Jim, Ted or Graham...DSP question?

[John Allardice]

In terms of picture quality, when stepping down from the RAW Bayer pattern to say, 2K, is the internal DSP going to be significantly better than scaling in post?
I know that the DSP is handling considerably more than just scaling and that there's also good solid bandwidth reasons for the '2k scaled' recording....just wondering if the final picture quality would be any different scaling in say Nuke or Shake, as opposed to having the DSP do it real-time?

 

[Graeme_Nattress]

The DSP will do broadcast quality downscaling. I don't see any reason why the in-camera downscaling shouldn't look utterly superb. BTW, it's Graeme, not Graham. Scottish spelling, remember?

Graeme

 

[John Allardice]

The DSP will do broadcast quality downscaling. I don't see any reason why the in-camera downscaling shouldn't look utterly superb. BTW, it's Graeme, not Graham. Scottish spelling, remember?

Graeme

Aw crap, sorry about that....I really should have known, being a transplanted Glaswegian myself.

The reason I asked was not, BTW, due to any doubts concerning the quality of the DSP, it was actually concerning the old question of 2.35 on the RED.

I was wondering if it would be preferable for the DSP to scale the 2.35 center strip of the sensor to the standard 2K anamorphic format(2048 x 1736, with a 2:1 pixel aspect). Rather than just record 4k then crop & stretch later.

Now obviously there's some pretty strong arguments in favor of this format in terms of bandwidth, recording time, etc. I just wanted to check that it would be optimum in terms of picture quality.

Think of it as Digital Super-35.

DS-35...hmm...that has a good ring to it, yes?

J

 

[Graeme_Nattress]

For super-widescreen, I think it's going to be best to crop / scale in post. I'd like to see if we can get variable cropping in-camera, but I'm not quite sure on that yet.

Scaling to to 2:1 PAR is a bit silly to do in camera, methinks (personally) if you don't have to especially if what you really want is the full rez from a 1:1 PAR image. Scaling to a 2:1 PAR is just a form of compression, and it makes post a pain as you have to deal with a non-square PAR format.

Graeme

 

[John Allardice]

Ok, it's just that, given the fact that (only if you're going back to film) you'll eventually HAVE to scale to a 2:1 PAR if your final format is anamorphic..

...God that was a hideous sentence, read it a couple of times through.

..when would the best point be to introduce the squeeze?

I know from working on a good few anamorphic & super 35 features that it can be done either way, full-ap or squeezed. Just looking for the best quality.

J