14
201112bit vs 14bit RAW and compressed vs uncompressed… Does it matter?
You know that to get the most of your DSLR you should be shooting in RAW, right? But these days Nikon cameras gives you even more options: 12-bit or 14-bit, and compressed or uncompressed RAW (NEF) files. Which should you choose?
Short question: Does it matter? Will you see any difference between compressed (lossy) and uncompressed (lossless) RAW? And between 12 and 14 bits?
Short answer: No it does not matter. Choose 12-bit compressed (because they take up less space) and forget about this topic. Or choose 14-bit uncompressed because theoretically you’re getting the “most” from your camera – you just have to live with the file sizes.
| Approximate RAW file size on a Nikon D7000 |
12 bit | 14 bit |
| compressed | 12.6 MB | 15.7 MB |
| uncompressed | 14.9 MB | 18.8 MB |
Not happy with the short answer? Then read on…
Basic computer science tells you that 14 bits store more data than 12 bits. To be exact: you can store 4 times as many shades of intensity in a given range, or if using the same step size you can cover a range of values 4 times as large.
Basic computer science also tells you that lossy encoding throws data away. So then it seems logical that images obtained from 14-bit lossless RAW files should have a larger dynamic range and be more detailed and nuanced than images from 12-bit lossy RAW files. The big question is whether these theoretical advantages are ever visible in real life.
This topic can get extremely complex. The best and most rigorous explanation I found online is in this article from by Emil Martinec: Noise, Dynamic Range and Bit Depth in Digital SLRs.
Conclusion: due to sensor noise you cannot see the difference between 12 and 14 bits, and neither will you see the difference between lossy and lossless RAW encoding.
This is also touched on by dpreview where they wrote “…it is easy to understand that [higher bit depth is advantageous] only IF the sensor itself has sufficient dynamic range.”
People have posted some experiments concerning this (for example see the D300 12-bit vs 14-bit comparison) but shooting a test chart in the dark is not totally convincing.
Curious as I am, I decided to see if I can experimentally find any difference in the recovery of over- or under-exposed “real-world” photographs. If maybe there is some advantage to be had from those extra bits in these extreme tests of dynamic range. After all – none of us know exactly how Nikon’s engineers implemented this. Since the Nikon D7000 has a good dymanic range it is as likely as any APS-C DSLR to show the advantages (if any) of higher bit depth.
For this I chose the following test scene:
Correct exposure (1/250 at F8.0, ISO 100). Click for more detail.
I then overexposed the scene by a massive 4 stops:
Overexposed by 4 stops (1/125 at F2.8, ISO 100). Click for more detail.
And then I underexposed the scene by an even larger amount, 6 stops. I went 2 stops further in the underexposure since digital cameras are better at retaining shadows than highlights:
Underexposed by 6 stops (1/8000 at F11, ISO 100). Click for more detail.
You ‘re welcome to download and experiment with the original .NEF raw files yourselves. (link to zip file, 101MB)
The next job was correcting for the over-and-underexposure in Adobe Lightroom 3.4, and comparing results between 12-bit, 14-bit, lossy and lossless encoding.
Crop from reference image (correctly exposed)
First, the overexposed examples. You’ll especially notice big white washed-out regions caused by channel clipping. It is interesting to see whether extra bit depth helps to lessen this effect. Also look for differences in detail caused by RAW file compression.
12-bit compressed (lossy), overexposed + corrected
12-bit uncompressed (lossless), overexposed + corrected
14-bit compressed (lossy), overexposed + corrected
14-bit uncompressed (lossless), overexposed + corrected
And the same sequence for the underexposed region. Instead of washed-out regions caused by colour channel clipping we now see a lot of noise, since correcting the underexposure requires massive amplification of a region with very low signal.
12-bit compressed (lossy), underexposed + corrected
12-bit uncompressed (lossless), underexposed + corrected
14-bit compressed (lossy), underexposed + corrected
14-bit uncompressed (lossless), underexposed + corrected
My conclusion is that even in this extreme example there is very little difference between either 12-bit and 14-bit or between uncompressed or compressed results.
If one had to seek differences I would say that the colour reproduction is less degraded in the uncompressed and for the 14-bit RAW images. It seems like the 2-bit difference between 12 and 14 bit is mostly applied to shadow detail and especially colour information in the shadow regions. But given how extreme my test example was and how subtle the effect, I would also call it a negligible difference.
Furthermore there seems to be effectively no difference in detail captured in any of these modes.
So I again conclude that THERE IS LITTLE OR NO PRACTICAL ADVANTAGE TO USING EITHER 14-BIT OR LOSSLESS RAW COMPRESSION.
, overexpose
Josh
Bad science. You should use different lighting conditions. There is a difference between underexposed sunny scene and shooting in pitch dark interiors or at night.
I did the tests myself and 14 bit has huge advantage under extreme dark conditions.
Compression concerns the highlights only and I assume it works the same way for the high-key scenes – white on white.
It’s true, in every-day outdoors shooting, 14bit/uncompressed doesn’t really matter.
fmalan
Hey Josh – thanks for the comments! Calling it bad science sounds a bit unfair. I will be happy to retest for extreme dark conditions, but I don’t agree with your statement… if the sensor is photon-starved it is photon starved. Doesn’t matter if you do this with dim lighting or with severe underexposure in daylight. Right? Only difference I can think of would be the relative larger contribution of integration noise (or hot pixels) if you use a long exposure in dim light. Or the fact that e.g. moonlight covers a smaller part of the visible spectrum. But the number of photons impacting the sensor would be similar in both cases (underexposed daylight or longer exposure in a low light scene) so indentical principles will apply.
