12-Bit or 14 Bit Lossless Compressed-Your experience

Msmoto

In our current state of affairs with the D600 body becoming a frustrating issue and not being well handled in the opinion of some, I have seen a few folks looking at the D800. From my perspective, the only real drawback is in the size of the files.

14-bit lossless compressed on FX is 41.3 MB
12-bit lossless compressed on FX is 32.4 MB
JPEG fine (L) on FX is 16.3 MB

14-bit lossless compressed on DX is 18.6 MB
12-bit lossless compressed on DX is 14.9 MB
JPEG fine (L) on DX is 8 MB

My question is....what functional end product difference is there in a 12-bit vs. a 14-bit lossless compressed image? If there are examples, please post a link to something like full size images on Flickr or elsewhere rather than fill the pages with small images we cannot tell mush about. Let's get ready to Rumble....

PB_PM

14bit images have two advantages right from the start, 1) They contain more colour information, 2) Greater dynamic range (than 12bit).

Do you notice those things in most images? No. In high contrast situations, where highlight and shadow recovery are important, will you notice? Yup. Is the extra file size worth it? Depends on how you shoot, and under what conditions.

The different bit rates have no effect on resolution, so asking for photo samples, converted to JPEGs, is pointless. You'd notice the difference in prints, but to point out the difference in two jpegs is minimal at best, due to the loss of data in jpeg conversion.

Gab

I think it only matters if you do extreme shadow recovery. I don't think the 14bit advantage is noticeable in the highlights and only noticeable in the shadows if you boost them by more than 4EV. (at least on a d7k)

Golf007sd

I shoot 100% of the time in 14-Bit uncompressed. I much rather have the data than not. Specially when I shot HDR. I'm not worried about disk space, give the price of hard drives these days, storage really shouldn't be an issue.

PB_PM

Not much point in shooting uncompressed when lossless compressed is an option. By nature lossless means no loss. Of course uncompressed can be faster, because images wont be stuck in the buffer waiting for the files to be compressed before moving them to the memory card.

Ironheart

Also, post processing of uncompressed images is faster as the CPU doesn't have to uncompress the data first.

PB_PM

The speed of your HDD/SSD would have far more effect on load times than the decompressing of compressed RAW image. Since uncompressed files are larger, the benefit is basically non-existent.

Golf007sd

@PB_PM I will have to try your theory out, but I have a feeling when I'm in the processes of combining 10 images in to a single HDR, the uncompressed version will go by faster.

The amount of information being read by my SSD, a 40MB file vs one that is 20-25MB is so small it is not even worth trying to calculate the milliseconds I lose. However, the CPU cycle in decompressing and having it load in LR is noticeable.

JJ_SO

When I started scanning my b/w negatives, I used 8bit gray. I'm redoing them (frequently) with 16bit. Too often the scale was broken. Now, instead of looking for smaller files, I delete the less good - and gain a lot of disk-space, but have never the problem to think, more bits would have been better. I'm not shooting with high burst rate all times. I don't believe, more shots will catch more, I try to get better in getting the right moment instead of shooting it death.

As for the rest - speed at import, speed at post process: nobody is pressuring me. So, amateurs looking for a D800 instead of D600 are still not consuming half a dozen HDs a week. But shooting with this beast can also be fun in caring for single shots instead of several hundreds.

JJ_SO

Golf007sd: 10 images of the dynamic range of a D800 for 1 single HDR? You're making me curious. That would be about 40 f-stops or something like that? Are you stitching panos?

Golf007sd

@JJ_SO I use this amazing device called The Promote Control System. It allows me to take up to 45 images automatically, of a scene with "up to 9.0EV steps between bracketed exposures." Thus for me having the images be uncompressed comes in handy.

PB_PM

Goodness, how much dynamic range do you need! Then again I'm not an HDR junky.

heartyfisher

Also, post processing of uncompressed images is faster as the CPU doesn't have to uncompress the data first.

Actually reading compressed files are probably faster cos the limitation is the disk IO.

Golf007sd

@PB_PM Not sure if you have tried HDR photography, but I'm telling you once you do, it will prove quite rewarding and you will find yourself becoming a "junky."

The topic at hand, does provide good food-for-thought. I just want to store as much information on my memory card as possible when taking an image. Speed and storage are important but the raw data in far more important to me.

PB_PM

I do use HDR from time to time, but only find it to have limited value for my shooting style.

As noted, with lossless compressed, you don't loose anything. The 14bit uncompressed files give more benefit than shooting 12bit uncompressed for example.

JJ_SO

45 shots, 9EV between each shot means 405 f-stops. In which dark caves with interesting spots of bright sunlight are you shooting macro with DOF from 2" to infinite? Just asking. I'm always too tempted to fool around when I read something real far beyond my personal horizon. Especially if it comes to subjects with HDR and I hardly know any looking not artificial. But as everything, it's a question of personal taste

MikeGunter

Hi all,

If one thinks that FX is any better than DX (and most do think dynamic range and color response increases), the magnitude from 12bit to 14bit is far beyond DX to FX. ;-)

The proof is in post processing, in which many folks don't get their toes wet. One can use a single image well exposed image and get a multiple HDR image from a 14bit RAW image with layers in CS6 pulling ±5 stops in .01 steps, rather impressive, and that's from one exposure.

