DNxHD v ProRes
 

I have just purchased the Atomos Samurai and have done some very quick connect & test shots using a Canon XL-H1, just to explore the Samurai. I captured a shot on both codecs at each resolution, purely for comparative purposes.

I stuck the clips in Premiere and added 3 secondary color corrections and pulled the colors as hard as I could push them.

What I noticed was that the DNxHD clips (even at 8 bit) seem to have far less noise than the Prores clips, even before my crazy color correction. Also, with the color corrections, the prores files seem to have more spill onto objects not part of the key, whereas the DNxHD clips play beautifully and the color corrections seem more contained and cleaner.

I know this is more of a subjective rather than objective test, but would be interested to know what others have experienced.

Re: DNxHD v ProRes
 

DNxHD has been through SMPTE qualification while Pro Res hasn't (I don't think) - most people choose ProRes if they are using FCP and DNxHD if Avid.

If you are finding DNxHD better I'd stick with it especially if you can eventually bypass QT in an MXF wrapper.

Re: DNxHD v ProRes
 

An old topic, but still of interest since not handled well in the public information I think.

Generally you may not find a significant difference between DNxHD and ProRes in terms of the two codecs.

BUT there can be a significant difference how the two codecs are hanled in the NLE that you use. Even if Atomos states that every NLE can handle both codecs - what is true if you install either the free Avid codec for DNxHD or the Quicktime player for ProRes. But even then there may exist significant differences how the codec is decoded in you NLE.

Especially the luminance behaves different in different NLEs. Tested with testbars, I have seen that ProRes is not handled correct in Vegas Pro 12 for example - the luminance is spread here; while DNxHD is handled in a correct way in Vegas Pro 12. In Edius it is the other way round - ProRes is fine in Edius, but DNxHD is not fine in edius in terms of the luminance handling.

So one should be tested with the system that is used - to my opinion.