PhotogDave
Offline
• •
Upload & Sell: Off
|
Steve Spencer>
The reason that you need to match resolution when comparing output for noise is that you may actually want to print your images and when you print the images you might even want to print at a specific size say 8 X 10 or 18 X 12 or whatever. If you care about printed images, then you don't care about what it looks like at 100% on the screen, you care about what the print looks like. And the problem is that if one camera has higher resolution often it will look noiser at 100% on a monitor, but may look less noisy in the prints that are produced. So when you look at a 100% on the monitor you are misleading yourself about what your prints would look like. That is why you need to do comparisons that match resolution, because making prints matches resolution. With that said the problem of not matching resolution is bigger the wider the discrepancy in resolution, so comparing 50D and 7D files won't be as bad as some comparisons, because they are fairly close in resolution. Still compared to what you see on a monitor at 100%, prints from a 7D will have less noise than prints from a 50D even if they have the same amount of noise on the monitor. ...Show more →
This is absolutely wrong. Making 8x10 prints from a 7D file vs a 5D file will not match resolution. The 7D file for an 8x10 crop would be 4320x3456(14.9MP) and the 5D2(17.5MP) there for, the 5D2 8x10 print will show less noise than the 7D because it has more pixels per inch in the same size print than the 7D does.
All you do is chop off a small portion of the pixel deminsions. Making the print matches the print size, but the higher resolution file has more pixels per inch in that print than the smaller, cause you must crop the smaller file to fit the aspect ratio also.
Its pointless to downsize the 5d file to match the 7D file for printing purposes, no one does this. If they want an 8x10, they crop it accordingly and are left with the remaining pixels.
They dont down size the 5D from 5616x3744(5D2) to 5184x3456(7D), then crop for 8x10 from there, its stupid. You just crop out 16% of the original file and done.
All I was saying is, it doesnt matter about how it looks at 100%, the 5d2 makes a 21mp image, the 7d makes a 18MP image, so compare what the sensor makes. You dont down res the bigger image to see if it compares with the smaller one.....BECAUSE IT DOESNT PRODUCE THAT SIZE IMAGE. So it doesnt matter, its pointless. Of course the downsized 7D file will be cleaner than the full size....so, thats a manipulated file. Thats not what the sensor makes. The 18MP file from the 7D should be compared to the 15MP file of the 50D and viewed at normal viewing size, not 100% and which ever is cleaner, then so be it. Canon doesnt produce images to be viewed at 100%. They make images to look a certain way when viewed as images, whatever it takes.
It produces what it produces. Its that simple. To me, looking at the 5D2 files side by side to the 7D files, they look very close except for the 5D banding. At 100% its a different story, but again, the image wasnt intended to look good at 100%, its supposed to look good at normal viewing size.
The real test would be set up a 70-200 on a tripod(cause the lens wouldnt need to be moved, just dismount the camera body from lens to swap.)then mount the 5D2 and set lens somewhere in the middle. Use f8 and whatevevr ISO, snap picture, then dismount, mount 7D and adjust zoom to match the FOV of the 5D, same settings, snap and view those side by side. Not 100%. This would be easier to see the difference, cause the RG pics are different images at different angles etc, so you cant really see the image side by side.
If you dont do it this way, and you just down size the image to match FOV or whatever, you change the playing field for one sensor by altering its pixels, let it produce the image its designed to make and see what wins.
In regards to my comment about the sensor specs, measurements....its all crap.
Those specs can be whatever, it doesnt matter until you see the real images. I dont care about what the SNR is, what the black offset is, what the sigma is, what the R and G and B channels look like....it doesnt matter.
When the rubber meets the road, what does my final image look like out of the camera. Canon can post all this they want. They did it with the 50D and everyone got all excited, they did this with the 1D3 AF supposed to be the best ever made and smoke anything in the market. I read all this when those came out here and on Canons site, and what happened. SPLAT!
WHen the rubber hit the road, all that crap that everyone thought they new about the cameras looking at specs, meant nothing once it hit the road. This is what I mean, we cant be sure until those of you or I get our hands on one just like the 50D and just like my 1D3 to see the real deal.
Yes, those specs mean something in producing an image, of course, but the specs never tell the real story, ask those who bought into the 50D that got burned.
It looks promising, but we cant think by looking at a couple images online we know the outcome in the real world. IMO, it already looks to be alot better than the 50D outcome, just dont get ahead of yourselves.....cause none of us really know yet.
|
come on.