is not a lot. It only affects IO (input output) people. mainly because you are coming off tape and not scanning film anymore. They have a long experience in LUTs (Look Up Tabe, kinda like colour profiles that photographers are used to but we'll often have many output destinations to check against) and it's just a matter of creating a LUT for the camera instead of for a specific film stock.