Florin Coter
Hi fmalan,
Josh is physically right, not educationally… In full sun the sensor is biased with a different voltage as compared to darkness. So the two situations are different. Moreover, in strong sunlight there is some photons spill from pixel site to neighbors. So one expects different color errors at full sunlight and in darkness. Finally, one does not willingly underexposes 6 stops… Although your test was technically correct, its conclusions are not good enough for real life situations of good photographic practice. One of the hardest lessons I had to learn as a Physicist was that is does not matter how easy a measurement is done, the interpretation of the numbers is an art in itself. Referring them to real life relevant situations is an even harder task. But, the start is promising, so keep testing!
Florin.
fmalan
Thanks for the correction and clarification, Florin! I also agree that underexposing by 6 stops is quite unrealistic in normal operation. The problem I foresaw was that in normal operation there would be no visible differences at all! (Which would support the same practical conclusion about which settings to dial into your camera.)
Jaimiee
In the list of specialties, I would Probably like to start at at a Family Practitioner. I beevlie this will give me a better opportunity to practice my skills , and once in a while my patience and expedience. After which I would more than likely feel confident enough to tackle any other profession. I would like to build on a solid foundation of skill and knowledge before taking on challenges I’m not well equipt for. The plain fact is , after my confidence is up, I relish i the opportunity to prove my worth.
FrisianPhotography
Thanks for the write-up and especially for the little test. Many theorize about it, but few do actual field tests.
Looking at the newly announced Nikon D800, I’m curious whether these results will be comparable with a 36MP sensor or whether there would be more of an advantage.
fmalan
I am going to blog about this… In short: the D800 has identical pixel density as the D7000 so should perform similarly on a per-pixel basis. However, if you rescale the 36MP image to 16MP it will look better than the D7000 @ 16MP, so actually the D800 will outperform the D7000 in terms of noise. Dynamic range should be about the same though.
Dilip
Good you posted it now :DI was going to write a tuaoirtl (with video) how to create video sitemaps (for external sites like vimeo and blip.tv too).Now I finally can combine both types of sitemaps in one file!Do you guys know any good Plug In which can create video sitemaps or MRSS? Didn’t find any…_wapr | screencastblog.net
fmalan
Just a short follow-up to say that the D800 has a pretty amazing sensor that seems to outperform the sensor of the D7000 on a per-pixel basis too! Still, if I had one I would be using 12-bit compressed RAW just as with my D7000.
Glenn
I’m looking all over the internet for information on this topic in regards to my newly received D800. I know memory is cheap and all, but this is more about how many cards I have to carry to a shoot. My Nikon D3s gets about 359 images on an 8GB card shooting 14 bit lossless compression. The D800 gets only 200 with a Lexar 400x 16GB card. I understand why and the resulting image is amazing, no small thing given my D3s is no slouch. However, short of going to 32GB cards, am I really losing anything going to 12 bit, and/or compressed both of which alone get me more images on the card and together get me even more? If I were going to chose one (and want to maximize shadow detail) should I reduce the bits or change from lossless to compressed?
fmalan
Dear Glenn
If you had to choose one of the two, I’d suggest switching to compressed (lossy) and keeping 14-bit. In my tests with the D7000 the bit depth was the only setting that seemed to have a marginally visible effect – in the colour accuracy of extreme shadow regions. I think it’s reasonable to assume that the D7000 and D800 will behave similarly in this regard.
If you want to be perfectly sure the best would be for you to test this on your own D800. Use a tripod and take 4 consecutive shots with 14/12bit and lossy/lossless. Please let me know if you find any surprises!
Jerry Suppan
Greetings. I read with interest when you commented the D800 has a pretty amazing sensor. I take that as a compliment to Sony as Nikon OEM’s their D800 sensors from Sony. The sensitivity of the Sony sensor’s (and perhaps FujiFilm) are some of the best in the industry. Sony not only OEMs their sensors to Nikon, but a number of other camera manufacturers. Since I heard Nikon has exclusivity (6 months?) on the Sony sensor technology for a period of time, I am looking forward to the upcoming Sony A99 (and beyond) full frame sensor camera from Sony, the Alpha A99. It should be announced before or at Photokina next month in Cologne, Germany. It will be a 24MP sensor with a 36MP to follow in 2013. Again, I think that may be related to exclusivity which Canon has on the D800 for a period of time. Which, is a good thing. Sony has an opportunity and advantage to learn how well Nikon does with the 36MP in the market, and can apply it findings to improvement on their own camera. BTW, it is also rumored to be 14-bit RAW and uncompressed too which brought me to this blog in the first place from SAR (Sony Alpha Rumors) website.
I currently own Sony Alpha A65, NEX-7 (both APS-C) and the Sony HX30V travel zoom category of P&S cameras. That camera currently occupies the #1 slot at: http://kakaku.com/search_results/hx30v+sony/?search.x=0&search.y=0
(if you can read Japanese) that does comparison of prices and popularity of nationwide throughout Japan. When first released in March, it was ¥39,800. Now, I got it for ¥21,500. :-)
Thanks for this analysis though because it provides information and a basis to make a decision in terms of design and/or performance features in making my camera upgrade decisions.
RPV
This is very interesting. I’ve been shooting my D7000 at 14-bit lossless, but I may change to 12-bit lossy in order to benefit from the smaller file sizes and the increased buffer depth.
Richie
Ed
Three options on my D700
1. uncompressed; 2. lossless compressed; 3. compressed. No “lossy compressed” option is shown.
