Tony, I never said that 14 bit increased the amount of photons reaching the sensor. The obvious way to increase "greater accuracy" is to provide the sensor with as many photons as possible. If you reduce shutter speed, open your aperture, or provide your own light you can accomplish the more photons to sample for better accuracy...and never even touch the ISO setting. I just wanted to clear up that misconception of my previous remarks.
TonyBeach wrote:
Using 14 bits does not increase the photons reaching the sensor, it's just a matter of more accurately recording them.
Different RAW converters do make different use of the greater accuracy offered by the extra bits, and the potential for future developments to eek out more from the files always exists.
We're going around in circles now, because I already concede that the advantages of 14 bits are mostly theoretical. I would note that Thom Hogan believes it makes a difference in the files that he says he sees but would be hard pressed to show. Anyway, I'm sticking with my bottom line offered in my previous post in this thread....Show more →
400? i see differences at 100. i'll do some test shots this weekend.
Herb...
James Markus wrote:
Herb, I ask you then to shoot a static subject at the exact same exposure, and post process using the exact same adjustments at both 12 bit and 14 bit and post the results. Plus post the raw files. So I can see these better shadow details that are buried so deep. If you deliberately under or over expose...then add one shot at the correctly metered exposure. How about at a common iso like 400...tripod...identical framing and lighting. I'd love to see the results.
The other difference I see is that continuous shooting (fps) slows down in 14bit. Makes no difference whether I have a very fast UDMA card or an old honker. That's a problem if you like to shoot birds, like I do.
As a technologist by training and profession, I lean towards getting as much information into the system as possible. Whether one can perceive that extra data, or even do something with it is debatable. But I try not to leave anything in the field, just in case.
Thanks everyone for the detailed technical explanations; I for one appreciate them in part because I know it takes quite a while to write it all down and check that it's right.
I chased up Thom Hogan's D3x review and he sees very little advantage in the 14 bits over 12 bits. However, I see a huge disadvantage in the greatly reduced frame rate. I reckon that high frame rate can be useful in landscape photography as well as action stuff because it allows bracketed exposures to be captured in a shorter period of time.
The 7D may be doing a better job than previous technologies considering the pixel density involved, but I gather that it still has less absolute DR than say the D700. One day Canon will just use the technology to improve the DR rather than the pixel density, which is pretty much what Nikon did with their D3 and D700.
about 2FPS vs 5FPS isn't something likely to affect my HDR shooting since many of my important landscape shots are well over a second. i have done HDR shooting where the longest exposure is 8 minutes.
Herb...
Alan321 wrote:
I reckon that high frame rate can be useful in landscape photography as well as action stuff because it allows bracketed exposures to be captured in a shorter period of time.
Alan - as one (or several) conclusions already should have shown, the difference only exists in certain conditions - you have to include quite a lot of the shadows in your pictures by actively choosing to lift them to be able to see them. Some people (like me, and apparently Herb too) like to do what many almost would say border on "one-shot-HDR" by including as much of the recorded picture information as possible in the finished output.
Once more - There ARE definitely differences to the 14-bit mode's advantage, the question is if the picture styles that you use make any use of them. The D3x is an extreme example where both the increased NR ratios in the darks and the amount of fps lost are both very "big". Cameras like the D300 show some difference, but not even close to the D3x amount. The difference 12/14 in the D300 is (relatively) bigger at higher ISOs, but not by very much.
i frequently run HDR tonemapping on single captures to increase the shadow contrast by lightening the lighter areas of the shadows. i also always use the Linear Tone Curve setting in Lightroom/ACR so i control just how the shadows fade to black. my prints have large amounts of shadow detail even though they may be quite dark in certain areas. this is going to be more and more a requirement as cameras reach and then exceed film's dynamic range. i can see a medium format back or something equivalent having more than 15-stops right out of the camera. to fit that onto the measly 5-6 stops available on paper already requires at least a little tone mapping. add another two or more stops and it can't be avoided. i can see those sensors producing 18 or 20 bits of output with information in all of them.
Herb...
theSuede wrote:
Some people (like me, and apparently Herb too) like to do what many almost would say border on "one-shot-HDR" by including as much of the recorded picture information as possible in the finished output.
