What's the correct pixel aspect of HuffYuv, mjpeg and flv?

[brassplyer]

Trying to convert a 4:3 flv file to avi eventually have it fit onto a DVD with other 16:9 material, also doing deshaking with Virtualdub. When deshaking with the Thalen deshaker and converting to HUFFYuv, what pixel aspect should I be using?

Eventually I want it to end up as DV avi since the app I use for DVD creation, Pinnacle Studio doesn't get along with Huffyuv.

[manono]

I don't guess your FLV is really 4:3 since that's a DAR. I suspect it's really 1.33:1 (or 1:1), some resolution such as 320x240, 480x360, 640x480, etc. At some point you'll want to prepare it to be encoded as 16:9, but you didn't say if you plan on adding black to the sides or crop from the top and bottom. If adding black to the sides, FitCD gives:

LanczosResize(544,480)
AddBorders(88,0,88,0)

Also, going from FLV to HuffYUV to DV-AVI to MPEG-2 seems to me to be a real waste of time, effort, and quality (a very little bit). If Pinnacle doesn't like lossless AVIs, use a different encoder and use Pinnacle only for the authoring if you have to. Unless you're doing some sort of editing of the FLV that can't be done in the script, I'd go directly from FLV to MPEG-2.

[brassplyer]

I don't guess your FLV is really 4:3 since that's a DAR. I suspect it's really 1.33:1 (or 1:1), some resolution such as 320x240, 480x360, 640x480, etc. At some point you'll want to prepare it to be encoded as 16:9, but you didn't say if you plan on adding black to the sides or crop from the top and bottom. If adding black to the sides, FitCD gives:

LanczosResize(544,480)
AddBorders(88,0,88,0)

Actually what I was referring to was the options Deshaker gives - Square pixels, .912 Standard NTSC, Anamorphic 1.215. Not sure which applies in which situation. And if it's possible I need to specify one for Deshake processing and a different one for output.

Also, going from FLV to HuffYUV to DV-AVI to MPEG-2 seems to me to be a real waste of time, effort, and quality (a very little bit). If Pinnacle doesn't like lossless AVIs, use a different encoder and use Pinnacle only for the authoring if you have to. Unless you're doing some sort of editing of the FLV that can't be done in the script, I'd go directly from FLV to MPEG-2.

Since I need to do both resizing and deshaking, there needs to be intermediate steps. And it seems HuffYUV isn't an option since it gives me weird output when I try to do the 16:9 pillarboxing and go to HuffYUV - the image comes out B&W, slanted and with weird framing.

The only way I've been able to make it work so far is Deshaking and going to uncompressed pillarboxed.

[manono]

Oh, the pixel aspect ratio for Deshaker? Sorry. As I mentioned, I expect your FLV is square pixel (when I said it's most likely 1.33:1 or 1:1). You haven't provided any information to confirm or disprove that, though. So, that's what you'd use in Deshaker. Since you're familiar with AviSynth (right?), I'm not sure why you're not using Depan (DepanStabilize) or even Stab if it'll do the job on light shaking. It'll save you from running 2 passes.

And it seems HuffYUV isn't an option since it gives me weird output when I try to do the 16:9 pillarboxing and go to HuffYUV - the image comes out B&W, slanted and with weird framing.

That's often an indication of messed up cropping and/or resizing. Perhaps if you posted the script (or the VDub settings?) that gave you that result, someone could help.