The proof is in the filming...

R

Ruzo

Well-known member
A rushed test between the D90 and D300s. Some settings might not be the same but I think the compression difference is obvious, isn't it? I'm getting a little more familiar with the D300s and I know the perfect trees that will challenge the compression when I make time for the next test. Don't ask me to make any advanced test because I'm no expert but I'll follow any suggestions if I can :). So, for the first "comparison":

D90 frame grab on Edius:
D90.jpg



D300s frame grab:
D300s.jpg
 
Clearly its much improved ! I don't see any stair stepping in the 300s clip and looks much sharper overall. Still it has that " this is how I think film looks " look to it. Certainly is much better then the d90 , and if this test is a fair one... it makes the Nikon video progress somewhat more interesting for sure.
 
Here is another one. I tried to match every settings possible. I locked the exposure before going into LiveView. I will redo the first shots when I have some time.

D90:
D90_2.jpg


D300s:
D300s_2.jpg
 
Last edited:
it does look much better, along the lines of what we've speculated based on seeing d300 clips, but the frames are too different to really tell. in the first example the light is completely different and in the latter it looks like it's not even the same f-stop? did you use the same lens and, as far as it's possible, the same settings?
 
Yep, same lens, same f-stop. The first set is admittedly not as prepared as the second one.
 
the compression is exactly the same. not only is the bitrate the same, the quantization tables are actually identical (i've posted them in a previous thread if you don't believe me). the difference in quality is obvious, but it's not the compression. and forgive me for saying so but i've been saying that all along. :-)
 
Matt,
I'm realizing more and more how much better it performs than the D90. What do you think could be the reason for the improvement?
 
a lousy debayer has always been my key suspect in the case of of the fuzzy d90 image, as well as possibly an equally lousy noise reducer. the lack of really sharp fine detail and lack of color definition especially around edges is my circumstantial evidence, and the reason i never believed it was the compression is both testing the same table on high quality data and simply zooming into live view and observing that it already looks that way before it's recorded.
 
This would truly mean that they are putting some effort into improving the quality of their video. I have a feeling that the D3s is going to be amazing. Let's hope they have the sense to include manual control this time. A dedicated Nikon Digital Film Camera may be next ;)

By the way, if you guys want more "tests" than let me know. Otherwise, it will not be fun to waste time on these if there isn't much interest...
 
mattsand said:
the compression is exactly the same. not only is the bitrate the same, the quantization tables are actually identical (i've posted them in a previous thread if you don't believe me). the difference in quality is obvious, but it's not the compression. and forgive me for saying so but i've been saying that all along. :-)
Click to expand...

Oh, I belive you. You are certainly more technically adept than me. I guess I mistook the lousy debayering for macroblocking.

I must say I find the D300s interesting, not because I would ever get one, but because the images shows great promise if we ever get manual control in a later Nikon model or firmware.

Alos, Mattsand, can you share some of your thougts on the Nikon look versus the Canons? To my eye, as I've said, it looks more organic and seems to handle highlights better. If you agree/disagree can you give us some technicallities as to why this is/is not the case?
 
I think Nikon went with a few changes on this camera. I'd say they focused more on the photography side. Maybe some internal adjustment to improve the jello.
They are probably going to do more on the next model next year. D400? Maybe a D700 upgrade this year.
I agree with you that the look of the Nikon done right is so much better. Just think if they can improve and get 1080 HD on the next camera!!! But I still see video that is all over the place. I'm not sold on the 7D either and I might keep waiting.
 
chagchag said:
Alos, Mattsand, can you share some of your thougts on the Nikon look versus the Canons? To my eye, as I've said, it looks more organic and seems to handle highlights better. If you agree/disagree can you give us some technicallities as to why this is/is not the case?
Click to expand...
i agree, and i think it's due to the preservation of color all the way in the highlights, but i can't really say why. probably the conversion from raw to rgb (the same that's causing the bad quality as well). there are many different ways of doing it. take a still and experiment in photoshop or whatever with different debayering, dithering, noise reduction, and so on, the look can change dramatically.
 
mattsand said:
keep em coming. a couple of test chart would be nice to look at, and some low light stuff.
Click to expand...

I'm afraid that test charts are over my head ;) I could try to get some if you point me to the right direction.

For a low light shot, I think this guy has a great one (even though the out-of-focus part seems a little long):

http://www.vimeo.com/6487886

If it has to be done another way, let me know.
 
A search is always easier when you know what you're looking for ;) I don't have 15.75"×8.45" paper so I'm not sure if it will work at 8.5"x11".
I'm gonna be a little busy for a few days but I'll see if I can do something by this weekend. One thing I observed is that putting the AEL on before going into LiveView doesn't seem to work the same way as the D90. It seems to setup its ISO more rapidly. We'll see...
 
Actually, I have a Dell 2405FPW 24" monitor, viewing the PDF full screen makes it much bigger then 8.5x11, would the white light mess it up?
 
You must log in or register to reply here.
Back
Top