I've done comparisons, and 14 bit does make a nice improvement in files vs 12 bit Raw, at least for Canon, especially at high ISO's. There is no reason not to use 14 bit since it doesn't affect FPS or buffer size or anything at all on Canon cameras.
Does FPS really slow down on the Nikons with 14 bit? Brutal.
Yes, unfortunately it does, due to the 12-bit out limitation of the Sony Exmor sensors. The system has to do multiple reads to get 14-bit accuracy, and this slows things down a bit...
This system does give some other things back though - like almost zero banding noise, and the fact that even the cheapest Nikon D5000 has the same/better blacklevel-noise than any Canon in existence, the 1D(s)-series included... :-) This is what gives Nikon cameras the DR-figures they have.
The D3/D700 use another system, and therefore slows down a little bit less. This is only natural - if you want to quadruple the sampling accuracy (+2bits) you have to do the A to D conversion process a bit slower if you don't want the higher resolved signal to be so polluted with noise/misquantization that the extra two bits are virtually useless anyway. You don't want a more resolved but noisier signal - then you could just as well leave it at the previous level.
Noise in AD-converters is roughly proportionate to the numbers of Mbits per second the system has to process... Now IF Canon had chosen to lower readout speed a little in 14-bit mode, noise performance close to black would most certainly be a fair bit better - and DR a bit higher.
Oops, sorry about that - you're totally right. I notice no difference when I checked now - I thought going to 12-bit was going to increase fps to 7.5-8. Checking the manual, the battery grip is the thing that makes difference on the D700. No difference on the D3.
The point has been made but IMHO should be "amplified:" you can't see "it" (benefits of 14 vs 2 bits-color space, tonal range whatever) on your monitor! This is a little different from the extra resolution that is claimed for prints-not apparent on a monitor. So, yes, you may see something in a print after gross adjustment but how can you manipulate tonal range and color space when you can't see to adjust it?
Where 14-bit helped me out most was with smoother sky gradations in my twilight shots (deep twilight shots in particular)...12-bit looks too "posterized" for my tastes. Also, I found that when I need to bring up my shadows there is more to play with with a 14-bit image.
Where 14-bit helped me out most was with smoother sky gradations in my twilight shots (deep twilight shots in particular)...12-bit looks too "posterized" for my tastes. Also, I found that when I need to bring up my shadows there is more to play with with a 14-bit image.
I prefer 8 bit but only recently. I have 40k plus NEF's on my computer. Most are 12 bit, but a good 10k of them are 14 bit.
But to answer your question with the D700 I can't tell any difference in a print (24"x36") between 12 and 14 bit. On my monitor I think I can, but I'm not sure. Now when I have non photographers look at my prints, I have yet to find one that can tell the difference between jpeg and 14 bit. Just and observation, not trying to start a discussion.
OK, finally I get it. NEF files are allowing me to make much better conversions to Black and White. 14 bit gives me even better flexability in the conversion.
Yeah, it takes me a while, but eventually I get it, and shoot in RAW (12 bit) w/ the D2X, and 14 bit with the D700. Oddly, I like the files from the D2X better than I do from my D700, which I admit is simply a matter of taste, rather than a technical observation. Like my preference for Nikon over Canon or others.
but i can see it even on an 8-bit monitor. that's because i know the information is there and how much and as i manipulate my black point, white point, and curves, i get more out of the image. the reason i shoot 14-bit mode is to bring out more shadow detail with less noise. i use a linear tone curve and try to set my black point so that almost nothing is clipped. those extra two bits become much more noticeable when shadow detail is the name of the game.
Herb...
SoundHound wrote:
The point has been made but IMHO should be "amplified:" you can't see "it" (benefits of 14 vs 2 bits-color space, tonal range whatever) on your monitor! This is a little different from the extra resolution that is claimed for prints-not apparent on a monitor. So, yes, you may see something in a print after gross adjustment but how can you manipulate tonal range and color space when you can't see to adjust it?
but only recently. I have 40k plus NEF's on my computer. Most are 12 bit, but a good 10k of them are 14 bit.