p.1 #1 · Compare High ISO Perfromace 5DII vs 5DIII RAW
Hi all, (Reviving this thread due to 5DIII announcement)
With lots of people throwing around insane high ISO numbers like 51k and speculating like on 1DX performance please enlighten me with a comparison I can wrap my head around. Note that I am only talking about RAW performance and am completely un-interested in any comparisons that involve .jpg.
I shoot a 5D II and consider the IQ at ISO 3200 very good for shooting events like comedy shows or bands playing in bars. This is my "max" ISO if at all possible. (for comparison, I consider my max ISO on 60D or 7D to be 1600 to get roughly equal noise performance to the 5DII at 3200) I am interested to know what cameras capture images at ISO 6400 or 12,800 with equal quality to my 5D II at ISO 3200. Will the 1D4 or the 1D3s? What about the 1DX? What about Nikon's current cameras or the rumored D800. Note that by equal quality I am not talking about resolution, just noise (I don't print large). This is the only high ISO comparison that interests me and it seems that when Canon and Nikon talk about ISO 51k and 102k they are talking about low quality noisy raw images I would never use and it is all just a marketing spec war. Please help me understand this and feel free to discuss anything related. I would love to have a "list" or some easy to reference equivalent like 5DII ISO 3200 = 7D ISO 1600 = 1D4 ISO 6400 or some such roughly estimated chart.
I realize of course that all of this is a gross oversimplification
p.1 #2 · Compare High ISO Perfromace 5DII vs 5DIII RAW
The nikon D3s seems to be the low light king, especially at iso 6400 and on, so if i weere looking to upgrade high iso shooting, there arent really that many other options. d700 is also slightly better at high iso, but just barely.
p.1 #4 · Compare High ISO Perfromace 5DII vs 5DIII RAW
I don't think we will be able to fully understand the 'extreme low light' capabilities that Canon is bragging about with this camera until we see some real sample shots at upper ISO's. We have not been able to see any full res samples to fully understand what this camera can do. It would be sweet to be able to shoot at 51k and have it equal your 5D2/3200. Time will tell.
p.1 #5 · Compare High ISO Perfromace 5DII vs 5DIII RAW
saneproduction wrote:
...I would love to have a "list" or some easy to reference equivalent like 5DII ISO 3200 = 7D ISO 1600 = 1D4 ISO 6400 or some such roughly estimated chart.
Ask and ye shall receive.
Some people here deride DxO Lab's tests, but one thing about them is that they use a consistant testing methodology. You might find their noise ratings useful as a way of comparing one camera against another.
I picked some cameras from their comparator including the ones you named and a random selection of Nikons. They don't have tests yet for the 1Dx.
Of those I chose, they rate noise-handling ability this way from highest to lowest:
Nikon D3s -- 3250 ISO (The current champ in noise handling.)
Nikon D700 -- 2302 ISO
Canon 5D Mark II -- 1815 ISO
Canon 1Ds Mark III -- 1663
Canon 1D Mark IV -- 1320
Nikon D7000 -- 1167 ISO
Canon 60D -- 813 ISO
Nikon D300s -- 787 ISO
My poor old Canon 20D doesn't fare too well: 721 ISO
The relevant section says, " Thus low-light ISO is the highest ISO setting for a camera that allows it to achieve an SNR of 30dB while keeping a good dynamic range of 9 EVs and a color depth of 18bits."
p.1 #6 · Compare High ISO Perfromace 5DII vs 5DIII RAW
RobertLynn wrote:
If you max the 7D at 1600, I have a feeling you'll never be happy :P I use it at 6400 like it is the job of the camera :P
Hi Robert,
I like the 7D and sometimes shoot it at higher ISOs no problem. I just felt like the files are around .7-1 stop noisier than my 5DII. My intention of this comparo is not to disperage any camera or high ISO shooting paractices, just to try to get a rough idea of comparative noise and what the future holds. The main merit of an image has nothing to do with noise, camera etc ; it is the photographer (and your work is excellent).
