Had a look at RG's ISO 12800 samples. Since these are pushed digitally, they tell a lot about Dynamic Range at lower ISO's, and pattern noise.
Well, the 7D seems free of banding/pattern noise, as apposed to the 5DII, which is useless to me at 12800 because of the banding. I do recognize the overall lower noise level from the 5DII, but even after NR and downsizing, the banding will be visible.
This will translate into wide dynamic range and good shadow noise performance at low ISO's for the 7D.
Jim Levitt wrote:
Everything about this camera seems great, except the 18mp on that small sensor. For people who work outdoors, at lower iso's (which, I will grant, is the majority, especially birders and sports shooters) this will be a benefit. I do most of my work indoors in low light. This same camera, at 12 or maybe even 15 mp, could have been the lord of darkness.
Larger pixels do not give significantly better low light performance. That is a myth. The factors which affect low light performance are sensor size, technological maturity, and manufacturer. This camera would have about the same low light performance whether it was 10 Mpixels, 18 Mpixels or 30 Mpixels. You, and the majority, are confused because you are considering noise per pixel instead of noise per square inch of print. Even famous review web sites like DPR make this mistake, so it is not surprising that many are misled. Noise per square inch of print is the real measure of a camera's ability to capture a picture in low light, and noise per square inch of print is largely unaffected by changing the pixel pitch, whereas it varies dramatically with sensor size, sensor age, and to a lesser extent, manufacturer.
I notice you often mention "on-print" with defraction limits and pixel size, etc. What (the few?) of us that Never print? It seems per pixel sharpness is more important to me when cropping for screen viewing. I sold my 50D at a huge loss cause images did not seem as sharp at the pixel level as my 40D or 1D mark III or especially, my 5D "classic". So if you do not print, seems bigger pixels ARE more important.
I notice you often mention "on-print" with defraction limits and pixel size, etc. What (the few?) of us that Never print? It seems per pixel sharpness is more important to me when cropping for screen viewing. I sold my 50D at a huge loss cause images did not seem as sharp at the pixel level as my 40D or 1D mark III or especially, my 5D "classic". So if you do not print, seems bigger pixels ARE more important.
No it doesn't if you compare images at the same output size (i.e. if you are going to view the images on a 1920x1080 display, you would resize the images to 1920x1080).
To add to what Brianiac said, the key factor for getting decent images when downsampling a higher resolution camera's image is the read noise of the high resolution camera vs. the lower resolution camera. As long as the read noise of the high resolution camera isn't significantly higher than the lower resolution camera (and it doesn't introduce fixed pattern noise) you will get equivalent high ISO performance by downsampling. However these things fall into the "technological maturity" category that Richard talks about.
All indications so far are that the 7D has impressive read noise and has virtually no fixed pattern noise.
abqnmusa wrote:
So we are looking at a print on screen, and clicking it to 100% to view detail in the images. That is where the per-pixel noise does become an issue.
To which I have to ask why? Why are you viewing an image on screen and then clicking to view 100% other than because its simply possible? Are you doing so to attempt to see fine detail? In that case, in order to adequate compare a low resolution camera to a high resolution one, you need to equality the playing field by upsampling the image of the lower resolution camera to that of the high resolution camera. If you do that, you'll probably find the high resolution camera, despite higher per pixel noise still has more image detail.
abqnmusa wrote:
You click on an image to view at 100% to see all the detail. I suspect everyone does that when viewing a full size image.
with print film you would print larger sizes to see the detail
with slide film you would project it to see more detail
with full size digital images you look at 100% to see all the detail
I don't understand this line of logic at all. You mention that "per pixel noise is an issue" when viewing at 100%, which I can understand, but again you're seeing an increased level of detail. For "normal" image viewing at current day monitor resolutions you're not going to see any increased level of noise. Right? So, beyond the academic purpose of viewing pixel level details I'm not sure why "per pixel noise is an issue." I suppose if you are sending 100% crops to other people to view....
I agree with Brainiac for the most part, but to my mind the main advantage of a googlepixel sensor is that you can either print very large and still view from close up, or you can crop significantly. Either way, per-pixel performance has some significance.
WillWeb wrote:
I agree with Brainiac for the most part, but to my mind the main advantage of a googlepixel sensor is that you can either print very large and still view from close up, or you can crop significantly. Either way, per-pixel performance has some significance.
Look at the advantage of a high resolution sensor this way:
- At low ISO, you can print very large and get superb detail
- At high ISO, (given Richard's pre-requisites) you can downsample and still get images with the same image noise as a low resolution camera
So then the only advantage of a lower resolution camera would be FPS and buffer size.
abqnmusa wrote:
You click on an image to view at 100% to see all the detail. I suspect everyone does that when viewing a full size image.
with print film you would print larger sizes to see the detail
with slide film you would project it to see more detail
with full size digital images you look at 100% to see all the detail
Sure, but when I present images digitally, I rarely zoom in to 100%. Unless I'm trying to show off my camera or lens, I'll downsize the photo to fit on my monitor/digital picture frame/hdtv. That's going to be the same size regardless of whether or not I used a 10D or a 7D to take the shot.