But I assume the article above, by “lossy compressed” means #3, am I correct?
fmalan
Hi Ed
YEs, “lossy compressed” means #3.
On the D7000 (on which I tested this) I can choose only lossy/lossless and 12bit/14bit. I thought that Nikon .NEF files were always compressed – but apparently the D700 also offers a uncompressed setting. There will be no image quality difference between lossless compression and uncompressed.
Glenn
Just a further follow up on my earlier reply to different post. With lossless compressed or uncompressed I get 259 images on my 16 GB Lexar 400x card at 12 bits. With Compressed I get 354. I don’t want to make it all about card capacity, I also don’t want to add another 2TB drives to my Drobo anytime soon. I’m just trying to understand what I’m tradiing off – I appreciate how you are addressing this issue. I don’t need lectures on memory is cheap, 14 is more than 12, blah blah. I just want to maximize images, save a little space, but give up little or nothing in image quality – the D800 is great, given everything else about its sensor, do I really need two more bits…?
fmalan
Hi! As reply to your second comment: you will theoretically be giving “something” up using 12 instead of 14 bits, but this “something” will most probably not be noticeable. I had to go to quite extreme lengths (massive under-exposure followed by post-processing compensation) and pixel-peeping to notice any difference on my D7000 between 12 and 14 bits. And even then the difference was very subtle. The D7000 uses more or less the same technology as the D800 so I expect the same for the D800.
Whichever settings you choose you will keep the advantages that the D800 has over e.g. the D700 or even the D3S e.g. resolution, video, and better auto-iso implementation. If I were you I’d choose 12 bit compressed if that gave me a real advantage in workflow usability (more shots in the field) or would save me money (postpone upgrading my Drobo). But take a few comparison shots and see if you can see the difference. If you can’t then the answer is clear, and this will convince you better than my word :-)
Mike de Klerk
No matter how many bits you have. If you over expose the sensor, it all comes down to the same story. But in good conditions (Meaning you are using the right range of sensor) you will never over expose. Meaning that you will never see the difference between 12/14 bit on your computer screen as that is only 24 bits per pixel and not 48 bits per pixel. But there is more information in 14 bit. If you need that information for some reason, you probably already know you do.
Storage is no problem these days! Shoot your images in 14 bit. Maybe for future 3D reconstruction it might be helpful ;)
RPV
I switched from 14-bit lossless to 12-bit compressed on my D7000 after reading this article, and thus far I’ve seen no real world difference when editing my NEFs. The smaller file size and improved buffer depth is a definite plus. I wouldn’t have considered this change before I read this article. Nice work.
Richie
fmalan
Thanks, Ritchie! Good to hear that it helped you decide too!
Jake
Are there any differences to things write speeds and buffering – obviously for the camera to compress a file it has to run some kind of process to compress the file which could effect how long it takes for the camera to process things, then again writing a smaller file to a card should take less time.. Wondering if there are advantages or disadvantages to any of these settings from any standpoint other than just the visual comparisons.
fmalan
Hi Jake.
Good question! I’d have to test this. AFAIK all the major image processing in a digital camera (raw conversion, jpeg compression, etc) is done in hardware by the specialized image processing chips in the camera. This explains why it is almost instantaneous in-camera compared to how long it takes on a computer, even though the computer has much more general-purpose processing power. Therefore I expect the time required for compression to be negliible compared to write times.
Usually the continuous shooting party-pooper is writing to the card once the buffer gets full. At this point the shooting rate drops dramatically because It is limited by the file writing speed, and smaller files write faster.
The number of frames you can store in the buffer depends on whether the compression and bit-depth reduction happens before it’s stored in the buffer (as I assume the case it) or if it’s done just before being written to card.
I expect that smaller raw files will make the buffer last longer before it is full, and once full will allow for a faster reduced shooting rate due to faster write-times.
RPV
I’ve seen an improvement in the buffer depth of my D7000 since changing to 12-bit compressed.
Richie
fmalan
Great to hear – that is indeed what I expected.
Afiavi
It would be cool however, if Google can prdvoie a tool that automatically does this – specially on the blogger platform.As webmasters, there are more and more things that we have to do now aside from creating the site itself and the content. It seems as we progress in technology, that we are getting more and more burdened with additional tasks.
Ron King
Hi,
I have a D300s and tried both 12 and 14 bit. The only real thing that showed up was that in continous mode and at high speed shutter rates the 12 bit kept to the high rate while 14 bit slowed it down by a considerable amount. This was regardless of it being lossy or uncompressed.
fmalan
Well there you have it – 12bit is the way to go. Especially on the older cameras like the D300s that probably has a slower image processor and data rates. But even on new cameras 14-bit will eat your buffer faster than 12-bits. And except for the D3/D4 series those buffers fill up very quickly.
RPV
The D300s drops to only 2.5fps when shooting 14-bit NEFs. The D3x drops to only 1.8fps I believe.
V L
Hi
You’ve proven that a really really bad file looks the same whether 12 or 14 bit, lossy or not. A blown highlight is white no matter how many bits it has to be sure.
What may be an interesting next step, is to do the same with just 1-2 stops over/under. This is more realistic to have in real life, and takes into account the 1-2 stop overhead of using RAW->JPG. I’m keen to see whether you’ll get the same results!
fmalan
Thanks for the suggestion. I pushed it this far because this is exactly where one would see the biggest difference; on the boundary between detail and no detail. The overexposed scene contains some details (it is not completely white everywhere) so one should have seen advantages in the transition zone. Still this argument is probably less convincing than a test.