My problem is that I am looking to get 1 more stop out of my tele lenses and if I could shoot 6400 or 12,800 with the same low noise as the 5DII at 3200 that would aid me in the kinds of caves that I end up shooting in a lot. The other option is just to get a super fast 200mm lens which I am also exploring. If I could get usuable ISO 6400, f2.8 200mm may be enough (which it is not now).
p.1 #7 · Compare High ISO Perfromace 5DII vs 5DIII RAW
BrianO wrote:
Ask and ye shall receive.
Some people here deride DxO Lab's tests, but one thing about them is that they use a consistant testing methodology. You might find their noise ratings useful as a way of comparing one camera against another.
I picked some cameras from their comparator including the ones you named and a random selection of Nikons. They don't have tests yet for the 1Dx.
Of those I chose, they rate noise-handling ability this way from highest to lowest:
Nikon D3s -- 3250 ISO (The current champ in noise handling.)
Nikon D700 -- 2302 ISO
Canon 5D Mark II -- 1815 ISO
Canon 1Ds Mark III -- 1663
Canon 1D Mark IV -- 1320
Nikon D7000 -- 1167 ISO
Canon 60D -- 813 ISO
Nikon D300s -- 787 ISO
My poor old Canon 20D doesn't fare too well: 721 ISO
Wow! This is exactly perfect! This explains the 1 stop noise difference I note between Canon's current 18MP APS-C sensors and the 5DII almost exactly. Suprised the other Canon cameras actually rate lower! The D3s is the current champ and it is only 1 stop better than the 5DII. Hope that means the new Canon stuff will be 1 stop better (similar to the D3s). The other thing that impresses me is that the 20D is only around 1 stop down from the 5DII and within 100 points of the 60D. To me this says that the differences between cameras are pretty exaggerated by marketing.
p.1 #8 · Compare High ISO Perfromace 5DII vs 5DIII RAW
I am not talking about resolution, just noise (I don't print large). This is the only high ISO comparison that interests me
A strange comment. I find programs like Noise Ninja have only a slight negative effect on resolution, while giving quite considerable noise reduction. Just take any digital camera image and sledge hammer it with the maximum that the noise reduction program allows.
The huge problem for 99% of us, is noise reduction while maintaining very high resolution, permitting large prints.
p.1 #9 · Compare High ISO Perfromace 5DII vs 5DIII RAW
I care alot about shadow noise, so for me looking at the "fully SNR" page on Dxo is far more telling to how I will perceive performance.
If you look at the 5d mkII at iso 100, you see its performance drop like a rock at the lower light values (shadows). Helloooooo banding / terrible noise floor.
The standard Dxo noise measurement is at 18% grey or some other fairly bright value.
p.1 #10 · Compare High ISO Perfromace 5DII vs 5DIII RAW
saneproduction wrote:
Wow! This is exactly perfect! This explains the 1 stop noise difference I note between Canon's current 18MP APS-C sensors and the 5DII almost exactly. Suprised the other Canon cameras actually rate lower! The D3s is the current champ and it is only 1 stop better than the 5DII. Hope that means the new Canon stuff will be 1 stop better (similar to the D3s). The other thing that impresses me is that the 20D is only around 1 stop down from the 5DII and within 100 points of the 60D. To me this says that the differences between cameras are pretty exaggerated by marketing....Show more →
Dxo measures RAW data, marketing and most internet opinion is based on JPG output from the camera.
p.1 #11 · Compare High ISO Perfromace 5DII vs 5DIII RAW
saneproduction wrote:
...The other thing that impresses me is that the 20D is only around 1 stop down from the 5DII and within 100 points of the 60D. To me this says that the differences between cameras are pretty exaggerated by marketing.
I picked up a T3i on Friday, and although I haven't had a chance to do any exhaustive testing one thing that I seem to see in casual comparisons with my 20D shots is the effect of resolution on visible noise.
