November 27th, 2007, 06:31 PM #271
A quick example. These are small portions taken from two of the photos I took. These are window blinds in the corner of my office, in shadow. The blinds and wall are white.
D300 at 1600 ISO
D200 at 1600 ISO
November 27th, 2007, 08:05 PM #272
Congrats on the new body! No fun not having it work as you expect out of the box, but good to hear it seems to be working now. As for the high ISO noise (or more accurately, the lack thereof) I have been reading that the D3 has fairly aggressive and sophisticated noise reduction algorithm going on in camera, and I wonder if the D300 has the same thing, only less so, but I have no idea how you would test that.
I have been itching to upload my Thanksgiving pics, but work has been an absolute bear this week (there is even an HDR sunset since you guys have been talking about them)
November 27th, 2007, 08:19 PM #273
Thanks! I think that there might be some truth to the idea that they are applying more sophisticated algorithms. But I think the fundamental design differences play a part too. In the case of the D300 it is moving to a CMOS sensor which inherently produces less noise. In the case of the D3 it is the larger sensor and micro lenses. Whatever it is I'm happy to have more sensitivity.
The D300 also employs an optional D-lighting function to preserve highlight detail, which is kind of intriguing. Lack of exposure latitude and highlight clipping is one of the things that has always has bugged me about digital camera images. I will have to play with it and see what it does.
Also, today's announcement by Nikon of loadable user settings (available on their site) that match color to the D2X is kind of cool too.
November 27th, 2007, 11:13 PM #274
Thanks for the sample pics. Amazing difference!
Anything worth shooting is worth shooting twice. Bullets are cheap. Life is priceless.
November 27th, 2007, 11:14 PM #275
November 27th, 2007, 11:37 PM #276
The images were shot using mostly default settings and saved as jpegs. I wanted to see what the camera's processing would do with the images. Although now that you mentioned it, I went back and checked the NR settings and the D200 was set to "Low" rather than "Normal" which would have been the default. So maybe the D200 could have done a little better. I set it to low for my normal shooting and do NR in Photoshop later and I forgot to reset it. So this comparison may not be totally fair, but I will say that my experience with the D200 is that auto white balance is a problem at high ISO and the D300 appears to be better at that.
This is a similar image shot with the D200 at the same exposure with NR set to 'NORM"
D200 - 1600 ISO with "Normal" high ISO NR
It is less grainy but detail is lost. The detail of the vertical lines of the blinds holds up much better with the D300. To me, the D200 NR looks blotchy.Keep in mind that this is a 465 pixel square out of a 3872x2592 image. It also looks like I might have a stuck pixel. Also notice color balance is different from the prior image. It's the consistency of auto white balance that is the problem at high ISO. I sometimes take continuous shots and they are all just a little different--this is under fluorescent of vapor lights at high ISO.
Last edited by DeaderFan; November 28th, 2007 at 12:03 AM.
November 28th, 2007, 12:18 AM #277
Ah, less noise for sure on this one.
Originally Posted by DeaderFan
I can see the problem the AWB is causing. It's acceptance of what is white sure is taking a lot of the white out of it.
November 28th, 2007, 08:55 AM #278
Here is an very technical (so probably not very useful in real world terms) comparison of the D300 and 40D sensors. But what is interesting is the D300 does some processing before the file is written to RAW (it looks like the 1Ds MkIII might do this as well now, which just seems to go against the whole RAW thing). Also I found it interesting that there isn't much difference in dynamic range between the 12 and 14 bit modes for the D300 (or for any camera), about 1/3 of a stop, which means that the overall size of the loaf of bread is about the the same for 12 and 14 bits, but with 14 bits your slices of bread will be thinner, which means the noticeable difference should be better gradations and less posterization.
On a completely unrelated note, it looks like the 1Ds MkIII will release Friday! At 21mp it will be interesting to see if they can keep the noise under control. If it does, it could give some competition to the medium format bodies for studio shooters, but with the distinct advantage of 5fps. You just need to be ready for the 25MB image size.
For 8 large, it better be good
November 28th, 2007, 12:32 PM #279
Interesting article, thanks. It looks like the D300 is a significant improvement over the D200 but the 40D is still has the edge in IQ. My understanding is that the D300 uses a Sony sensor that does some NR on the chip. Sony has manufactured mostly CCDs in the past and so perhaps they are a little behind Canon in CMOS technology. The article appears to confirm that the D3 is likely a state-of-the-art benchmark camera.
I am also anxious to see what the 1Ds Mark III can do. The density of the photosites would be comparable to a 14MP DX/APS sensor which would lead you to think noise might be a problem. I will be interested to see what kind of technology they employ to address noise and light fall off in the corners of the image. I'm not sure I'm ready for a 21MP image in a 35mm format though. The workflow would consume a lot of resources and you really can only use the very top quality lenses, which is another huge investment. Although it is still probably cheaper than most medium format digitals.
November 28th, 2007, 12:49 PM #280
I'm looking forward to it, as well.
Originally Posted by DeaderFan
The non-S Mark III got a firmware upgrade today. Did they not recall these in the US?
Originally Posted by Canon EOS-1D Mark III Firmware V. 1.1.3
Last edited by Unfiltered; November 28th, 2007 at 12:52 PM.