Stoffer wrote:
Banding/pattern noise @ ISO 250. There is a touch of that in the shadows, but it is somewhat random and not really that strong IMHO. It is not Sony EXMOR clean, but way better than 5D Mark II. Not a big concern for me at this point, but we obviously only have one shot to go by. It might even be better in the final production camera, if they didn't calibrate the demo camera 100%.
Indeed. I think it's better than I expected. I pushed it 3.5 stops and could definitely see noise in the shadows that I *suspect* wouldn't be there on a Sony sensor.
So be it... it's significantly better than the 5d2 though - by a very long way.
For me and my shooting style, the ISO3200 shadow noise is the biggest improvement followed by the ability to shoot at ISO12800 and have a usable file. I wouldn't even touch ISO6400 on my 5d2.
I don't think it will register as a 2 stop increase over the 5d2 - I think it'll be about 1 stop better in RAW - but the noise characteristics are SO much nicer that I get another stop of usable range. And I'm very impressed with the colour rendition at 12800 as well!
Even if you push the NR up at 12800 the LR algorithm really does handle it well and leaves a good amount of detail in the file.
For me this is where sites like DXO miss the point. They can only tell you what there is and isn't - not how usable it is.
Well I'm incredibly pleased with my (pre)purchase! ISO3200 is as far as I go for most of a wedding and that's looking pretty close to noise free now and the shadow noise has improved incredibly. Well done Canon!
PhilDWedding wrote:
I don't think it will register as a 2 stop increase over the 5d2 - I think it'll be about 1 stop better in RAW - but the noise characteristics are SO much nicer that I get another stop of usable range. And I'm very impressed with the colour rendition at 12800 as well!
Even if you push the NR up at 12800 the LR algorithm really does handle it well and leaves a good amount of detail in the file.
For me this is where sites like DXO miss the point. They can only tell you what there is and isn't - not how usable it is. ...Show more →
Agreed! And Phil; great wedding photos on your site. I wish i could have hired you for my wedding 5 years ago.
PhilDWedding wrote:
I spoke to Canon. They said RAW noise improvement over the 5d3 was about 1/2 stop.
People keep quoting JPEG improvements which is where all of the unlikely differences are coming from. Physics simply wouldn't allow a 1.5 stop increase unless Canon had found something unbelievable..
yep, I can believe 1/2 stop in RAW. IMHO, shooting jpegs with this $7K camera should be a mortal sin punishable by decameration...
PhilDWedding wrote:
I think it'll be about 1 stop better in RAW
You're talking about the 5D3 here.
I just saw a report by a guy who seems very knowledgeable. (He wrote a book on DSLR shooting.)
His finding was that, when comparing the 5D2 RAWs side-by-side with the 5D3 RAWs, there was almost NO improvement. I saw some of the shots he had right next to each other magnified, and luma noise looked almost identical at high ISO! And in some cases, the 5D2 high-ISO files actually looked better in regards to luma noise!
I don't remember the guy's name but he just put up a YouTube video detailing his findings. Perhaps someone else here would have the link to that.
D. Diggler wrote:
You're talking about the 5D3 here.
I just saw a report by a guy who seems very knowledgeable. (He wrote a book on DSLR shooting.)
His finding was that, when comparing the 5D2 RAWs side-by-side with the 5D3 RAWs, there was almost NO improvement. I saw some of the shots he had right next to each other magnified, and luma noise looked almost identical at high ISO! And in some cases, the 5D2 high-ISO files actually looked better in regards to luma noise!
I don't remember the guy's name but he just put up a YouTube video detailing his findings. Perhaps someone else here would have the link to that....Show more →
Cool but is there any way you can post an ISO100 RAW? A frame that is pure black, body cap on (if a lens is mounted it has to be at least 2.8 or SLOWER, as the camera tinkers with RAW file exposure when it sense faster lenses and it makes hard to know what is really going on), in dark room, say ISO100,1/8000th is best but even just any ISO100 file is somewhat OK since the masked areas can be used to an extent.
It's easy to figure out a rough idea about dynamic range and shadow quality from a low ISO file.
I'm curious to see if it has the same old DR as the old or if they managed to do something new for the 1DX.
Banding aside, it did't seem like the 5D3 even had a half a stop on the 5D2. Maybe 1/3 or even less.
One thing to keep in mind is that the 5D2 also applies like 1/3 stop less gain at each ISO so if it looks 1/3 stop better at the same ISO for each, in reality, it is more like 2/3 stops better SNR.