(SR5) A first for Sony: 14bit RAW files on the A99! | sonyalpharumors
[…] More 12 vs 14 bit tests: here and here. […]
Archivos RAW de 14 bits en la A99!
[…] de los tamaños de archivo ligeramente más grandes. Más pruebas de 12 vs 14 bits: aquí y aquí .El motor JPEG se ha mejorado también. Pero habrá que esperar a la prueba real para ver los […]
Dr. Nicolas Rao
I own a d7000 and recently was asked to help out the main photographer at a humoungous Indian wedding. The change to 12 bit compressed was a real boon. We were able to even screen our shots while wedding was in proceess on huge screens provided. My frriend used Canon 5D MKll and with lightroom 4.1. I was happy to see little or no diffrerence between the two. These were projections on giant screens and not a calibrated moniter, yet I was happy for the huge number of shots avaiable in my memory bank of cards.
If pressed to work on a large print, say 40 X 30″ I may try the 14bit lossless,but then the number of photos would not be large and of little consequece.
Loved the debate so far.
Let us see what the foture has to offer! Thank you.
Dr. Nicolas Rao
Pardon the spelling and typos. It wasn’t easy on my touchscreen tablet. I am 61 almost 62 years old.
Kawalta
each and every sitemap has its own XSD, but all, incuidlng the Geo sitemaps uses the standard XML ‘layout’. You supply the schema for the Geo reference the same way you’d supply a News schema@Arthur, a standard web sitemap (sitemaps protocol) is the base of this new feature, so yes, you can continue using web sitemaps
igor
I know it’s fun to theorize about 12 vs 14 bit, but you have to remember that it is 12 vs 14 bit per RGB channel. Most monitors are 10 bit, some are 12. This is not the same as the color depth you set in windows. The 16 vs 32 bit you set in windows is the sum of RGB bits. What this all boils down to is that 12 vs 14 bit raw files will look almost the same on most monitors since monitors effectively down-sample them to 10 bits per channel.
fmalan
Hey Igor!
There are two aspects to the bit depth – the one that you mentioned (in you comment above) is the D/A conversion – displaying the colours once they are recorded. Here you are completely right. 14-bit doesn’t really improve over and above 12 bit.
The more important aspect in my post is the A/D conversion that happens at capture time — whether the original scene is captured accurately. Especially in highlight or shadow regions you want to capture the limits of the camera’s sensitivity range in as many steps as possible so that correcting the exposure (i.e. stretching the information-containing extreme edge of the captured range to fill the whole visible range) doesn’t show posterization artefacts. However we see that the sensor’s noise generally prevents that from becoming obvious even in those extreme cases.
Irene Kotov
Hey Igor and Francois,
A question for you – how does all you’ve said so far relate to being able to export the same 12 and 14 bit files from Lightroom or Capture One as 16-bit TIFFS?
Does that mean that 16 bit TIFF is a bit of a lie, then? Where do those extra 2 or 4 bits suddenly come from?
Irene
fmalan
Hey Irene. It’s not really a lie. In audio processing, for example, pros also always use higher bit rates for intermediate processing, otherwise you may lose a little bit of quality at every step that ends up summing up to a large total loss. This is especially severe if you compress or stretch tonal ranges – exactly what happens in the tone curve conversion from RAW to TIFF or JPEG. However for final output you can lower the bit depth to sane levels again.
Also remember that the usual alternative to 16 bit image processing is 8 bit, and you will be guaranteed to lose data when downcasting to 8 bit. There is no TIFF format for 12 / 14 bit so the next one up is 16, so that is a better choice for intermediate processing steps.
Dr. Nicolas Vincent Rao
My first question – what is the minimum bit level difference perceptible to man with perfect eyesight?
In what medium – CRT or LCD?LED monitor?
What is the ultimate impact on the photo – infinitesimal? Few of us even know how to get exhibition quality prints these days.
In the old days of film and paper, I knew the art of fi9ne art prints.
What is the best presentation today – a brief show on most inefficient digital projectors or a print you can leave in your will?
Dr. Nicolas Vincent Rao
This is something deep in my heart. I spent a small fortune, almost enough to buy a home to equip my darkroom, a large room, with all that was required. I tried to buy everything Ansel used and did use a Reseller 23 C II with great lenses to do my printing.
I had no less than 6 timers, 3 for exposure and 3 for developing B/W in red light and most of the time no light.
My developers were measured in science labs before I mixed them.
Was I nuts or are you playing around with hair splitting?
fmalan
Wow, respect for your dedication! If you weren’t earning money with that equipment there is no objective way to measure if it was “worth” it. If you could afford it then you should answer that question yourself, as the worth is something personal. Technology is beautiful in itself, but in the end the beauty of you photography is about the effect it has on the viewer. And for that you don’t need expensive equipment per se. It just needs to speak to the viewer.
per
Hello, nice investigation! I am thinking of buying a D800 so file size is a concern and your investigation is good news.
Here is a question, though. You have over/underexposed photos whose basic dynamic range is not huge. But what if you take a photo where the dynamic range is huge – say that in the same picture you have a bright sun as well as a coal cellar. Would 14 bits help alleviate this situation or does your results still apply?
fmalan
Dear Per
If you shot a scene with such a huge dynamic range it will result in locally overexposed and underexposed areas. If you want to recover those areas it will be the same process as in the examples I showed.
