Best workflow to work with Hdv footage

Gaara

New member
Hi everybody!

I've been looking several post about the famous 60i to 24p... about how to do the deinterlace with this or that software, the timelines at 23.967, etc...

but I still have one doubt... how it's your workflow with your footage???.. I mean... do you capture the video, then edit, after that it comes the work with effects and then you do the deinterlace???... or there's a recommended workflow to do that???

I work with avid 5.1 and shake 2.5.

I hope someone can help me... thank you very much :happy:
 
Yeah I wanna know this too. Also is it possible to convert 1080i z1u footage into good-looking 720/24p?
 
If you're going to use an adaptive de-interlacer to get 60i -> 24p conversion, you must recognize that proper conversion *the Magic Bullet way* has absolutely enormous render times. You are best off doing a very good edit at 60i to get rid of all the footage you won't need as 24p.

From there, convert ONLY the video that will be in the final render to 24p (render to a lossless codec/uncompressed), and make a frame-accurate 24p edit.

Depending on your post-production/special effects requirements, you may be better off doing stuff like keying before or after the 24p conversion - you'll have to experiment to find out, and it may vary between shots.

The result of a proper adaptive deinterlacer should be 1080p24. From here, you can downsample to whatever resolution you want - including 720p24 and 480p24.

-Spiff
 
So, because of the "extreme" long time that takes to do the convertion, the proper way to do a de-interlace should be:

- Capture the video
- Edit
- Apply effects (keying, text, glow, blur... stuff like that)
- Final cut (I mean.. to do the final edition, not the program :P)
- De-interlace footage to output whatever media you need.

Thank you very much for the answer :thumbsup:
 
In the Articles section there is an excellent post by Jarred Land entitled "Converting 60i to 24p. A Real World Case Study." Additionally, there is a real gem listed in the comments below the article - (keep scrolling down) posted by Loi Banh using Apple's Compressor to do the deinterlacing and converting to 24p. I completely agree with his comments and my experience has shown the same thing.

As to the converting itself - you want to convert the smallest amount of video possible because of the render times. For example, the other day I submitted 2 clips to convert. One was 1 min 27 sec, the other was 1 min 32 sec. I was using a dual 1.42 G4 with 2 Gig ram and it took 31 HOURS! (I am waiting for Apple's announcment next week and will then be ordering a new computer to start my own cluster!)

The render time depends on the overall quality that you want. I needed very good quality for those clips above. Obviously, the better the quality, the more time it takes. I have found that the quality is not noticeable in sections where there is little movement, so I will often break the final edit into sections to be processed, trying to keep the render time to a minimum.

Good luck,
Dean
 
You've got it almost all correct! Except I just want to clarify this order:

- Apply effects (keying, text, glow, blur... stuff like that)
- Final cut (I mean.. to do the final edition, not the program :P)
- De-interlace footage to output whatever media you need

Very often effects work on interlaced footage (keying, blurring etc) is done on a deinterlaced image. Often the final render from this ends up only 30p, which makes it impossible to again "de-interlace" to 24p. Proper 24p motion estimation REQUIRES the source to be 60i - otherwise you'll be SOL. As such, you either have to do your effects work at 60 fps - which is very time consuming.

I would suggest you treat effects shots on a shot-by-shot basis. Render them out to 24p and see if the effect can be done effectively at this resolution, if not, do the effect at 60p - which will be a lot more work.

Also, I would suggest you do your final cut on 24p-converted material, since cuts on a 60i timeline may not produce exactly what you want in 24p.

-Spiff
 
Gaara,

You should upgrade to Shake 4. For one thing, the QuickTime support was totally overhauled for version 3.5. Version 4 has new deinterlacing, retiming, morphing, and you can generate Shake scripts out of FCP 5 that automatically bring in the files with edit points.

ACStudios mentioned Compressor; I believe the latest version has some of the same retiming and deinterlacing logic used in Shake.

Unless you're on Windows, in which case you're stuck at version 2.5.
 
Back
Top