Can someone explain if there is a big difference in dynamic range between the older Canon 1-series bodies like the 1D or 1Ds and the new canon bodies like the 5D, 1Ds Mk II, or 1D Mk IIN.
Is the difference so great that it is actually visible in the images or is this more hype.
What would one look for when trying to point out dynamic range on a digital SLR. I had a D60 that produced some of the nicest images at ISO 100 compared to any of the other D-SLR's I have used. For whatever reason, the D60 images were simply film like smooth which I have not seen with other SLR's. Was dynamic range a player on that. Any opinions ?
From everything I've seen and experienced it's pretty darn close. In other words, I don't think dynamic range is a useful way of discriminating between these cameras.
Realistically the digital cameras with the best dynamic range are the ones with the lowest noise. That's because DSLR sensors have a very brittle highlight tolerance, but they register shadow detail over a long ambient exposure range. The problem is that shadow detail is replete with noise, and it's very hard to resurrect detail from there. So cameras like the 5D, with their great noise tolerance, will probably give you an extra stop of shadow detail. But technologically they don't differ from one another that much. The camera with hands down the highest DR is the Fuji S3, and that's because it uses a sensor in which 50% of the photosites are ND-filtered to capture 2 or 3 extra stops of highlight detail.
But trust me on this point: most of the time your exposure needs will be met by the current cameras if you know where to meter and how to expose, and it's uncommon that you'll need an expanded dynamic range. I do a lot of zone system photography using black and white film, and this requires that I spot meter all over the place. It's uncommon that I have a scene in which there are more than 5 or 6 stops that require detail (and Ansel Adams himself usually shot scenes with important highlight and shadow details falling over a 5 or 6 stop range). The rare exceptions are sunsets (where I need to grossly underexpose the sky and therefore the ground to get color in the sky) and shots like cathedrals with a dark interior but important detail in the windows.
Yes, there are scenes with 9, 12, 15 stops of dynamic range, and there are methods both with DSLRs and with film for capturing it all. But if you go around with a spot meter, you'll find that it's extremely uncommon that you need to capture even 9 stops of detail in an image.
Yes, there are scenes with 9, 12, 15 stops of dynamic range, and there are methods both with DSLRs and with film for capturing it all. But if you go around with a spot meter, you'll find that it's extremely uncommon that you need to capture even 9 stops of detail in an image.
Yes, and moreover, the real range limiter--for both film and digital--is print paper.
Whenever I hear people talk about a "12-stop" dynamic range for film, when I question them closely, they admit that's "with compensation in processing and printing." (I already knew that, because I was also a dyed-in-the-wool Zonie.)
Well, if we're going to allow compensation in processing and printing, then we should allow it with digital as well. With the proper exposure and the full range of digital post-processing methods, I can pull as much or more range out of the 5D than any ordinarily available film...especially any color film.
Sure, I mean if you bracket enough stops into an HDR image you can capture detail over 15 stops. That's not the challenging part. The challenging part is taking all of that range and making it look like 1) you were photographing the real world, and 2) that your scene actually did have that huge dynamic range.
Similarly, with B+W film you can get 15 stops of detail onto a single negative, then use highly dilute stand development and low contrast grade printing. Bruce Barnbaum is a master printer, and he uses a method to pull 7 stops in negative development, and still manages to evoke it in the print.
But as you say, the challenge is not capturing the entire range. It's making a print that looks realistic and good.
As I and RDKirk make pretty clear it's a lot easier to capture all that information than it is to use it effectively.
With digital you need only buy a tripod and learn how to bracket. With b+w and color negative film you can capture 15 stops in a single handheld exposure if you know how to meter.
The challenge is to get all that info into a useful output. Capturing it is far easier.
High DR provides more nuances when "distributing" raw colors into a subset, like jpeg.
It is also useful when the picture is under/over exposed.
It is why it's important to shoot raw...
Yes, I think the DR increased. For instance, under exposed 1Ds (mk-1) pics are less forgiving than 10D's ones.
Finally, I agree about the "magic" aspect of these old cameras images, D30 and D60.
Could be the pixel size, or the way the AA filtering was performed at that time - don't really know why.