Henrik
Very nice article – thanks for helping me save a good 30% of storage :)
ouki
Not noticeable by human eyes does not mean the information is not there. As the technology advances over years, the photos you shot today using 14-bit uncompressed may become a bit more beneficial when future technology can extract and convert that bit of invisible into visible. Same concept as HDR, photograph can see what man cannot see. Thus, As long as storage is cheap, I stick with 14-bit uncompressed.
fmalan
Hi Ouki
The author of the technical piece I linked to explains that the information is basically not there. Or rather: the signal is obscured by the noise floor.
So it is not really a question of technology, but rather the way it is encoded in the first place. But I won’t advise anyone to stop using 14 bits if the theoretical knowledge of saving the most data is important to you. Just point out that in practical terms you’re not gaining any noticeable image quality. And I don’t see this changing anytime soon.
Regards, F
CS
Hey FM,
thanks for the writeup. I was doing some research on the topic as file size can be a concern with a D800. In my humble opinion, the whole “storage is cheap / filesize doesnt matter” argument comes from people who don’t take any photos, simple as that.
Anyway. After wading through hundreds of stupid posts I eventually found your post and the from some other dude featuring actual evidence. I also did a test series for myself to be sure, using both high and low ISOs and different combinations of 14/12 bit and lossy / uncomprossed. Bottom line: If I were to mix the shots up I couldnt tell the difference and I’d dare anyone for money that they could.
So finally one easy answer at photography, yay. 12bit gives me 700 shots on a 32gb card (as opposed to 400 shots on14bit / losless) hence I could safely shoot any event with the backup-function without having to switch cards! Theres gotta be hook, right?
And of course there is. On my i7QM2630 cpu, hybrid state drive and 8GB RAM (it’s still reasonably fast machine, probably way above the average home computer), uncompressed RAW files, albeit beeing much larger, load significantly faster. While the zoom from full image to 100% takes a good 5 seconds for a compressed file, the uncompressed files load instantly. I have been shooting lossless compressed all the time (and dealing with the same delays while editing) and it was quite shocking how fast those uncrompssed files can be.
God damnit, why does everything has to be such a stupid tradeoff? I whish the option wasn’t there so that I could stop worrying about it and just take photos.
fmalan
Wow CS, thanks for the positive and insightful comment! I agree with you about the “storage is cheap” comment btw!
Also thanks for the info about the loading speed. I haven’t yet tested this, and it also surprises me somewhat. Makes sense, but still surprises me. What software do you use? I’d like to test this on my machine too.
Of course you could still shoot compressed and then convert to uncompressed DNG on import – if you use Lightroom, right? Then you at least save space on your SD/CF card and in your camera buffer, where space is most limited.
CS
Hey again,
I forgot to mention that I’m importing my photos into LR (4.2).
It would be great if you could run some tests on it, maybe I’m just seeing ghosts. It is a problem that it’s hard to run such a speed comparison test in a scientific manner, too many variables (like memory allocation, how’s the CPU beeing throttled, whats running in the background etc.) cannot be controlled.
Your suggestion sounds great, but I have no idea where I could do that. I couldnt find such an option on the import screen.
fmalan
Hi
Well to be honest I don’t use DNG since some camera-specific metadata (like focus points) are lost on conversion.
However all the important aspects of NEF are preserved, and one could argue that DNG is a better file format for portability and future-proofing your photos.
(I don’t feel too strongly on this matter – something I intend to write about at some point)
Here is a description on how to batch-convert your imported NEFs to DNG.
http://livedocs.adobe.com/en_US/Lightroom/1.0/help.html?content=WS0700C48D-7ACD-4639-9825-CCB6A22C0056.html
Note that you can choose to use lossless compression – the alternative is a new “lossy DNG” format which offers high compression with some loss of spatial resilution.
Leaving this unchecked implies uncompressed DNG.
I’ve run some speed tests and don’t notice any significant difference between loading times of compressed 12-bit NEF, lossless DNG, or lossy DNG in Lightroom 4.2.
Regards
F
Emon
I have noticed that with ADL on (my d90) that it aucatlly exposes the photo less than if the ADL was off. In most cases if the photo were turned off in post processing the photo would end up looking WAY under exposed as opposed to it if the ADL was never on which would simply leave the shadows somewhat underexposed. For example your boy’s face would not be quite as dark were the ADL left off, but the sky would likely had also been slightly blown out because the photo would have simply metered in a way that left the photo exposed longer. Just FYI, as I am aucatlly a fan of ADL, but your article and photo examples seems to portray a more exaggerated benefit than ADL aucatlly deserves.
Anne Wallace
Thanks for taking the time to thoughtfully reply and engage in useful dialogue. I am going to now take your recommendations to set my D-800.
Why do you suppose Nikon would offer a relatively meaningless option? Surely they run their own tests and would be aware that there is negligible difference visually?
Cheers,
Anne
Dr. Nicolas V. Rao
This is very comforting. Thanks Anne.
I have another question. Not sure this is the best place to ask.
But here goes anyway,
I use a D7000 and Macro Nikkor 105 f 2.8 for a lot of macro work with insects.
Will the compressed file lose anything in this area of work. I do crop my files often after conversion. I have to do it to remove unnecessary objects.
Dr. Nicolas V. Rao
Just wanted to say, I take back my earlier comments on equipment etc. You have been right about everything said. I also must stop making comments after a couple of beers. I can say that now after quitting the habit for two months at least.
So my apologies to the offensive comments made earlier and my thanks for improving my life with my much loved D7000. I think your earlier answers cover my question about the macro format too! If there is a reason to use the 14 bit, I am not going to be able to see it and hence not bee able to edit it either, so I think nuff said.
Thanks again for the excellent article.
