Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Tutorials for Newcomers to Premiere Pro

Exporting from a 4:43 5kFF sequence to 1080p prores 444. CUDA renderer.

With RR enabled in Pr and MRQ checked it says aprox 1h to export.
Without MRQ it says 8m.

With RR disabled and MRQ checked it says 2h.
With MRQ unchecked it says 18m.

So the RR does seem to be working. I just don't understand what's happening with the Max Render Quality option...

Hopefully someone from Adobe would step in and answer this question?
 
Brad Allen said:
Hmm, not sure on this. I was under the impression that the RRocket would assist debayering at any resolution. It's just that below 4k it will accelerate it faster than real time (whereas 4k is about realtime and 5k is almost realtime from memory)
Click to expand...

I'm not exactly sure how it works but I believe the RR ALWAYS debayers at full resolution and the computer then has to downscale to whatever you have set. So theoretically based on that assumption using the RR should be the equivalent of MRQ, that's why I'm confused over these results.

But I could be wrong.
 
Tehben Dean said:
Exporting from a 4:43 5kFF sequence to 1080p prores 444. CUDA renderer.

With RR enabled in Pr and MRQ checked it says aprox 1h to export.
Without MRQ it says 8m.

With RR disabled and MRQ checked it says 2h.
With MRQ unchecked it says 18m.

So the RR does seem to be working. I just don't understand what's happening with the Max Render Quality option...
Click to expand...

This seems to confirm my suspicions. It's not that the RRocket isn't being used, it's simply that the downscale from 5k>1080p is a very intensive operation and demands a hunk of processing time from your GPU or CPU (which is dependant or not on whether you have CUDA enabled on a supported GPU or not). Based on this post from Todd, it seems that you can use your 650M either in OpenCL or CUDA - it might be worth checking out both of these to see if one of them performs better than the other?
 
Tehben Dean said:
I'm not exactly sure how it works but I believe the RR ALWAYS debayers at full resolution and the computer then has to downscale to whatever you have set. So theoretically based on that assumption using the RR should be the equivalent of MRQ, that's why I'm confused over these results.

But I could be wrong.
Click to expand...
Yeah, I can see why that would add to your confusion. I'm almost certain the rocket is still in effect when debayering at less than the original frame size. For example, think about when you are in Premiere on the timeline and watch things back in 1/2 res or 1/4 res. If you have your Rocket enabled, then previewing at these resolutions doesn't mean that the rocket is no longer used and it defaults back to the CPU, it simply means that the RRocket is then used to debayer at 1/2 which it's much faster at doing. (I doubt the CPU would be able to keep up with 1/2 debayer of 5k in a MacBook Pro). Same thing in RedCine-X, when looking at footage you will see that the pre-buffering bar fills up faster than it would if only using the CPU even when previewing at 1/2 res etc.

EDIT: I was completely wrong on this point. Red Rocket debayers at full resolution whenever in use.
 
Last edited:
Brad Allen said:
Yeah, I can see why that would add to your confusion. I'm almost certain the rocket is still in effect when debayering at less than the original frame size. For example, think about when you are in Premiere on the timeline and watch things back in 1/2 res or 1/4 res. If you have your Rocket enabled, then previewing at these resolutions doesn't mean that the rocket is no longer used and it defaults back to the CPU, it simply means that the RRocket is then used to debayer at 1/2 which it's much faster at doing. (I doubt the CPU would be able to keep up with 1/2 debayer of 5k in a MacBook Pro). Same thing in RedCine-X, when looking at footage you will see that the pre-buffering bar fills up faster than it would if only using the CPU even when previewing at 1/2 res etc.
Click to expand...

That makes a lot of sense :)

Now I just need to do some tests to see if there is really any noticeable difference exporting from the full resolution or not when mastering to 1080.

Thanks for all the help!
 
No Prob at all.

You will probably find that your export results when using R3D's at non-max render quality are still quite good. They can sometimes almost look sharper (in the same way a 5D by pixel cutting can look sharper) than a full debayer. I usually advocate using non-max render quality for anything but where the absolute quality is desired. ie for any previews of cuts or copies to send to clients etc.

Then I will also use an unsharp mask when doing the final Max Render Quality to ensure that I have a decent amount of sharpness on that export too :)
 
My guess would be Premiere's downscaling algorithms are more complex than RedCine-X's? But I'm really just taking a stab in the dark here

Here's some info regarding scaling in Premiere Pro that's probably worth reading up on.
http://blogs.adobe.com/premierepro/2010/10/scaling-in-premiere-pro-cs5.html

I think it's definitely worth contacting Wes who frequents the forums here and run the performance numbers by him that you are seeing. He would definitely be a more definitive answer on why we are seeing what we are. Plus if there is a big that is slowing things down, helping him find the problem will be the best way for the Adobe team to fix it :)
 
Brad, thanks for the educations videos. Any updates here in 2013? More videos?
 
You must log in or register to reply here.
Back
Top