And here's the image converted with Aperture, without any NR or other steps applied. Just eyeballing the WB and exposure... http://fm.jirman.com/3200-B-APE.jpg
I don't see the big deal here... There's no question the 1Ds MkIII is an awesome camera, no one doubts it (except maybe the dark side folks
Just for curiousity sake, I also tried the following pic at 3200 with a -2 Exposure, which would yield the 12800 ISO, heres what I got, straight out of the camera with NO post-processing except for exposure increased in photoshop:
Camera: Canon EOS 5D
Exposure: 1/6400 sec
Aperture: f/1.2
Focal Length: 85 mm
ISO Speed: 3200
Exposure Bias: -2 EV
Flash: Flash did not fire
- Bibble Ev set at 0
- "Really low" tone curve
- Sharpening at 150 out of 400
- Sensor correction enabled (auto dead/hot pixel removal)
- Sean Puckett's "Sharpie" plugin enabled
- Sean Puckett's "Shady" plugin enabled and set to Ev to +1.75 , smoothness 1, colour 0.258
- Everything else basically at defaults
Output from Bibble to TIFF and run through Neat Image. Auto profiled. http://www.cogitech.ca/photos/838_3_crop.jpg
So far I'm still digging my DPP version, though Aperture comes closest to improving on it. Neat Image seems to just make the image more plasticky. If I'd wanted that look I would've bought a D3! ;-) I don't think it would improve a print. It's interesting how Aperture has retained the colour of the plectrum better than all of them. Maybe stan didn't push it so far.
epuja wrote:
I don't see the big deal here... There's no question the 1Ds MkIII is an awesome camera, no one doubts it (except maybe the dark side folks
Just for curiousity sake, I also tried the following pic at 3200 with a -2 Exposure, which would yield the 12800 ISO, heres what I got, straight out of the camera with NO post-processing except for exposure increased in photoshop:
Camera: Canon EOS 5D
Exposure: 1/6400 sec
Aperture: f/1.2
Focal Length: 85 mm
ISO Speed: 3200
Exposure Bias: -2 EV
Flash: Flash did not fire
Thanks epuja. By my reckoning the only shadowy area of your shot, on the left, looks pretty chopped up. The rest of the frame is brighter, so noise and streaks will be less apparent there. Also, detail looks low. I suspect that if my guitarist shot had been taken with a 5D, the shadow areas would be patchy to the point of ruining the picture. I use both cameras regularly, and I reckon that in difficult low light conditions my 1Ds3 is about a stop better than my 5D. However, you only see that if you downrez the 21 megapixel file to 13 megapixels before comparing them.
brainiac wrote:
So far I'm still digging my DPP version, though Aperture comes closest to improving on it. Neat Image seems to just make the image more plasticky. If I'd wanted that look I would've bought a D3! ;-) I don't think it would improve a print. It's interesting how Aperture has retained the colour of the plectrum better than all of them. Maybe stan didn't push it so far.
I agree with your taste, and considering that Aperture managed that with no extra work, it seems like really useful software as an investment. Might have to consider it, even before I can afford a 1Ds3! Could be good with 1D3?
Thanks epuja. By my reckoning the only shadowy area of your shot, on the left, looks pretty chopped up. The rest of the frame is brighter, so noise and streaks will be less apparent there. Also, detail looks low. I suspect that if my guitarist shot had been taken with a 5D, the shadow areas would be patchy to the point of ruining the picture. I use both cameras regularly, and I reckon that in difficult low light conditions my 1Ds3 is about a stop better than my 5D. However, you only see that if you downrez the 21 megapixel file to 13 megapixels before comparing them....Show more →
The above image is not the original -- It is smaller - hence the 'choppiness', I can post a 100% crop later today.
I would be very interested to compare the same resolution (i.e. downrezd 21mp image, or uprezd 5d image?)
I have no doubts the 1dsMkIII is better quality..its a fabulous camera.
RobertP wrote:
Lookin' good. Your image doesn't really have any bad color noise, and it has great detail, so I honestly don't mind the luminance noise, especially at that ISO.
Most NR software would probably make the image lose some of that "crispy" sharp look to it, so I'd keep it the way it is. You have a winner.
Yeah, PS and LR do fine with color noise. It is the Lum noise that catches us all in the end. When someone develops the algorithms to piece back together an image at 12800 ISO like it was taken at ISO 50, that will be like the guy who invented donuts (and how many donut shops do you see going out of business [Krispy Kream excluded]?).