I'm not very smart for I regularly shoot the sun as part of the scenery and doing this with a 10D. What's the DR range, sun to shadow, without blowing out the blue of the sky or the depth of the shadow; single image, no blending?
While ago I proposed a method that one can measure the DR easily and consistently w/o any equipment or chart. And FMers volunteered to post their result. The numbers are pretty consistent with DPR, Imatest and digital imaging resource numbers.
When this method was proposed I had mentioned some methodologic problems with that manner of assessing DR.
My problem is that this methodology didn't assess noise or detail fidelity at the extremes, and absolutely critically it didn't generate a response curve. It also wasn't based on Ev readings of a high DR scene with a quantifiable output measure (i.e. one that could separate a noisy electronic signal with faint image detail from meaningful detail with photographic resolution).
The response curve, again, is critical. At a few Ev above metered middle gray you get a sudden and complete loss of detail capture. And of the total Ev that you can capture with any of these sensors, be it 7 or 9, around 3 stops at the bottom end encompass only about 10-15% of the total image data (hence the common advice to "expose to the right").
So I don't quite believe that the 9 2/3 stops measured by this test completely corresponds to a camera's performance in a 9 2/3 stop scene with critical detail across that entire range.
Paul, that's why I posted the two images as they weren't posted for purposes of vanity. Both images have the disk of the sun in them for reference purposes. I'm not into "DR," so much as I'm into getting the shot and controlling, via GND filters and CirPl, harsh conditions, so as to be able to accomplish this goal.
What's the DR of a capture, going from the white of the sun, not blowing out the blue of the sky, to the depth of the shadows, not blocking up the shadows in the process? These are real world image that should allow one to be able to answer this sort of question.
Etadam wrote:
High DR provides more nuances when "distributing" raw colors into a subset, like jpeg.
It is also useful when the picture is under/over exposed.
It is why it's important to shoot raw...
Yes, I think the DR increased. For instance, under exposed 1Ds (mk-1) pics are less forgiving than 10D's ones.
Finally, I agree about the "magic" aspect of these old cameras images, D30 and D60.
Could be the pixel size, or the way the AA filtering was performed at that time - don't really know why.
Actually this was the question i was trying to get an answer to.
1Ds needs to be metered just right to not get an underexposed image was one of the issues I was trying to get my arms around.
1Ds images are a lot more unforgiving when it comes to underexposure than say for instance 5D images. I was equating this to the fact that the dynamic range of the 1Ds was less than that of the 5D which could be one possible explanation.
From Pondria's table of findings, it certainly seems like the 1Ds has just a little less DR than the 5D.
The findings, although as someone pointed out may not be extremely scientifically accurate, also coincides with my experience with the 1D - which was a little more forgiving than the 1Ds when it comes to underexposed images. I am guessing DR has something to do with how much an image can be corrected.
The other aspect of the DR that is of question is sorta explained in the images Beeman posted. Being able to capture the color of the sky, the sun, and the detail in the shadows tells me how many stops of DR can be captured. The same image would have taken greater effort with my old D60 than with a newer 5D or 1D Mk II.
I remember my Pentax spotmeter often confronted me with values from ev 17 to ev 6 when shooting landscape. We may need a DR of 20 in our cameras before braketing and HDR will become obsolete. However we are very well off today with a camera DR of 8 as Film was much more difficult to use when having high DR
Naturally the body of the sun will always be blown out under normal shooting circumstances, but the nice aspect in your shot is there is a quick and smooth transition to normal sky in that first shot (even with a pleasing rim around the sun). The blocked shadows are not objectionable, and in fact in the larger one on the left there is a hint of detail.
If you actually spot metered that scene including the sun, you'd have a huge range. But if you look at the brightest part of your image that has detail (probably the highlights on the pavement) and the shadow areas that still have some detail (because the darkest shadows, like on the ground to the left are blocked), I'd imagine you've captured image detail over around 7 stops. That's based on a lot of experience spot metering scenes like this. Again, the darkest shadows and bright sun itself don't count as captured dynamic range, because your camera wasn't able to capture any detail there -- but it doesn't matter, because that isn't critical detail anyway, and the shot works completely well without that detail.