Nick
fmalan
Hey, thanks, Nicolas. Appreciated! :)
yoda
I always used lossless 12 bit setting and now more than ever. The big thing with me is file sizes- its just ridiculous. On my D700 its just under 12mg and now with the D7000 its a whopping 15mg and this is with the setting I mention. I have read in the past of this very same topic. In my style of shooting I see no real advantage. I can only guess what the newer cameras will bring.
Yung
Hi everyone,Since over a year has pasesd since we published this post, we’re closing the comments to help us focus on the work ahead. If you still have a question or comment you’d like to discuss, free to visit and/or post your topic in our .Thanks and take care,The Webmaster Central Team
Sandu B
Hello. Just some thoughts:
– the files (raw) should be available for download. I would really like to make some comparison in other software (bibble, silkypix, capture one). No matter their sizes.
– The same experiment I would perform on PS CS6. It might make some difference
Size (not always) matters. My beloved Fuji S5 makes the RAW at 25Mbytes…
fmalan
Sandu – thanks for the tip. I’ve added a link to the raw files in the post’s body. All NEF files compacted in a single ZIP file of 101MB.
12Bit Vs 14Bit Raw And Compressed Vs Uncompressed… Does It Matter?
[…] Francois Malan is a freelance photographer formed in The Netherlands. You can follow his blog here and his 500px here. This post was creatively published here. […]
Noticias breves de la semana (18) - ALTFoto
[…] 12bits o 14 bits, ¿hay alguna diferencia? […]
Noticias breves de la semana (18) | 18minutos
[…] 12bits o 14 bits, ¿hay alguna diferencia? […]
Stepan
Thanks for interesting test!
But for testing compression probably scene with more detail and high frequency details (and very sharp lens) would be better. And also with some sharpening the difference may start to show.
Does anyone know more about Nikon NEF compression? If it is in spatial domain or colour domain, … ?
Axn2@
Interesting write up. This article helps me skip some of the early step i need to do myself when i asked myself the same question. I am however still have a list on my own, which didn’t answered or maybe mentioned in your test which are:
1. how it is in relation with the color space you used (you did not mentioned whether you are using sRGB or AdobeRGB?). Can’t dig that out of your original NEF files. Those color clipping, whether this color space aspect took part on this?
2. what RAW converter software you are using to come to this conclusion? I suspect you are using Adobe product rather than Nikon proprietary RAW converter for NEF (CaptureNX2). Please NOTE that camera manufacturer never disclosed their secret code recipe for their proprietary RAW files, and all of the RAW converter software maker only use their own algorithm to convert them to match its perceived real color rendition made by its camera maker as best as possible. Nevertheless, the raw converter from respective camera manufacturer is still the best to do it. I just throw the possibility to this discussion that we might only could harvest the benefit of the 14bit advantage in practical photography when it is converted using their own manufacturer’s RAW converter.
Another thing i have in mind to add the complexity, does anyone noticed that the 14bit NEF format has a more reddish & magenta in its color rendition? i noticed clearly that in my D700 LCD of my photo taken on 14bit format and so in your magnified crop. This is lead to another question of mine whether Nikon’s image processor has more sensitivity on this color channel? (i suspect it is derived from Nikon’s characteristic of their image processor chip since their 1st digital, rather than from its digital image sensor medium. My experience from my old D70s and other old referral on early Nikon digital e.g. D1, D100)
I guess i should start doing a test shot myself then, and do the comparison using CaptureNX2 :-). In the mean time, it would be interesting to see how people find that in this discussion.
Cheers!
fmalan
Het Axn2
I used SRGB, and Nikon’s ViewNX2.
Actually I believe that Nikon and Canon do share (or sell / license?) some codec info to Adobe; at least on how to decompress their raw files correctly. But I may be wrong on this one.
Would like to know your findings.
Regards,
Francois
Terry
I have just found this thread. Very interesting. Honestly there is a place where 12 vs 14 etc. does make a difference. In the processing of HDR especially with large array Panoramas where you “could” print a 4 meter by 6 meter print at 300 dpi and a file size in many gb. The processing the files, the software “sees” a much great bit depth than the monitor, printer of eye can interpret. Then Tone Compress “flattens” the result to a usable depth. Lightroom now has the ability to work with greater depth with the same result. Files are always “flattened” to match the printer or monitor. Monitors or printers which can show greater bit depths are available but very expensive and of limited use to most photographers. So the answer may be in the end it makes no difference but it might (under very special circumstances) in the middle.
fmalan
Hey Terry
Thanks for your comment. I agree that is is good to use high bit depths (higher than 3x8bpp in any case) for post processing, but this is a different issue. Nikon NEF is not an intermediate format: it is decoded once, and then the rest happens.
My experiment seems to show that there is essentially no additional image information in a 14bpp NEF compared to a 12bpp NEF, theoretically explained by the observation that any additional signal is drowned by sensor noise.
The bit depth of your display device doesn’t matter if the image information wasn’t there to start with.
I used 3×8-bit (=24 bpp) sRGB images to show the differences between modes, but I use these 3×8 bits to show everything that is contained in a small extreme subset of the original spectrum, therefore it shows all there is in this band. Decoding at e.g. 3×16 bpp should show exactly the same info.
To conclude, whether you start with 12bit or 14bit NEF, you should choose not to save them at unnecessary low bitrates at any point in the subsequent post-processing chain (e.g. use “16-bit TIFF” (= 48 bpp)) if you want to retain maximum theoretical image information, because additional clipping and quantization cab happen in post-processing. Whether you started with 12 or 14 bit RAW images won’t really matter.
Terry
Thank for the more in “depth” explanation. I t seems to agree with what I understand. Do you feel using sRGB had any effect on the result compare to using RGB of ProRGB? I understand the use of sRGB for the web. Other words if I were to print with a ProRGB file, theoretically more color depth, would I have a difference in a 12 bit compared to a 14 bit?
fmalan
To be honest I don’t expect to see a difference. The only difference that I know of is that by default sRGB has a smaller gamut and higher contrast, but this is a feature of post-processing, and not of the image data captured by the NEF file. In this example I zoomed in on a part of the histogram in such a way that it would fit in either the sRGB or ProRGB gamuts.
I did also process the images in Adobe Lightroom (which uses ProRGB) and didn’t see anything new or different.
fmalan
To be honest I don’t expect to see a difference. The only difference that is know of is that by default sRGB has a smaller gamut and higher contrast, but this is a feature of post-processing, and not of the image data captured by the NEF file.
I did also process the images in Adobe Lightroom (which uses ProRGB) and didn’t see anything new or different.
Terry
thanks you the information
Alex
I dont know if is there any real difference between losless and compressed nef’s but (that is what i was searching for) but there is a huge difference between the 12 and 14 bit nef’s. On extreme editing or photomanipulation there is a big IQ difference between the 2 type of raws. Also the test where inacurate. The conclusion should have been “always use 14bit if you know what you are doing”.
Paul
Hi, I agree. I would expect 14 bit images to be much better after some post processing than 12 bit images. As the 14 bit image will have 4 luminance levels for each luminance level in the 12 bit image, so there is less blocking together of very similar luminance levels, probably most noticeable when manipulating in high or low key images with small dynamic range.
If Nikon have provided 14 bit processing then you should assume that the rest of the optical pipeline (sensor, ADC & processing) are designed for it and therefore use it.
Photoshop (well elements anyway) always tries to flatten your images to 8 bit for certain manipulations, but it will not always be so, processing power & memory are always increasing and todays images won’t be that large in 4-5 years time, so keep all the data.
12bit Vs 14bit RAW Files
[…] always shoot the best i can so 14bit, see the test below. 12bit vs 14bit RAW and compressed vs uncompressed? Does it matter? | francoisMalan John. Fuji x20 Black Reply With […]
Dr. Nicolas Rao
As I have let some time pass and in the meantime using creative Photoshop a little more, with a foray into digital art also. I find there are very noticeable differences in editing.NEF RAW files in Photoshop and Nikon NX software.
Even View NX2 with its limited tools does do a more impressive job on the .NEF files. I am sure Nikon does know more about its own software on camera and in its software than it probably discloses even to Adobe whose users may be more in number than just pure NX software users.
I know make a simple change in using 12bit and 14bit. Where time and space is of essence, its 12bit and when I know that my pictures are going in for some extensive editing its 14bit in Nikons own software for me now.
Later once the color profiles, each of which can be downloaded from Nikon free, is set and the file saved. I then take it into Photoshop to create my photos or digital art. There has been an overall improvement in noise profiles and of course color which I did not get by processing directly into Photoshop CS6.
I guess someone more qualified than myself can test out this for themselves. I would be interested to note what they found.
Hans Loepfe
Interesting article, Nice work, thanks. BUT the comparison should be done using Nikon Software to take advantage of the 14-bit instead of ACR (Adobe Camera RAW) used in Lightroom.
These 2 RAW processing engines produce very different results.
My recommendation when shooting with a Nikon: shoot NEF 14-bit uncompressed, develop in Capture NX2 (not in Lightroom), colorspace RGB (not sRGB), buy disk space.
I also agree with the advice given by Dr. Nicolas Rao in his comment above.
Alex
The 14bit raw it’s 14 bit in both Capture NX and ACR but ACR can only read the TIFF related contend of the NEF, can not read the Nikon proprietary contend: AD-L, Picture Controls, vignette control, noise reduction, etc. I use Capture NX and View NX, by far a better choice than Lightroom. But still, on heavy photomanipulationsi import the image in Photoshop as a 16 bit tif by using ACR, just for speed.
fmalan
Thanks for the comment, Hans. You, and some others, have mentioned that I should have used Nikon’s own software to extract the 14-bit goodness. I still have the .NEF files for this experiment and can re-process sometime; it will be interesting to see.
I stand by my statement because many professionals rely on Lightroom for their whole digital development process from start to finish, so even if NX can make the difference clearer, it will not be of any practical use to those photographers.
Secondly, none of you address the very fundamental analysis by Emil Martinec, to which I linked, that explains that the usefulness of increasing bit depth (for a given dynamic range) can only be useful if this finer gradation is not drowned in noise. Which it seems to be, for the current state of the art DSLR.
Keep this site posted for updates to the article using ViewNX2 (I do not own CaptureNX, but assume that it uses the same NEF engine as ViewNX)
Fero
Gary, Joe: there will be no problems! Okay that rlelay means the problems that happen will be dealt with in an expeditious manner which is just like no problems in my mind Things will change a wee bit of course, given the changes to the blog engine. MOST are improvements in depth and scope, but some of the visual changes will be noticed immediately by frequent visitors. Not a lot of worry on my end though cuz ya know what? The killer bee killer is gonna clean out my infestation and I’m going hang gliding next weekend. Life is beautiful. Code is ugly and pernicious, but life is beautiful.
nicolasvrao
Eventually after conversion from .NEF RAW, I have no choice but to go to Adobe CS6 for my photo-editing and my graphics work. I cannot imagine doing frequency separation levels on any other software other than PS CS6 now or in the near future.
So like I said unless time and space are the question 12bit will work out fine. People like Ken Rockwell actually make a living without going into RAW!
But for me and my art work, like I mentioned frequency separation with skin, my mind rests a little easier shooting 14Bit, RAW conversion in NX software and ultimate print prepared in PS CS6.
Alex
You know better what’s the best workflow for you, if you rely more on editing, post processing and manipulation you don’t have to use NX at all as long as Photoshop it’s still the best tool but if you rely the most on the photographic process, if you want to get the best result directly from the camera and you spend a lot of time on that than NX is the only tool you can rely on. Either way, 14 bit is recommended in both cases. Even if you don’t know why the 14 bit photos are better and how to use those 2 bits in you advantage maybe in the future you will know and you will feel sorry if you already shot all your photos 12 bit. SD
nicolasvrao
I guess, only time actually tells, as even from the time of the first post, things have changed so much. My Nikon D7000 has been replaced with the D7100. More and more large Mp cameras are hitting the market everyday.
Entry levels with more than 20 MP, the estimated – at one time – equal to film- seems like tens of years since I heard that! It does seem that even the 14 bit we are dissecting now will be nowhere near where we will be in the near future.
Larger memories and higher processing power is going to make everything we say here redundant in not much time.
In reality however, I think we are actually only limited by what we can see, and that one process, unfortunately for people of my age is something that does not progress! It dims! Sigh!
Alex
We all are limited of what we can see but some people can see more than the others. If some people are satisfy with a RGB mediocre print some people use CMYK with custom color profiles, better than Adobe RGB and still they think they can do better for their clients. If you think that what you see is the limit that perhaps you have a lot to learn and any better gear won’t give you a better result as long as you don’t use the gear you already have at full potential. Cheers.
Stepan
I think the REAL PHOTOGRAPHY isn’t about 12b vs 14b. It can help you sometimes make photo little bit better but that is all.
Hans Loepfe
You certainly good a very good point here.
Alex
Yes, indeed, but what if DaVinci would have painted Mona Lisa with some bad paint that would have decay in time, as a lot of artists from renaissance did. The painters from renaissance where preparing their own paint, that was thier 12bit/14bit debate. There are a lot of masterpieces that did not survived to this day because of their experimentation with paint.
What if you would have made your supreme masterpiece photo with the phone with a bad IQ and the incapacity to print it. Of course that “real photography” is not about that but you would be just ignorant if you ignore this and many other technical stuff.
fmalan
Hey Alex
Firstly: thank you all for your lively comments.
The issue with 12/14 bits is not decay, obviously. If your photographic masterpiece looks great today in 12 bits, it will not look any worse 500 years from now if you were able to save your files all those years.
Alex
Hey fmalan, thanks for the appreciation. The point about DaVinci was that not only the artistic spirit, the talent and the dedication are important regarding art but also the boring technical part. Stepan said that “REAL PHOTOGRAPHY isn’t about 12b vs 14b”, i told him that is about that too, thats why we are talking about it. DaVinci was a genius and had a lot of talent but he was also a master of the technical part, preparing his own paint, that is what i meant. If he ignored that he may have wasted all that talent. If you would have shoot jpg you would have also wasted a lot of talent just because those jpg photos would have lost a lot of information from the photo and would have been almost impossible to process.
nicolasvrao
I have no doubt that there are better printers than me. I was only decrying my age and my declining visual sight. Ha ha! Technically no problem. I printed all my color prints on photo paper using subtractive printing and in CMYK mode only, so thats not anything new to me.
I shoot in Adobe RGB mode as it is available to me. If a bigger gamut was available in camera or there was choice, I would use that.
I don’t think there are too many IQ problems here. I am a physician, studying medicine was not too difficult. My fellowships took a little more trouble and post graduation did too. I presume most subjects are possible once one has applied themselves enough.
Anyway I digress. My complaint was eyesight not insight.
Alex
Haha, i am glad we are talking about the same thing this time even though my other post is also valid, sometimes the things you cant see makes all the difference. Just to be sure, you are talking about your IQ not about IQ as Image Quality, right? :D
Sorry for the confusion, dont take it personally but i thought this thread was about the insight on the outside, not about u :D
Expat
Can anyone share a 14 bit lossless Nikon D7000 file uncompressed (circa 20 mb) ?
Dr, Nicolas Rao
I wonder now that Nikon has so many new model with even larger file sizes, if there is any discernable difference.
We have been discussing the D7000 ad infinitum. There must be more to the new models other than Mp. The camera software must be upgraded and perhaps the use of 14 bit is more apparent now.
Would be good to know. I am also curios about the D4, it is the lowest Mp of the lot and is the top model.
I agree its meant for sports and speed, but does it lack quality from say the D800/800e?
Are there differences in the 14 bit and 12 bit files in it. I am very curious.
Post processing images - The final step to getting that awesome click - Page 6
[…] Posted by izzikio_rage @raja: i didn't know that. would love to read about it though. Here is the post. There could be a lot of arguing over his methods and results but I think he has made a good case […]
James
With the cost of memory cards and hard drives there is ABSOLUTELY no advantage to shooting anything less than 14 bit uncompressed! Your argument is pointless. In fact it is a nonsense
Sherita
This blog was… how do you say it? Relevant!!
Finally I have found something which helped me.
Thanks a lot!
Terry Mieger
For many of us there is a huge disadvantage to 14 bit. It slows your shutter rate down to unbearably slow. NOT good for wildlife. 12 bit is no problem
關於影像大小的選擇 | 大倫攝記
[…] 12bit vs 14bit RAW and compressed vs uncompressed… Does it matter? […]