I own an HV20, so I know the answer to some of your questions. Recently this camcorder was voted as the camcorder of the year btw. It literally sells like hotcakes.

So, HV20's 24p is not real 24p, because the HDV standard does not have support for it. So, it saves the 24p in a 60i stream, using this sequence: 18 progressive frames, 12 interlaced (3+2 in succession). When you put the 12 interlaced frames together by de-interlacing, they become 6. So at the end you get 18+6=24 truly progressive frames.

Unfortunately, Vegas does not do this job for you though. Neither VMS, or the full version of Vegas. Except After Effects I think, no NLE supports the HV20's kind of 24p so far (Premiere is buggy with it too). There are capturing programs that do so in one go (capture from tape and pulldown removal in one go), like NeoHDV, which unfortunately sells for $250. For those who don't have the money to do that, I have written a tutorial on how to go about it using freeware tools:
http://eugenia.gnomefiles.org/2007/07/13/canon-hv20-24p-pulldown/

There is also a new freeware Vegas plugin that saves more time (currently in beta), but it requires Vegas Pro to work, not VMS. The original method transcribed in my tutorial will work with VMS.

>I've been making hd-wmv videos using a framerate of 59.940

I am usually using 23.976 or 29.997 when I export...
http://eugenia.gnomefiles.org/2007/08/11/from-dvhdvavchd-to-ps3xbox360appletv/

Correct me if I'm wrong but I thought that the HV20 recorded true 23.xxx (24p frame based) and to get it into the HDV format spec on the tape they add 3+2 frames to equal 29.97 fps, 3 frames in the beginning & 2 at the end (I forgot the exact sequence). The same as converting film to 29.97fps ntsc.

Then when you render out you do the reverse of this to extract the true framerate and frames, dropping those extra frames that were added to fit the video into the HDV wrapper.

Nope, it's not like this. If you cut the last 3 frames and the next 2 from the next second, then the overall video wouldn't be smooth, it would be jerky because 5 sequential frames would be missing. The HV20 records only 18 progressive frames, and the rest 6 come from de-interlacing the rest 12... So it's recorded like this in the tape:
3 progressive + 2 interlaced + 3 + 2 + 3 + 2 +3 +2 +3 +2 +3 +2 = 30 fps (actually, 29.997)
Then, you use various utilities (ranging from $0 to $800) to keep in tact the 3 progressive frames, and blend the 2 interlaced and transform them to a single progressive one. This will result at the end of having a video at 23.976 fps, after the required audio delay modification too.

Vegas and VMS don't support any of this at all. You must use external utilities to remove pulldown and save to a lossless intermediate format (you see, you re-encode the .m2t stuff so you must use a lossless format to do the re-encoding), and then import these true 24p newly created clips on Vegas/VMS to do the rest of the editing as you normally would. That's how everyone in the HV20.com community does it atm.

One thing to remember is that the HV20 does not have "attributes" in its stream to tell the capture application which frames are progressive and which ones are interlaced, so the application must really do the extra work of analyzing the footage and decide to keep some frames as is, and some to de-interlace. I called Canon a few months back, they don't have plans to upgrade the firmware to include attributes, so when a capture app decides to support the HV20, they will have to be careful with the audio delay and the frame analyzing.

Though I don't film in HD, I've been reading this thread for sake of the future. You are talking about a Canon HD camera that is not so well supported by the Vegas software. Could it be that a Sony HD camera will be better supported?

Thanks for the answer/clarification.

It's the HDV standard's support to blame, not Vegas'. Manufacturers have to play tricks to get a 24p stream on a 60i one, otherwise apps can't read from the tape.

As for the Canon HV20, it is supported fine by Vegas. Only the pulldown removal is not. You only need to do pulldown removal if your are shooting in 24p mode (25p/50i/60i modes also available and are don't need pulldown removal), and if your exporting video is not meant for the TV. If all your videos are meant to be viewed back on a TV, you don't need to do a pulldown removal at all.

But if you want to export and playback your 24p footage on devices and computer screen, then you better do the pulldown removal using some of the available ways, because otherwise it will look like this:
http://www.flickr.com/photos/missilemike/494788313/
instead of like this:
http://www.flickr.com/photos/missilemike/494788323/

Ivan, I don't think there is a Sony consumer camcorder right now that records in 24p like the HV20 does... This is why the HV20 sells like hotcakes -- plus it has great picture quality.The HV20 broke ground with this camcorder and Canon stroke gold.

Eugenia,
I'm a relative novice in this area. I have an HV10. Do you have a simple explanation of the benefits of the HV20 and the benefits of your inverse telecine please ?

While the two camcorders use the same sensor, the HV20 has less noise (which means that either the lens is better, or the HV10 had a software bug or unoptimization). The HV20 is more expensive than the HV10, but it has a huge community already and the following additions, copied by another HV20 user forum:

1. HDMI out

http://www.hdforindies.com/2007/02/canon-hv20-as-crash-cam.html

and one more:
6. CINEMODE with 1/48th shutter speed control and better dynamic range.

As for the inverse telecine/pulldown removal, it's only needed if you shoot in 24p mode. If yes, then you must remove it ONLY if your videos are going to be viewed anywhere but via the HV20 itself. If you don't remove it, and you leave the 24p inside the 60i stream, then it looks really blurry in fast motion scenes.

Answering my own original question, under Vegas, the best option to de-interlace is "interpolate", not the default option "blend fields":
http://eugenia.gnomefiles.org/images/interlaced.png

It depends on the material which de-interlace method will give the best results.

As the manual says, when using high-detail, low-motion material then Blending information from each field will look best, but when using low-detail, high-motion material the Interpolating the "missing" lines using just one field from each frame will look best. When using material that isn't characterised as low-motion, high-detail, or high-motion, low-detail, then you will need to experiment to see which method suits your particular material best.

Both Blend and Interpolate are very simpleminded methods of de-interlacing. Better results can be obtained using one of the many smart de-interlacing algorithms available with VirtualDub, for example.One of the best of these, Mike Crash's Smart De-interlacer, can also be used as plugin in the full version of Vegas. I'm not sure of this is possible in VMS.

Mark

>It depends on the material

Well, usually a material has both kinds of footage, and you can't use de-interlacing selectively.

>Mike Crash's Smart De-interlacer

I will give it a try.

ok, that "Smart De-interlacer" filter is REALLY BAD.

No matter which options I select, and no matter which combinations I select, it performs really bad on fast motion footage. And when I kinda fix it for the fast motion, it pukes on normal footage...

Overall, Vegas' own "Interpolate" option seems to be the best overall deal.

I don't like de-interlacing anything if possilbe.
Go with an interlaced format and let the player de-interlace the video IF for computer viewing.
I can't believe how well WMP11 de-interlaces my hd-mpeg2, it's excellent.
Matter of fact I don't think it's de-interlacing it, I think internally maybe it's playing it back at 59.94 treating each field as a frame.

Thing is, I want all my viewers to be able to playback my footage, not just Windows users. This is why I always export in h.264. And thing is, except the indeed very good de-interlacing that WMP does, no other media player does it so. So, it is up to me to deliver a deinterlaced video to my viewers. More over, youtube, revver and other sites don't de-interlace for you, so if you upload there -- like I do-- then the result is really bad.

In other words, while WMP does excellent de-interlacing, you should not leave that to the user, but it should be taken care by the videographer.