The files are large for a reason, and whether one will be willing to sign up for it is something else.

There are some 'film' affectionados who take their film to labs or go into the darkroom where they soup up their film and try to manipulate in the enlarger and dodge and burn the negative on the paper. Photoshop isn't much different - just cleaner and easier on the fingernails, maybe a tad harder to learn, but infinitely more rewarding.

My best,

Mike

DenverShooter

I run the 14 bit file size on my D800E. My theory is that hard disks are cheap and getting cheaper (and smaller) every year.

Once you shoot a picture in 12 bit you are locked into 12 bit forever...

Denver Shooter

safyre

I think it only matters if you do extreme shadow recovery. I don't think the 14bit advantage is noticeable in the highlights and only noticeable in the shadows if you boost them by more than 4EV. (at least on a d7k)

+1
here's a good article about it already: http://www.earthboundlight.com/phototips/nikon-d300-d3-14-bit-versus-12-bit.html

Bottom line, it won't make a difference for 99% of photographers, other than filling up your hard drive and slowing down your fps speed. For most people, just use 12 bit lossless compressed and stop worrying about it.

tc88

The proof is in post processing, in which many folks don't get their toes wet. One can use a single image well exposed image and get a multiple HDR image from a 14bit RAW image with layers in CS6 pulling ±5 stops in .01 steps, rather impressive, and that's from one exposure.

Mike, regarding creating HDR from combining multiple image files created from a single file, procedure wise, I know how to do it. However, I'm just wondering if there are additional benefits compared to just changing the curves directly from the original single file? Thanks.

Golf007sd

@JJ_SO: First things first, I have never taken 45 shots automatically, most of my HDR is created by 10 shots at a time give a specific perspective, with about +/- 1EV. Sometime I choose 1.7 or 2 +/- EV.

As for Mikes recommendation, each to his or her own. Many photographer also use PS to add filters as well in post, I on the other hand have filters that I use on the lens as I take the shot.

In the end if the photographer is happy, as well as those he or she shares it with...that is all that counts.

pippigurl

+1 to your last sentence Golf!

MikeGunter

Hi all,

@TC88, I'll post something soon to show you how to do that with RAW and layers in CS6 and if you are using Adobe Elements 10 you can do a variation of it, too, just not as elegant nor quite as extreme. Some people (I would be one) like to put multiple layers so as to add or subtract depth to the photograph, but sometimes just two layers with variance in grays will do.

@Gofl007sd, you're exactly right. To each his own - There's a lot of Crayolas in the carton for a reason. CS6 has an "Automate to HDR" under File that allows for a slew of files to automatically create an HDR file that can be wickered with into a PSD format with each of the files separately adjusted in post.

Seeing what you shoot before you push the shutter release helps. You can add the light where you need it so you don't have to add or subtract so much light. Ansel taught me that much. ;-)

My best,

Mike

spraynpray

If the driver for this thread is the size of files and the storage of them, then the problem goes away if you have a workflow that includes deleting all images that are not keepers. I import the lot and spend quite some time with them then when I reach the point where I feel I have got the best from the shots I have, I delete the rest. All this talk of sqillobytes of storage always makes me wonder how much cr@p is being kept. I would go so far as to say that not being ruthless at the delete stage would result in - say - a D7000 raw shooter keeping more Mb than a D800 raw shooter that deletes all non-keepers. Most people say that 5% is a typical keeper rate - do the math.

[Deleted User]

Can anybody explane to me the -12 and 14-bit- part - for HDR.

My HDR workflow at this moment in time, because software become better and better:

Camera set to 14-bit, the only reason, it gives me the most information.
PP in Lightroom.
Combine them in PS-CS5.5 in 32bit, (PS always gives me the best alignment).
Then I go directly to HDR Efex Pro, for me far the best HDR program.
Save it as a -16bit- TIF in Lightroom and get an enormous dynamic range to play with.

Works pretty well.

@ MikeGunter I'am curious to your - RAW and layers - post.

TaoTeJared

HDR should be it's own thread - especially if Mike is going to divulge some goodies

My question is....what functional end product difference is there in a 12-bit vs. a 14-bit lossless compressed image? If there are examples, please post a link to something like full size images on Flickr or elsewhere rather than fill the pages with small images we cannot tell mush about. Let's get ready to Rumble....

I'm coming back to MsMoto's original question.

With my D800 the only thing I can tell, is bringing back highlights and shadows (more highlights though) with the 14bit over the 12. Snow this winter is something I have noticed the 14bit helped. Not by much, but enough that I just leave in on 14bit and call it good and something I don't need to worry about. Honestly I'm not sure I could even post something as once the image compresses to jpeg, I don't think you could see the difference at all, but it is there.