So you've shown very well that one can take a camera with a finite dynamic range and with deliberate metering and subject placement still make great use out of a very wide ambient DR.
bathman wrote:
However we are very well off today with a camera DR of 8 as Film was much more difficult to use when having high DR
That I don't really agree with -- high DR scenes are extremely easy to capture with single handheld film exposures, and since I shoot much more film than digital these days it's actually been a huge relief to me. Using color negative or B+W film I can easily capture more than 12 stops without any special procedure, and for greater than 12 stops I can just use a two-developer process (a regular developer for a few minutes, then a dilute developer for the remainder). For color negatives I just ask the lab to do the same.
I just expose such that the important shadow detail falls on zone 3, then take off 30% of development time for every highlight stop that falls above zone 8. This cathedral shot is one in which there was a 12 stop dynamic range and I was able to easily capture it with a single film exposure.
I've learned to selectively lasso and level individual areas, using the history brush to blend the lines of demarcation so as to expand the DR of a capture. This shot with the extremely limited D30; about a six and a half stop range sensor. I lassoed and leveled this image three times to go from the blue of the sky, the harsh reflected sun light on the walls (walls are not blown according to the color pick information in the original) to the depth of the walkway shadows and not blow the highlights.
This year, looking to replace the L-358 meter with a L-758DR (I have but don't use the PW capabilities) for it's spot metering capabilities and will pick up a Lee filter system to augment/replace the screw on GND filters currently being using.
To me, it's incumbent to control the light to our advantage.
Edited by BeeMan458 on Feb 28, 2007 at 07:37 AM GMT
Is there such a standard way of measuring so called Dynamic Range? I am not aware of any standard posted by the ASA or similar organization.
Also, I agree 100% that what does it really matter on the 'academic' dynamice range if you cannot translate it to a printing paper? How many zones are in the zone system and how many zones out of the 10 zones are really meaningful in the real world? I would rather look at DR as a 'relative' meaurement rather than an 'absolute' number, similar to how we discuss the 'sharpness' of lenses.
When one shoots in raw and have to go through a data conversion using certain softwares, how much impact is on the so called DR just because different software is used? Would the coversion software is as effective/critical as in the BW film development methods that can expand and contract the DR on films?
DrPablo wrote:
That I don't really agree with -- high DR scenes are extremely easy to capture with single handheld film exposures, and since I shoot much more film than digital these days it's actually been a huge relief to me. Using color negative or B+W film I can easily capture more than 12 stops without any special procedure, and for greater than 12 stops I can just use a two-developer process (a regular developer for a few minutes, then a dilute developer for the remainder). For color negatives I just ask the lab to do the same.
I just expose such that the important shadow detail falls on zone 3, then take off 30% of development time for every highlight stop that falls above zone 8. This cathedral shot is one in which there was a 12 stop dynamic range and I was able to easily capture it with a single film exposure.
That sounds much more difficult to me. Just imagine most 35mm photographers would enter the cathedral with less preparation and probably do not want to develop the whole film N-3.
Slau, there are 10 zones in the system. it does not define DR but it is a tool to handle the transition between the scenery and the final print. It is about visualisation
Yeah, the zone system isn't a measurement of real world dynamic range. It's a planning system for the placement of different tones in a print. In general, for a scene with a typical brightness range, each zone will be one stop apart from the next.
By definition middle gray falls on zone 5, and a reflective meter will give an exposure reading that (with normal development and printing) will give a zone 5 placement. And you go from there -- if you want a shadow to fall on zone 3, then you just meter off that shadow and reduce exposure by 2 stops. If that shadow placement causes your brightest highlight (with important detail) to fall on zone 7, then you're done -- you just expose, develop, and print normally. If your brightest detail falls on zone 9, then you would want to use contracted development (which will darken highlights but not shadows). Etc. So the zone system is flexible for different scene brightness ranges, and you use different metering, exposure, development, and printing strategies based on the scene's requirements.
Andi -- yes, you're right that it's harder to be quite so creative in the darkroom if you're shooting a 36-exposure roll. Sheet film is easier for that. But with my Hasselblad, and 12 exposures per 120 roll, it's no big deal to shoot all 12 in sufficiently similar conditions so as to control development that way.
Also, variable grade contrast paper makes it very easy to print negatives that you haven't gotten 100% right in development.
