My delivery is mainly to computer LCD screens. Would you leave as interlaced and let the monitor work it out?

Mike

My delivery is mainly to computer LCD screens. Would you leave as interlaced and let the monitor work it out? Depends on how well & what mode the program uses to de-interlace. Personnally I would never de-interlace the video.
My hardware video card ATI X1650 has HD hardware accelleration & hardware de-interlacing. Works perfect for computer displays. De-interlacing HDV can really make it look terrible (my opinion) whening panning & motion.
If you have to de-interlace for strictly computer viewing only there are a few good de-interlacers out there.

...but not the Cineform or Vegas ones....... you are saying, right?

duplicate deleted

For mostly static shots I find it generally ok to leave Vegas to take care of deinterlacing implicitly. In settings I leave deinterlace method as "Blend Fields" and that works fine most of the time. As far as I can tell (having done web research and tests) Vegas' deinterlacing is automatically invoked whenever needed e.g. only when footage is Interlaced (obviously) and Project or Render settings are Progressive or else when doing doing Pan&Scan etc. (e.g. to zoom into the HD). Also for certain FX I believe.

If I want better results then I pre-process the media using a motion-compensated (MoComp) deinterlacer like AlparySoft DeInterlacer (VirtualDub Plugin) or TomsMoComp (AviSynth plugin), in either case saving the result to CineForm. You need to convert from m2t to mpg first (e.g. using the free Mpeg StreamClip), and use VirtualDubMod rather than VirtualDub, or if you use AviSynth then it gets complicated (but possible) to input the mpg. Also when saving from these tools, might have to explicitly tell the Codec it's (now) progresive or else may have to tell Vegas explicityl (via media Properties) that it's progressive.

As you may gather, this kind of process is fiddly and only worth it if the scene contains significant movement (when blending fields leaves edge ghosts). For example, for static tripod shots of lectures and interviews I did not see any advantge. On the other hand for fast moving things like aircraft or the scene itself (shaky camera) the advantage of the MoComp approach becomes clear.

Most editors (not just Vegas) only have simple deinterlacing like Blend Fields, presumably because of the CPU load involved. Don't see why in principle at least MoComp shouldn't be a render option (for the patient) though. There is also in principle the Intelligent Deinterlacing technique but all that does is intelligently switch between Blend Fields and Interpolate - not in the same league as MoComp.

Well the CCDs on a lot of HDV cameras are 960x540. The cameras capture this sixty times a second and the captures alternate as your even and odd fields. Because of this, what I like to do is make a 960x540 master where one complete field is dropped and use this to render any computer viewable video that I need. I used to do this uncompressed, but lately I've been using the $28 http://www.pegasusimaging.com/picvideomjpeg.htmPegasys MJPEG codec[/link] at a really high quality setting. It looks exactly the same as uncompressed and is easily encoded by convertors for DivX, Quicktime, Flash, etc.

I noticed that iMovie is starting to use the same 960x540 dimensions. Somebody at Apple must have been reading my earlier posts ;-)

I looked at AlparySoft Deinterlace and see it is not for Vegas but a plug-in for VirtualDub. I capture with Cineform Neo, so how do I integrate these three together? An outline workflow would be appreciated and I'll try this.

Mike