While there may only be 92 points difference in the DxO Mark rating between the two (the T3i uses the same sensor as the 60D), the T3i (18Mp) has more than twice the resolution of the 20D (8Mp), so when displayed at the same size the noisy pixels from the T3i are smaller than those from the 20D, and so the image looks much cleaner than the numbers alone might suggest.
p.1 #12 · Compare High ISO Perfromace 5DII vs 5DIII RAW
I am sure that the Dxo site is measuring something scientific and significant. In my opinion, however, the real world results don't match up. I grant the D3s as the high ISO king (ranked #1) but if you look at the list of the top ten or so cameras, It doesn't make sense to me. The 1Ds mkIII and original 5D are rated better at high iso than the 1D mark IV? I don't see that nor have I ever heard it from another source.
What am I missing?
p.1 #13 · Compare High ISO Perfromace 5DII vs 5DIII RAW
System noise is what count at the end. Not the sensor noise alone, Canon needs to spent more money to get the over all system noise down. Who needs DSLR IQ when zero read noise can be done in point and shoot camera.
p.1 #14 · Compare High ISO Perfromace 5DII vs 5DIII RAW
Here are some sample 1:1 crops from un-processed raw images (just imported into LR3 with the default settings, which I know has a little noise reduction applied).
p.1 #15 · Compare High ISO Perfromace 5DII vs 5DIII RAW
Don Clary wrote:
A strange comment. I find programs like Noise Ninja have only a slight negative effect on resolution, while giving quite considerable noise reduction. Just take any digital camera image and sledge hammer it with the maximum that the noise reduction program allows.
The huge problem for 99% of us, is noise reduction while maintaining very high resolution, permitting large prints.
I have not explored noise reduction plugins or really anything outside of what LR3 can do. I will check into this more. Maybe I am just being a stop to picky on the noise I am seeing.
p.1 #16 · Compare High ISO Perfromace 5DII vs 5DIII RAW
mttran wrote:
System noise is what count at the end. Not the sensor noise alone, Canon needs to spent more money to get the over all system noise down. Who needs DSLR IQ when zero read noise can be done in point and shoot camera.
p.1 #17 · Compare High ISO Perfromace 5DII vs 5DIII RAW
Here is an example from the darkest concert I ever shot (at a speakeasy with an extra lamp the band brought because I asked about lighting (you can see it on top of the right speaker) ) Also unprocessed as above.
p.1 #18 · Compare High ISO Perfromace 5DII vs 5DIII RAW
mttran wrote:
System noise is what count at the end. Not the sensor noise alone, Canon needs to spent more money to get the over all system noise down. Who needs DSLR IQ when zero read noise can be done in point and shoot camera.
amen,
and to the crops posted, if you look at the full snr link, read noise issues in canon are more prevelent in the lower iso range, resulting in reduced DR at low iso and some banding (in some cases with some bodies).
At higher iso, the signal to noise curve is much flatter (which is good).
p.1 #19 · Compare High ISO Perfromace 5DII vs 5DIII RAW
pjbuehner wrote:
I am sure that the Dxo site is measuring something scientific and significant. In my opinion, however, the real world results don't match up. I grant the D3s as the high ISO king (ranked #1) but if you look at the list of the top ten or so cameras, It doesn't make sense to me. The 1Ds mkIII and original 5D are rated better at high iso than the 1D mark IV? I don't see that nor have I ever heard it from another source.
What am I missing?
The 1D mark IV has less color sensitivity than the 5D and the 1DsIII. Using weaker color filters is unfortunately one of the methods used for improving noise performance.
The DXOmark low-light ISO value catches this. If you look at the individual measurements for SNR, DR and Color sensitivity, you will find that color sensitivity is in fact the limiting factor for the ISO value. BrianO referred the criteria in his post. I prefer the individual measurements over the low-light ISO.
p.1 #20 · Compare High ISO Perfromace 5DII vs 5DIII RAW
pjbuehner wrote:
I am sure that the Dxo site is measuring something scientific and significant. In my opinion, however, the real world results don't match up. I grant the D3s as the high ISO king (ranked #1) but if you look at the list of the top ten or so cameras, It doesn't make sense to me. The 1Ds mkIII and original 5D are rated better at high iso than the 1D mark IV? I don't hhsee thhingat nor have I ever heard it from another source.
What am I missing?
+1
Something fishy about those results. They are surely per pixel noise
and not how you would compare cameras at all.
