Know that's a pain but I guess you'll soon remember that each time you have to set the parameters! Does this matter when you're generating proxies? I'm not clear whether your final product is SD or HD.

Yes this happens when I'm generating proxies. The final distribution is SD but I'll be making a HDV backup for myself. The deinterlace method tab needs to be set for both the proxy render and the final SD mpeg 2 render for DVD.

What happens is that if you set the deinterlace method to either blend fields or interpolate, Vegas will realize that the video is interlaced and resize the even and odd fields separately before refolding them back together on alternate lines. If however you set the deinterlace method to none, Vegas will resize the video as if it was a single progressive image. The result is that the combing interlace artifacts on motion are resized and look awful. None of this has anything to do with deinterlacing, but that's how it works. You would think that Vegas would just look at the project properties and see if it was set to interlaced or progressive and resize according to that setting, but I can tell you from experience that that is not how it works. Instead, Vegas looks at how the deinterlace method is set in order to determine how to resize. This confused the heck out of me at first because I thought that selecting a deinterlace method before resizing would deinterlace the video then resize it. This is not the case however.

For anyone who doubts this, try the following experiment. Select "none" as your deinterlace method, then render HDV as SD and view the resulting file through a 1394 box like a Canopus ADVC-100. If you pause the video in the middle of motion or a fast pan, you'll see a static image with ugly wavy vertical lines.

Now go back to the original HDV, select either "interpolate" or "blend fields" as your deinterlace method, render to SD again and look at the video on a 1394 monitor again. This time you'll see that the downrezzed SD video is still properly interlaced. Pause in the middle of motion or a fast pan and you'll see the image alternating at 60 hertz between two images a sixtieth of a second apart with no wavy vertical lines whatsoever. In other words, the video was properly separated into fields before resizing and then the fields were folded back together into a 60i image that looks great on a standard 60 NTSC TV set.

Laurence, you might have hit on something else that's been bugging me for ages.
Then again don't remember if you mentioned if you're rendering at Best which you should if you're rescaling.
But, last night i was animating a logo and as I was going to do a lot of other work with it down the track I rendered to uncompressed HD and the result when it was scaled down to SD was woeful, hm, makes me wonder what was going on.

Bob

Yeah I render at best when rescaling, even with proxies. Gearshift defauts at good so you need to make your own SD template to do this and name it with a GSP prefix so that it shows up in Gearshift.

Selecting a deinterlace method (it doesn't matter which one as it doesn't actually deinterlace) is really important any time you resize interlaced video. This is something you need to do every time you you downrez HDV to SD.

This doesn't address your question. I just wanted to share an Avisynth script I use for converting HDV to SD (interlaced). I was not satisfied with the interlacing artifacts I saw when using Vegas to do this directly. I experimented for two days doing this every conceivable wrong way before getting it to look correct.

--------------------------------------------------------
# Note: rename file to 'foobar.avs' for use

# This avisynth script changes HDV 1080i into SD 480i with better
# quality than a direct conversion in Vegas 6d.
# Script written by John Beale 3/27/2006 www.bealecorner.com

# Script requires Avisynth 2.5 and 'reversefielddominance' plugin from
# http://www.geocities.com/siwalters_uk/reversefielddominance.html


AVISource("C:\Video\HDV.avi") # frameserved HDV, 1440x1080i RGB24 mode
SeparateFields() # work on fields separately
BilinearResize(720,240) # convert 1440x540 HDV fields to 720x240 SD fields
Weave() # rejoin fields to form interlaced 720x480 SD frame
ReverseFieldDominance() # HDV (upper field 1st) -> SD (lower field 1st)

# -------- optional smoothing & sharpening -------------

# Blur(Horizontal, Vertical): negative value sharpens
Blur(-0.75,0.25) # slight blending of fields (vert. blur) looks better
Blur(-.5,0.0) # yet more sharpening horizontally

# -------- optional noise reduction --------------------

ConvertToYUY2() # smoothing filters don't use RGB formats
FluxSmoothT(6) # reduce noise on smooth surfaces for MP2 encode
ConvertToRGB24() # MP2 encoders prefer RGB formats

# get MPEG2 with CCE Basic with Adv Video Setting->Luminance 16-235

I used to do something similar with VirtualDub and a chain of filters. Then I found out that once I enabled a deinterlace method, the results straight from Vegas looked just as good. Your script is doing the same thing as my VDub filter chain was: separating the interlaced fields before resizing. Vegas does this too, but only when you choose a deinterlace method.