deinterlacing - before or after editing?

[jrockford1974]

I shoot with a Canon HV-10, capture with hdvsplit (to m2t files), and edit and render with Vegas Pro 8. Almost all video I take is high-action sports. Editing is limited to fade transitions, and an occasional speed change (slow down, speed up) of some clips on the timeline.

Question: When I want the final product to be deinterlaced, is it best to deinterlace before bringing into Vegas, or deinterlace as part of the rendering process in Vegas?

I get bogged down in the "it's just so easy to do it all in vegas" vs "I'm sure some external, more complex approach will result in the best final product". So, I've got a pro/con list:

deinterlace before editing in vegas:
PROS:
I get to pick my deinterlacer.
It just seems like editing and rendering video that has already been deinterlaced is "just better" (my baseless assumption)
CONS:
Yet another workflow step
Frameserving (at least with the free or open source options) seems complex
Converting to deinterlaced files (like m2t to avi) takes a ton of time and diskspace
deinterlace during vegas render
PROS:
Easy. Doesn't require any additional time, complexity, or diskspace
I paid for Vegas, so I might as well use everything that comes with it.
CONS:
Vegas stock deinterlacer likely not as good as many of the open source, free, or commercial deinterlacers on the net.


Any experiential, anecdotal, or factual feedback on my question is greatly appreciated. Maybe it will help others, and I'm sure there's no single, holy-grail answer. Thanks in advance.

[jagabo]

AviSynth's TempGaussMC_beta() is the best bob/deinterlacer I've seen. It's not hard to use once you've learned a few AviSynth commands. But it's very slow.

import("\path_to\TempGaussMC_beta.avs")
WhateverSource("filename.ext")
TempGaussMC_beta() #bob to double frame rate ie, 30i -> 60p
SelectEven() #throw out half the frames to restore the original frame rate

See the link at the end of this post for some deinterlacing examples:
https://forum.videohelp.com/topic365493-30.html#1949705

[poisondeathray]

^ I completely agree with jagabo's comments, and would like to add a few:

-If quality is important, you will not use Vegas' deinterlacer (either blend or interpolate), as it is not even average, it is well below in terms of quality. There are no settings or adjustments either (e.g. no bob mode which would be ideal for sports footage like jrockford is using)

-Frameserving out is easy with debugmode frameserver, a great way to avoid using Vegas' crappy h.264 encoder (e.g if you were encoding to a flash format)

-To get the best results, you should deinterlace before as a general rule (this doesn't apply solely to Vegas but to other editors/compositing tools/effects programs as well) , because any effects, overlays, color keying, etc.. are generally not as precise on interlaced foootage and the edges can get distortions if you deinterlaced later in the workflow.

[lordsmurf]

Deinterlace in VirtualDub or AVISynth beforehand, export to uncompressed AVI.

Import that into Vegas, have at it.

[zoobie]

seems like going from m2t to avi, editing, then back to m2t to burn would be 2 extra and unnecessary re-encodes...
doesn't every re-encode lose some quality?

[lordsmurf]

No, uncompressed encodes won't visibly lose anything -- only in theory. Even pro studios do it.

Some of the hobby crowd gets a little "too pure" for their own good sometimes.