Would we want that, 12800 ISO at the rate of 50? I've got a great ISO 800 B&W print hanging on the wall that I pulled through noise reduction in LAB color mode: clean as a whistle. Know what? I miss the noise...
Emile Gregoire wrote:
Would we want that, 12800 ISO at the rate of 50? I've got a great ISO 800 B&W print hanging on the wall that I pulled through noise reduction in LAB color mode: clean as a whistle. Know what? I miss the noise...
We would definitely want the option, but not the requirement (i.e. reduce opacity slider).
simonella_viru wrote:
...what do you think is a fair comparison between a super-mega-pixel camera like the 1ds3 and a camera like the d3 that has about half of the pixels?
Print them as a photographer would, not as a pixel peeper would.
32067dlm wrote:
Print them as a photographer would, not as a pixel peeper would.
that's a very obtuse, shallow response. it's akin to saying:
- why don't you go out and shoot pictures
- why don't you switch to nikon
you're out of context of the discussion. people on this forum tend to look for technical, empirical reasons for why product-A is better than product-B. "pixel peeping" is one way to compare.
not everything is printed medium. there is something called the Interweb...
simonella_viru wrote:
that's a very obtuse, shallow response. it's akin to saying:
- why don't you go out and shoot pictures
- why don't you switch to nikon
you're out of context of the discussion. people on this forum tend to look for technical, empirical reasons for why product-A is better than product-B. "pixel peeping" is one way to compare.
not everything is printed medium. there is something called the Interweb...
Actually it's about the only way to do it. Just keep printing larger and larger until one of them falls apart. The main advantage to more mega pickles is larger prints. Other than that there's not a lot of difference between the 1 series bodies.
Gary Petersen wrote:
Actually it's about the only way to do it. Just keep printing larger and larger until one of them falls apart. The main advantage to more mega pickles is larger prints. Other than that there's not a lot of difference between the 1 series bodies.
no way! it's not the ONLY way to compare. as discussed earlier, you can up-scale, down-scale, etc. you CAN print, but you don't have to. also, taking printing out of the equation allows for a more fair comparison.
This thread almost makes me want to stop waiting for a new 5d or contemplate the d700 and get an used 1ds Mark II (I guess that the iso quality of the II is not THAT far behind the III, or what?)
Pixel Perfect gave me a great pointer on how to improve and make easier my noise reduction processing in photoshop, since this thread was started. Here's a 450D shot at iso 3200 (pushed +1 in DPP), and another at 6400 (pushed +2 in DPP), although I slightly missed focus on the 2nd one:
DPReview is helping to perpetuate the popular delusion that the D3 is better at high iso's than the 1Ds3, by comparing 100% crops. I uprezzed the iso 3200 D3 crop so that magnification matches that shown in his 1Ds3 crop, and I shifted the 1Ds3 colours a bit to get close to the D3's: http://cyberphotographer.com/1ds3/dpr_D3v1Ds3_iso3200.jpg
Left is the 1Ds3 crop, center is the modified 1Ds3 crop, and right is the D3 crop. It is fairly clear that the Nikon is doing some kind of colour blending to remove colour noise, but it hasn't entirely succeeded because the large green and red splodges in the upper and lower halves of the tree are artefacts, not true data. Good colour noise reduction would improve the 1Ds3 output here, I think, and it wouldn't need to be quite so brutal as the D3's. It is clear to me from these crops that the 1Ds3 is performing at least as well at iso 3200 as the D3, and it's even capturing significantly more detail.
DPReview didn't test the 1Ds3 at iso 6400 or 12800. Since it outperforms the D3 at iso 3200 in DPReview's test, I see no reason to expect the 1Ds3 to capitulate at 6400, 12800, and possibly even 25600, provided it is allowed some sensible chrominance NR in post.
Impressive image as well. The 1Ds Mark III would complement my 1D Mark III well, other than the fact that the three of us would have to live in the doghouse.
Helimat
I have a 1D3 and a 1Ds3, and they do compliment each other very well. Noise is better controlled on the 1D3 though, obviously because of the larger pixel/density. I don't hesitate to push the 1Ds3 either, but ideally if I'm going into a high ISO shooting situation, like weddings, concerts etc, I grab the 1D3 for its high ISO performance. But for commercial projects, controlled/studio lighting etc, nothing beats the 1Ds3... The two make a perfect pair of tools for the professional that shoots a wide variety of environments and projects.
(I guess that the iso quality of the II is not THAT far behind the III, or what?)
