Can Twixtor interpolate 23.976p to 29.97p in 29.97p project?

[Matt Carlson]

From my experience Twixtor does not work well if the Vegas project settings are greater than the frame rate of the event the plugin is on. It will create a bunch of clumped static frames at intervals. Because of this using Twixtor solely as a standard frame retiming plugin seems like a bad idea.

If you are dead set on Twixtor because you are not happy with any other frame generation plugin’s output the only solution I can think of is a ton of work. Set the project properties to 23.976 and size the event to the amount of frames it would be at 29.97. Set Twixtor to 80% and render the event as an image sequence and re-import at 29.97 after changing the project properties back to 29.97. Twixtor will not have any static frame clumps that way and the extra frames generated will be better than Vegas’ native resample. If you have a lot of 23.976 events then this is probably more work than you are willing to do.

[Rick Anvican]

Hi Matt,

Thanks for sharing your experience, my current workaround is that I would speed up the 23.976p footage to 29.97p and since I’m doing some processing in Virtualdub for the 23.976p footage, might as well do that there, render it back out in lossless and re-import that version into vegas, then “twixtoring” it. This would probably give me the best workflow. Anyway, thanks again for your suggestion.

[Ian Pearson]

I don’t have Twixtor, but I have successfully accomplished this with great results with a free tool called SVP (Smooth Video Project). The tool imitates the “motion flow” interpolation effect on a pc. While it was designed for playback, it does have options to output the results to a file. I can’t remember the exact process, but I think it will output an Avisynth script, then you can use that in whatever encoder can use a script. You mentioned Virtualdub which can import an avs file. You may be able to do all your processing in one step in Virtualdub.

[John Rofrano]

[Rick Anvican] ” my current workaround is that I would speed up the 23.976p footage to 29.97p and since I’m doing some processing in Virtualdub for the 23.976p footage, might as well do that there, render it back out in lossless and re-import that version into vegas, then “twixtoring” it. “

Why don’t you just place the 23.976p footage in a 23.976p project and use Twixtor on it and then drop that project into your main project as a nested project? This would solve the frame rate problem with Twixtor and eliminate the need for any rendering.

~jr

http://www.johnrofrano.com
http://www.vasst.com

[Norman Black]

Another non Twixtor option is Prodad ReSpeedr. It is stand alone and thus you get a render out of it to use.

It is an optical flow resampler like twixtor, but since it is stand alone you can specify the output frame rate of its render.

If you do not change the speed slower/faster then in effect you just get a frame rate resample. Currently selling for $80 US with a free trial.

[Rick Anvican]

Hey guys, thanks for all the suggestions, I will have a go at them soon. Initially, I had the 23.976p footage slowed down using velocity envelopes where duplicate frames are distributed evenly, but I switched to time remapping where smooth motion is required and duplicate frames are not wanted at all.

Twixtor has shown unwanted warping (e.g. straight edge becomes curvy) on a few frames even with keyframed adjustment, considering the footage had quick panning and distinct edges, so I’ll see if the alternatives suggested give better results, again, thanks for the help.

John, nesting would work just requiring a bit of recalculating speed values but would take more time to render, especially with projects having many event FX/transitions and layer compositing, pre-rendering the “twixtored” footage would cut down rendering time I suppose.

[John Rofrano]

[Rick Anvican] “John, nesting would work just requiring a bit of recalculating speed values but would take more time to render, especially with projects having many event FX/transitions and layer compositing, pre-rendering the “twixtored” footage would cut down rendering time I suppose.”

I agree. I thought it was just one event. If there are a lot, I would pre-render too.

~jr

http://www.johnrofrano.com
http://www.vasst.com

[Rick Anvican]

John, can you clarify three things for me please:

1) Nested projects would be used like if you want to apply like a single transition on a composited shot (e.g. made by multiple layers on different compositing modes), right? Kind of like an After Effects composition…

2) Has Vegas 12 resolved the green tinted bar that affects GPU rendered output on Mainconcept AVC 480p templates? I know this had something to do with width (854) not being divisible by 16 and would be fixed by using width divisible by 16. Any known reasons why it occurred?

3) I tend to think that giving Sony Vegas source video with non-standard resolution dimensions (e.g. 1258×544) would exhibit problems like garbled renders, would it be better to convert it into resolutions with one commonly used dimension (e.g. 720×310 or 1280×554)?

Thanks.

(Ever had the situation where you have HD AVC/MOV footage where the video won’t show up in the preview window on preview qualities Draft/Preview(any mode) but would show up with Good/Best(any mode)?)

[John Rofrano]

[Rick Anvican] “1) Nested projects would be used like if you want to apply like a single transition on a composited shot (e.g. made by multiple layers on different compositing modes), right? Kind of like an After Effects composition…”

Yes, exactly! Whenever you would use a Pre-Comp in After Effects, think Nested Project in Vegas Pro.

[Rick Anvican] “2) Has Vegas 12 resolved the green tinted bar that affects GPU rendered output on Mainconcept AVC 480p templates? I know this had something to do with width (854) not being divisible by 16 and would be fixed by using width divisible by 16. Any known reasons why it occurred?”

I’m not sure if I remember that problem but you need to render in a resolution that the codec can handle so this will never be “fixed” because the codec requires it.

[Rick Anvican] “3) I tend to think that giving Sony Vegas source video with non-standard resolution dimensions (e.g. 1258z544) would exhibit problems like garbled renders, would it be better to convert it into resolutions with one commonly used dimension (e.g. 720×310 or 1280×554)?”

If for some reason you need to deliver in a non-standard resolution, you need to be prepared to use a codec that supports nonstandard resolutions. If you plan to use a standard codec to render then you need to get your footage into a standard resolution.

It’s all about planning. You can’t just “do whatever you want”. If you plan to render with a codec that only supports resolutions in multiples of 16 then you should acquire all of you video in a resolution that is a multiple of 16. Otherwise you are going to need to convert your video after the fact.

~jr

http://www.johnrofrano.com
http://www.vasst.com

[Rick Anvican]

Many thanks for the clarifications, John.

Edit: I just had this in mind – if a video layer’s track motion doesn’t retain source resolution like event pan/crop, wouldn’t it mean that 3d source alpha would affect resolution too? Any method to do a 3D pan/crop like putting an event in 3d space yet still retaining original resolution when zooming in and such?