mjcabrera wrote:
Come on guys, how many pictures have you taken since Nikon's announcement?
Go out and enjoy the equipment you already have. Have fun!!!
Well, since Friday just about 2200. And today, my 20D had a hell of a time AF'ing and my genuine defective canon grip locked it up on me.
Does the D300 really drop to 2.5 fps when using 14-bit NEFs?
While none of Nikon's current published documentation suggest this, it does appear that the Nikon subsidiaries have been verifying that: frame rate drops on the D300 if you shoot 14-bit NEFs.
oops 2.5 fps
This probability plus the D3 ISO 3200 images? I say let the bashing begin!! That's the natural order of things anyways. We pick out the deficiencies here. The Nikon guys praise the camera strength over on their forum.
obviously you have not spent time there and why would you anyway? at present i am waiting to get on any bandwagon till Real bonafide pictures are released. I do the same when canon releases a new cameraq, don't you?
as for the 14bit issue, i am waiting to see how that pans out as well , tho as we have seen in Pondrias? question a few weeks ago No One was able to show a real benefit to 14 bit tho the speculation was rampant here in This forum
jmcfadden wrote:
obviously you have not spent time there and why would you anyway? at present i am waiting to get on any bandwagon till Real bonafide pictures are released. I do the same when canon releases a new cameraq, don't you?
as for the 14bit issue, i am waiting to see how that pans out as well , tho as we have seen in Pondrias? question a few weeks ago No One was able to show a real benefit to 14 bit tho the speculation was rampant here in This forum
J
I haven't spent time where? Hogan's site? I stay current with what he writes. I didn't see Pondrias test. I have no idea how you could come up with a comprehensive test of possible 14 bit benefits. I do remember a lot of talk of dynamic range. There's seem to be constant confusion of how 14 bits relates to dynamic range. I'm more interested in a more detailed expression of meaningful analog values off the sensor. It seems that is what makes a thick file. Accurate tonal range that can be manipulated. The purported D3 files are very thin. Anyone who works with better files knows that at a glance. But I'm commenting on only those two files, not the D3.
My comment you quoted was simply referring to the over enthusiasm for the Nikons in the Canon forum. There's a good possibility that the high ISO Nikon files will be heavily processed, so great enthusiasm doesn't make a lot of sense at this time.
I believe the more troublesome camera for Canon will be the D300. I can say that now because it's an improved D200, which causes Canon problems.
The whole 12-bit v 14-bit can be summed up quite easy: 4 times more stored intensity data. That leaves more headroom before you run into banding issues.
What irks me about this whole issue is that it doesn't make sense. Going from 12 bit to 14 bit RAW means just that. 14/12 = 1,16666(repeating). That's just 17% more data. How the [expletive] can that bring a framerate of 6 or 8 down into 2,5 ?
The only thing I can think of is that the Sony sensor can't do 14-bit ADC with the AD converters that are added to each line of pixels. So the high fps is done with the 12-bit ADCs while the high accuracy 14-bit ADC needs to take place off-sensor. But that's just me guessing.
Tentacle wrote:
The whole 12-bit v 14-bit can be summed up quite easy: 4 times more stored intensity data. That leaves more headroom before you run into banding issues.
Assuming the analog data is accurate enough to be converted into that many significant digits. If a camera is pulling dirty readings off the sensor, it may not even really be "using" 12 bits (although it may be a 12 bit A/D). That's why doing post capture noise reduction is so inferior to having technology that allows a CMOS sensor to accurately capture data.
Assuming the analog data is accurate enough to be converted into that many significant digits. If a camera is pulling dirty readings off the sensor, it may not even really be "using" 12 bits (although it may be a 12 bit A/D). That's why doing post capture noise reduction is so inferior to having technology that allows a CMOS sensor to accurately capture data.
Well, the usability of 14 bits depends on the noise present, that's true. But if I shoot a very low contrast image at base ISO, I can boost the contrast a lot more from 14-bit raw than I ever could from 12-bit. 16384 or 4096 discrete intensity levels to work with can be the difference between having visible banding or not.
Why not just take a 14bit raw file and set the two least significant bits to zero and compare it to the full 14bit version?
I think there are a few people on this board with enough knowledge about raw file formats to hack a file like this.
35+ pages and not much in the way of tangible evidence in the form of a Picture that could prove any Real benefit. Real here being the operative word.
J
Interesting way to interpret that thread. I remember the discussion as trying to learn more about the data off the sensor. I'm sure there's tangible evidence of 14 bit data being more expressive than 12 bit. What I would like to see is a meaningful difference in print. I don't think thee was a single image in that thread.
It's easy to demonstrate that dark areas do not "block up" as easily on the 1DIII as the 5D. That was the first thing I looked at when I got the new camera (after playing with the 10 fps, of course) But there are many changes between the camera in addition to 14 bits.
I like MichaD suggestion. It removes all variable except the expressiveness of the data.
Jeff wrote:
Unfortunately there's no 'perfect' way to demonstrate it, without truly identical files from two different cameras (since no camera I know of produces both a 12-bit and 14-bit file). Slapping a given lens on two different bodies won't do it, because differences in light falloff at the corners can easily influence the perceived posterization. There's just no perfect way to demonstrate it at this point in time (that I've discovered).
Doesn't the D300 have options for both 12bit and 14bit RAW?
I shoot for fun, not work- and lately i've been having more fun w/ a simple Nikon FE2 setup than my 1DII and L's... now that the market is shaking up, i've decided to ditch all the EOS gear (sold sold and sold), pay off some debt (good to go!), and buy a few more AIS lenses.
I'll wait for the 5DII, and unless it is really fantastic, i'm going to go D3- unless the D3 doesn't live up to the hype. If that's the case- then damn i don't know what i'll do. Keep shooting film i guess?
Canon EOS 5D and 50/1.2L : ~$3200 used
Canon EOS 1V and 50/1.2L: ~$1800 used
Canon F1N AE and 50/1.2L: ~$600 used
Nikon FE2 and 50/1.2 AIS: ~$370 used
Thinking about it again you would probably need to use some simple dithering (i.e. without error propagation) by adding maybe 1.5 bits of random noise to the 12bit data to make it a fair comparison. Otherwise you'd get pretty nasty posterisation in the lower bits.
Every modern A/D converter would do this.
this really isn't a great way to express what we are talking about here . It is a good example of what Might be true WRT our discussion but too many variables are at play here to make a statement
Jeff wrote:
It's better than anything else anyone's come up with so far. If you have any suggestions, I'm all ears. Have you worked with any 14-bit files, John?
Seems to me everybody (over your way) is getting caught up in DR discussions, which really don't directly apply to the 12- vs. 14-bit issue.
Regards,
Jeff
But is it even valid? For example, what is 12 bit Raw to 16 Bit down sampled to 8bit compare to 14 bit Raw to 16 bit down sampled to 8 bit.
The comparison as described is 8bit JPEG out of Camera compared to a post processed 14 bit raw converted ot 16 bit (tiff or psd I presume) downsampled to become an 8 bit JPEG.
:
