The quiet revolution in photography
Shutter speeds? Yaaawn. ISO speeds? Oh-god-not-again. Megapixels? Oh puh-bloody-lease, that’s so 2003. The newest frontier of digital photography is dynamic range - and it’s arguably the most exciting (r)evolution that’s happened in dSLR-world so far.
Interestingly, most manufacturers are continually improving the dynamic range of their cameras, but somehow seem to forget to tell us about it - which means that we’re witnessing - or should I say not witnessing - a quiet revolution.
It seems as if ‘dynamic range’ gets forgotten in PR world, where a bigger screen, better battery life or Live View is an easier way of getting regular consumers exited. The real technological leaps have been happening under the bonnet, though, and the result of the ongoing improvements will mean that your next camera will be significantly better than your current one - but you wouldn’t be able to tell from just reading its specification sheet.
So, why, exactly does this make a difference to us as photographers? All will be revealed…
Is this the same as HDR?
Well, we’re still talking about ‘Dynamic Range’, and higher-dynamic-range-than-before at that, but when people are usually talking about ‘HDR’, they mean multi-shot HDR photography, which we covered thoroughly a few weeks back.
Multi-shot HDR is very exciting stuff, and it’s a taste of what is to come, but this time around, we’re talking about single-shot HDR photography and how cameras have been steadily improving over the years.
The improvements have happened so slowly, it seems, that nobody has really noticed - but grab a 5 year old dSLR and compare it to a current-day snapper, and the biggest difference in picture comes from the dynamic range.
Let’s start with a quick poll
Last year, I ran a poll to see how many of you used JPG versus RAW. As it turns out, 28% used exclusively JPEGs, while 27% used mostly JPEG files. Pure RAW users were about 22% (see graph to the right). So let’s try again…
(to see the full results of the old poll, and of all the other polls we’ve run in the past, check out the creatively-titled ‘past polls‘ page.
Whatevz, can we start from the the beginning, please? What is dynamic range?
The human senses have an absolutely incredible dynamic range - think about it: when you’re inside a concert venue at a rock gig, you can hear every note and enjoy every instrument.
When you’re in a quiet room, you can hear water flowing through your radiator heater, and the extremely faint buzzing of the phone charger is loud enough to notice. More incredibly still, you can keep a conversation going with someone in the next seat while the jet you are sitting in is taking off, which is a testament to the width of dynamic range which is possible - although that particular example has more to do with psycho-acoustics than your ears themselves.
If you thought your hearing was amazing, well, your eyes are even more incredible. On a bright day, you can see perfectly, but you can also see things by moon- and starlight on a clear night. Not impressed? Try taking a photograph or do some filming by starlight without using a tripod…
Now, these examples of your eyes’ dynamic range come with a caveat - you cannot stand in a dark room and look out into a sunny landscape and see both perfectly - you’ll either be able to see the indoors, or the outside, with the other being over- or under- exposed, respectively. For photography purposes, the important thing is how much dynamic range you can see simultaneously.
Allow me to illustrate: Take a landscape photo where the clouds are nearly over-exposed. The dynamic range of the particular imaging-chip you’ve got, decides how much details you get in the darker parties of the image. The higher dynamic range, the more shadow-details you can expect.
A theoretical camera with perfect dynamic range wouldn’t need shutter speeds - you would select an aperture to get your depth of field, and you could select any shutter speed you need. The sun would have texture, and the deepest, darkest shadow parties of your image would still have detail in them, too. Of course, perfect dynamic range is impossible (for now…) but that doesn’t mean that increasing dynamic range isn’t a great thing…
Riveting, I’m sure. But is it really that different from 5 years ago?
Back when I first started taking digital photographs in the mid-1990s (I know, we still listened to The Cardigans, Tracey Bonham, Marcy Playground and Tonic…) and I did some playing about with shooting in RAW format, comparing it to just using the JPEGs straight out of the camera. Back then, I decided that RAW was a complete waste of expensive memory stick space, because it was nigh-on impossible to spot the difference. I didn’t know why that was the case back then, but I think the answer is pretty clear right now: The dynamic range of 8-bit JPEG photographs was, in fact, pretty similar to that of the imaging sensor inside the camera.
A couple of years ago, I believe when I got the then-brand-new-on-the-market Canon EOS 20D, I decided to switch to RAW. I spotted that the quality was better than with JPEG, and I stuck with it. Mostly, I did it because I could never quite get the white balance right, and with RAW, you defer the decision until you’re at your computer, which always suited me quite well.
More recently, I upgraded again, this time to a Canon EOS 450D, and the difference is quite noticeable - right from the start, I felt that the 450D was taking better photos than my old 30D, but I was struggling to figure out why. Ignoring the resolution (the 30D delivers 8.2mpx files, whereas the 450D has a slightly smaller imaging chip, but saves 12.2 mpx files to the memory card), the 30D is a better camera in all possible ways: Better top ISO, faster top shutter speed, better autofocus, quicker continuous drive, magnesium body, and so on and so forth. Nonetheless, I swear I was getting better photos with the 450D than with the 30D, on quite a consistent basis.
Then, finally, a few days ago, the penny dropped. I did some test shots on my 450D, setting it to shoot JPEGs, and then some more shooting RAW. The difference was vast - by using Photoshop’s built-in RAW editing tool, I was able to pull incredible amounts of extra information out of the RAW images from the 450D, compared to the ones from my 30D.
Now, add to that the fact that the Canon EOS 450D is Canon’s entry-level digital SLR, and that Canon’s R&D department haven’t been twiddling their thumbs in the meantime either - but as always, they save the best goodies for the people who cough up serious cash for the more hard-core semi-pro and professional lines of cameras.
I haven’t had a chance to have a go myself yet, but it’s rumoured from various fora that the Canon EOS 5D mk2’s RAW files (in addition to being full-frame and significantly higher resolution) have a 15-bit dynamic range which is completely out of this world.
What’s in it for me? How does this mumbo-jumbo make my photos better?
Much in the same way as how tastefully done multi-shot HDR photographs can look realistic and fantastic at the same time, single-shot HDR photographs can do a lot of good for you as a photographer.
Already, photographers all over are shooting in RAW instead of JPEG, because they’ve noticed that it’s a Better Thing - but only rarely do people stop to think why that might be. The reasons are above: you simply gain a lot more flexibility by having a higher dynamic range to play with, than if you limit yourself to the 8-bit limit of JPEG.
This extra flexibility isn’t just camera geekery either: It’s extra raw data in your photograph which you can genuinely use to deliver better final photographs. When I’m out taking photos in difficult lighting situations (such as dancing, concert photography or similar), I now routinely dial back the exposure by a full stop.
Yes, I know that it means that I lose some data in the top end, but because I’m shooting in RAW, I get away with it: The software will save me, and, more importantly, I can get a full stop faster shutter speed. When you’re out taking photos at a concert, the difference between 1/60th of a second and 1/100th are significant.
It isn’t just in poor light that the benefits are obvious, however - In fact, I can’t think of a single genre of photography where extra flexibility isn’t a good thing.
Look at it this way: If you go on holiday and bring two sweaters, you can always choose to wear the thick one or the thin one; when you’re shooting in RAW, you can always decide to go with the suggested automatic choices, and at worst you’ll have lost a few megabytes of storage space for a few hours (or days, or weeks, depending on how long you keep your photos on your camera), but seeing how cheap memory cards are these days, that’s hardly a huge problem - the extra flexibility is there when you need it, and it’s better to throw away data when you don’t need it, than to wish you had more when you don’t.
OK, I’m convinced, how can I join the fun?
So, how can you gain from all this extra goodness already? Easy - just set your camera to RAW. Stop reading right now, and set your camera to RAW. Yes, you. Yes, now. Then experiment. See how much your photos allow themselves to be tweaked without losing significant quality in the process.
What’s coming up in the future?
Frankly, I think digital SLR cameras can only go in one direction.
We already have higher resolutions than we know what to do with (I stick to my opinion that we never needed more than 6 megapixels, although it’s nice to be able to crop your images when needed without quality loss), and the top professional cameras are currently better, both in terms of dynamic range and resolution, than the top film-based SLR cameras ever were (and are rapidly closing in on medium format film cameras), and there are DSLR cameras that can film in full 1080p high definition video.
The way forward, is that dSLRs drop further in price, as the components that go into them get cheaper, and people get bored of the megapixel race. Canon, Nikon, if you’re listening, start selling a $200 8-megapixel sub-entry-level, and you’ll make enough money on the licensing of your lens mounts to make up for the loss in body sales; much like the way printer manufacturers (including Canon, interestingly), sell cheap printers with expensive ink refills.
The other boundary that needs to be pushed is dynamic range - I want a camera with completely ridiculous dynamic range, please, and I don’t mind if I have to sacrifice a bit of resolution or ISO speed to get there either. 20 bits worth would be nice. 24 bits if I can get it, so the dynamic range of my camera matches that of my screen.
Having such a camera means that I can become sloppy, but I can still rescue any photo unless I really balls it up. More importantly, however, it’ll allow me to do stunning HDR photos in a single click of the shutter. And, finally, it’ll be the last nail in the celluloid for those poor sods who still hang on to their film with a desperation which is inversely proportional to their dignity - and directly proportional to the grin on my face.
But seriously - start using RAW now, you might be amazed at how good your camera really is.
Do you like this article? Tell the world using one of these:
[
Digg
|
Facebook
| 
del.icio.us
| 
Stumble Upon ]
reddit_url='http://photocritic.org/dynamic-range-in-photography/'
Categories: PC articles | Photo Theory | Photo tips
Tags: Dynamic Range | HDR | High Dynamic Range | poll | RAW
This entry is © Haje Jan Kamps / Photocritic.org, and was posted on Thursday, November 13th, 2008. You can follow any responses to this entry through the RSS 2.0 feed. You can skip to the end and leave a response. Pinging is currently not allowed.
#1 - November 13th, 2008 at 01:51
I completely agree. I think what HDR has done is shown people what you can really properly do with a RAW file. Theres so much data in there that you never really use and if you run it through HDR software you can bring that out. Who knows but in 5-10 years time it’ll be automatic. Its just a question of storage and CPU power really. Theres an advert for some compact camera that even mentions about retaining shadow and highlight detail. Its a consumer compact camera for fun days out with your friends. They don’t care about shadow and highlight detail but it is becoming a feature to market. Dynamic range is definitely going to be the next big step.
#2 - November 13th, 2008 at 09:21
great article! Ends the RAW vs JPG debate once and for all!
#3 - November 13th, 2008 at 16:33
With regard to the “RAW versus JPG” debate, it doesn’t end anything. The choice between the two is really dependent upon your situation and your expected results. Like choosing an automatic transmission versus a stick-shift for a car - there are situations when one is more appropriate than the other. I agree that RAW is superior - but I’d lose of a lot of event shots waiting for my camera’s buffer to clear if I shot every photo in RAW. I use it when it is appropriate for me to do so.
With regard to the new frontier of high dynamic range, you should consider looking into what is really on the horizon - not just high dynamic range, but mixed or selective dynamic range.
Film has an ISO. The whole roll is stuck at that ISO (discounting pushing, pulling, etc). Digital has the advantage of changing ISO on the fly - but you are still using a single ISO for any particular image. High dynamic range just gives you a wider gamut - which is great, a step forward, but there’s something better on the horizon.
There are several patents listed and some new products pending that purport to offer mixed ISO on the same image. That is, dynamic range where the actual sensor sensitivity, and thus the low and high end of the dynamic range actually differ on the same image. This is more like the human eye sees - our eye is not stuck at a single ISO. Not only do we have ‘high dynamic range,’ but our brains also selectively enhance parts of any scene we view - and suppress others.
This will be a true step forward in dynamic range - not just ‘more’ but ‘better’ dynamic range.
#4 - November 13th, 2008 at 16:43
Good article but remember that the sensor dynamic ranges you quote are per component - a 12 bit sensor captures 12 bits of red, 12 bits of green and 12 bits of blue giving you about 36 bits per pixel (let’s forget about Bayer patterns for simplicity). Similarly, a JPEG encodes 8 bits per component for a total of 24 bits per pixel. Like the JPEG, your screen (or, more accurately, your graphics card) is most likely 8 bits per component (24 bits per RGB pixel) so your camera RAW file already has a far higher dynamic range than your display card can render.
#5 - November 13th, 2008 at 18:01
Great article! My camera is one of those older models (Nikon D70) where I don’t see much difference between RAW and JPEG so I’ve always shot JPEG. Now that I’ve upgraded to using Aperture 2, my workflow really won’t be much different shooting either RAW or JPEG except using up memory card space faster (which, as you pointed out, really isn’t a factor with prices today).
When I upgrade my camera, I will definitely revisit RAW and the next time I am in a tricky lighting situation will flip it to RAW and see what it can do for me.
Here’s to the future!
#6 - November 14th, 2008 at 13:59
The Fuji S5 Pro has great dynamic range, when I bought it, the dynamic range interested me more than all the other numbers. Then I sold it and went back to film, realizing the dynamic range of digital isn’t in any way important to my photography.
But I agree that Dynamic Range of a digital camera should be advertised moreso than megapixels, as well as buffer size and the card write speed or whatever it is.
If any digital camera can produce an image like the one on my ground glass, I will buy it, so far I’ve only found my eye which can produce an image like that.
#7 - November 14th, 2008 at 16:02
Very interesting article that will definitely have me using raw more often. I agree that a camera’s dynamic range abilities are not touted nearly enough - largely because the public is uninformed on the topic…way to bring about a discussion. Wigwam Jones - great additional points.
#8 - November 14th, 2008 at 20:58
Question time for the uneducated. I use a Powershot S5 IS. Will I notice the differences with my lower end camera enough to justify switching over from JPEG to RAW?
#9 - November 18th, 2008 at 04:34
Wow, I never knew about this stuff. That’s really interesting and very exciting. Thanks for this article. Unfortunately I don’t have the technology to easily upgrade to RAW format with my shitty computer and lack of photoshop at home so it will be a while before I can integrate this into my personal (non-professional) workflow.
Now when it comes to your last bit about film, I agree that digital will eventually take over. For the time being, the quality of a $50 film rangefinder is still astonishing, so don’t be a dick. Point and shoot digital cameras have got nothing on what was available in film.
#10 - November 18th, 2008 at 06:32
Once I began learning to use raw, I use it whenever I can. I keep my camera set to Raw + Jpeg in fact.
Back when I was going to school for photojournalism, two years ago, we were told not to bother with RAW. Big, slow, and with timeliness jpgs were the way to go.
So I keep my camera on Raw + JPG currently for that reason. I find Raw far better in most regards, but the times I need quick turn around (half hour at times) my computer works faster with jpgs.
Then I can turn to the raws to process these photos for the more polished versions.
#11 - November 18th, 2008 at 11:25
Dave Wilson, I agree with you about the bit depth of the monitor, but remember that what is important isn’t how much information you can see but how much information you can work with “to deliver better final photographs”. Remember that one stop can do a lot of difference.
#12 - November 19th, 2008 at 01:55
Rui,
I agree 100% - the more bits of resolution I have in my source data, the more I have to work with later in Photoshop or wherever. The point of my comment, however, was that the colour resolution of RAW files already well exceeds the ability of most graphics controllers and displays to show them. To see what I mean, open a 32 bit HDR file (not the tone-mapped output, the original version with 32 bits per sample/96 bits per pixel) and see how terrible it looks on the display.
#13 - November 20th, 2008 at 07:54
Correct me if I’m wrong, but I think bit-depth and dynamic range are different concepts:
If you think about a brightness histogram, the dynamic range is the difference between the lowest and highest values on the X-axis, while the bit depth determines how many individual samples exist along that axis, i.e. how many digital buckets the A/D converter splits the analog signal into.
If this is right, it should be possible for a manufacturer to increase the dynamic range of a sensor without changing its bit-depth. You would end up with a more quantized brightness signal, but you would still have a wider dynamic range. Hence, I wouldn’t see bit-depth increasing exponentially as you’ve suggested, because for every doubling of the dynamic range you would only need to add one bit to maintain the same quality of signal quantization.
While tone-mapped HDR imagery can produce some really awesome photos that were never possible with film, I think it’s important to remember that a wide dynamic range is only really useful when capturing high-contrast scenes. In fact, one impact of having a higher dynamic range in your camera is that scenes that don’t naturally have a wide range of light levels will appear quite flat (contrast- and colour-wise). Whenever we open up the Levels or Curves dialogs in Photoshop and pull those little arrows inwards to “give it some more punch” we’re quite often *reducing* the dynamic range of the image (although the major manipulation is over the contrast).
I love what HDR can do, but I can’t agree that it’s going to change the course of photography.
#14 - November 20th, 2008 at 07:58
By the way, I think you’re spot on about RAW. If you’re going to do any post-processing that changes colours, having 14-bits of colour depth is vastly superior to 8.
#15 - November 21st, 2008 at 05:07
A generally excellent article, but your massive blooper at the end regarding screen bit depth spoils it somewhat.
In addition to what Dave Wilson has already said, I’d like to add that simply increasing the bit depth of the RAW file won’t necessarily make things much better by itself, that’s a red herring. The lower bits (or higher depending whether the RAW data is big- or little-endian) of the 12/14/15 bits per channel that RAW files already contain is mostly random noise, and it’s noise that is the enemy of dynamic range (at the bottom end of the intensity spectrum anyway), as it limits the number of discernible shades of near-black. Eliminate more noise and you have more shades of black to record the image with.
One way to radically improve dynamic range is to combine high-sensitivity CCD pixels with low-sensitivity ones, as Fuji did with previous incarnations of its “SuperCCD SR” technology:
http://en.wikipedia.org/wiki/Super_CCD
Their latest version is somewhat different and even more exciting, as you’ll realise when you read about it:
http://www.dpreview.com/news/0809/08092210fujifilmEXR.asp
In the past Fuji licensed the D200 camera body from Nikon and stuck in their own SuperCCD (and software) to create the S5 Pro:
http://www.trustedreviews.com/digital-cameras/review/2007/04/30/Fujifilm-S5-Pro/p1
Hopefully they’ll do the same again with their new sensor and a new Nikon body (e.g. D3 / D700 / D90) to give us the best of both worlds… or perhaps the other way round, Nikon/Canon/someone else might licence the latest SuperCCD technology from Fuji? Fingers crossed!
#16 - November 21st, 2008 at 11:55
Also meant to mention this useful new tool/site from DXO Labs that (among other things) lets you compare dynamic range between cameras:
http://www.dxomark.com/
#17 - November 28th, 2008 at 09:32
Great article. Does throw up some interesting questions around your method of underexposing by one stop. My understanding is that when a DSLR exposes for a given scene, it looks for the brightest area and allocates half it’s available resources to that part of the image. It then allocates half of what’s left to the second brightest area, half of what’s left to the next brightest.. and so on. Meaning that by not making use of the right end of your histogram, you are losing an awful lot of capacity. My D3 does a far better job of recovering highlight detail than recovering shadow detail - which tends to support that theory, in my experience. If my images are over exposed by a stop or so, I can nearly always end up with a perfectly acceptable result - if under exposed, noise in the shadow areas (if recovered ‘heavily’) is a problem. Maybe it’s because I do a lot of my D3 shooting at high ISO’s?
Michael Reichman (of Luminous Landscape) is also a big advocate of ’shooting to the right’, i.e. exposing so that your histogram makes the most use of the right end of the scale.
Would be interested in anyone else’s experiences.