VirtualDub and Framserver help

Paul_Holmes wrote on 4/16/2003, 10:43 PM
I've installed Satish's Frameserver and have been serving to virtual dub. I would like to try temporal filtering as mentioned by John Beale in one of his posts.

So far so good. I open the frame-served file and apply the temporal filter. I am then able to preview it as it is frame-served but can't figure out how to create a file that I can then use again in Vegas. I tried saving as AVI uncompressed and as AVI Intel . . something codec. With the uncompressed when I tried to use it in Vegas it came up with some message about DFSC not serving or something. I can get it on the timeline but then I have to rerender it back to NTSC DV.

So the question is, anyone know how to frameserve a file to Virtual Dub and get an avi that is ready for use in Vegas?


satish wrote on 4/16/2003, 11:58 PM
If you are okay with rendering a fully uncompressed AVI in Virtualdub and importing into Vegas, everything should be okay. The problem comes only when you want to frameserve again from VDub to Vegas, which can only be done if you use Virtualdub's frameserver to serve to another instance of Vegas.
DDogg wrote on 4/17/2003, 2:19 PM
For instructions on how to Serve from VDub TO vegas see:

This would allow you to (using Satish's FS) to serve to VDub and then using above to Serve back to another instance of Vegas.

Frankly I think you would be much better off just opening Satish's signpost file in Avisynth and using one of the multitude of noise filters available although that is a little more complicated and requires simple scripts to accomplish. That may be something you want to investigate after becoming more familiar with this whole frameserving "thing" :-)

Paul, I would be happy to assist if I can be of service.

Paul_Holmes wrote on 4/17/2003, 5:56 PM
Thanks, DDogg. I was looking at Avisynth last night on the web (in fact I may have it downloaded and ready to run). I appreciate the offer of help. I'll read up on it a little. It seems like the only thing I can get out of Virtual Dub is an uncompressed file. If Avisynth allows me to render to an NTSC DV file that can then be used directly in Vegas without re-rendering, it would be the ticket.
BJ_M wrote on 4/17/2003, 8:22 PM
in v-dub click the video tab and then "compression" .. select your compresion method .

i suggest you install and use either the main concept DV codec or canopus DV codec or Panasonic DV codec (if space is really limited) or install and use the excellent huffyuv lossless codec as your 2 choices of compression method ..

THEN after selecting the compression method - file -> "save as avi"

DDogg wrote on 4/17/2003, 8:43 PM
"...allows me to render to an NTSC DV file that can then be used directly in Vegas without re-rendering..."

What I was suggesting was (in one operation) serving from Vegas to VDub and then at the same time serve from VDUB to another instance of Vegas. No need to actually render until you output a finished file.

Paul, but if you insist :), as said you can compress to whatever format you want in VDub and then load that file into Vegas (why?)

Actually and just for the heck of it I just FS'd from Vegas to VDub back to another instance of Vegas and then from Vegas to another instance of VDub. Pretty crazy. Fun is where you find it when you get old I guess :)
Paul_Holmes wrote on 4/17/2003, 9:27 PM
DDog, I need your help on this. Can serve from Vegas to VDub, then I start frameserving in VDub and the 2nd instance says it doesn't recognize the file format (I named the VDub served file "VDubtoVegas.avi." Is it because I need to set certain parameters in VDub, telling it what kind of file to serve. I'll keep going over help, but I'm new to all this so I've probably got settings or codecs or something wrong. As far as serving to VDub and applying filters, then previewing them, I don't have a problem.
DDogg wrote on 4/17/2003, 9:37 PM
FS'ing from VDub to Vegas is a little more complicated than the gift Satish gave us. Cough, I guess you didn't take a look at the link I put above? :-) There are explicit instructions for VDub to Vegas using VFAPI and READAVS.dll with links and installation instructions. It is easy but takes a few minutes and some concentration for readavs.dll (allows reading and wrapping of the VDub vdr signpost file).

Paul_Holmes wrote on 4/17/2003, 9:46 PM
Whoops! Overlooked your very explicit earlier post. Will go there immediately!
Paul_Holmes wrote on 4/17/2003, 10:24 PM
Well, I'm trying my first experiment after following your instructions -- Vegas frameserved to VDub frameserved to another Vegas, and am rendering the 2nd Vegas file as an avi. Then I can compare them side by side and see the difference the temporal filter makes at 60%. Hopefully I've got it all right. Thanks for the help, DDog!
DDogg wrote on 4/18/2003, 8:47 AM
Paul, now that you have VFAPI and readavs installed you can do a few other things. As well as using VDub, you might also want to "play" with:

1> Load the signpost.avi in TMPG. Do advanced settings for noise etc.. Save as a tpr project file. Wrap the tpr with VFAPI (TMPG does not need to be running). Serve to where ever.
2> Use simple avisynth script (use AviSource command) to open signpost.avi. Noise filter, resize, etc.. Wrap AVS file with VFAPI. Serve...

With either of the above or VDub you can serve multiple timelines to another instance of Vegas. By "muting" the individual timelines one can do a frame accurate comparison between various served video. This allows a good comparison of different settings, scripts, noise filters, etc. Also, remember you can "play" the signposts in WMP to get a feel for how you have effected the video with your settings, filters, scripts, etc.
mikkie wrote on 4/18/2003, 8:59 AM

If the reason for using V/Dub is to use it's noise filters...

I've found it usually more efficient to get the files in V/Dub first, apply any filters, save the filter/VDub settings, do any cropping/resizing (sometimes DI or IVT), then render to avi using whatever codec (there's a few high-qual lossless codecs out there).

V/Dub is much faster due to it's nature when cropping & resizing, and any filters in most any app can be applied more or less depending on the scene characteristics. Reviewing the processed video in Vegas, mark scenes as nec. to be tweaked, then re process those scenes from the orig source in V/Dub, importing the fixes into Vegas before final render.

The point of all that anyway is that the amount of noise filtering you can use varies quite a bit depending on the scene characteristics, and I've rarely found one setting that works all the way through. It's not such a big deal using a filter that works the same in every case like the std. deinterlace, but when you have something that bases it's filtering on the amount of in frame and frame to frame motion, different backgrounds and light levels really have a very big effect.

With the temporal & similar filters for example, stronger settings that work on outdoor scenes with well defined background subject matter, will likely cause some jerkiness of motion in dark scenes with poorly defined backgrounds, as the software thinks something like the shadows of moving facial characteristics are unwanted artifacts. With actual rendered files it's not a big deal to tweak the filtering in such scenes using V/Dub & the source footage, then use the results to patch over the same scenes in Vegas.
DDogg wrote on 4/18/2003, 12:03 PM
That's a good point but I think I see a different method ...Hmm, what might be interesting is to open the original source in a second instance of Vegas and then from the first instance of Vegas serve to VDub and serve from VDub a second time-line in the second instance of Vegas. Now you could slice, dice, mix, and match portions of the original with portions of the filtered time-line for final render .... Hmmm, off to play around with that.
Paul_Holmes wrote on 4/18/2003, 1:01 PM
Great suggestions. One great thing about VDub is I can serve back to Vegas, then see a preview on the computer (low frame-rate of course) and get a sense of what the filter is doing. As Mikkie says, something like the temporal filter isn't perfect. Using 60% it did a little too much smoothing on highlighted hair but overall it worked real nice in fast-moving scenes smoothing the grass and not the runners. I'll experiment some more at say 40% and see if it's acceptable. The whole reason for using the temporal, in my case, as in John Beale's would be to be able to lower the bitrate going to DVD in cases where you wanted an hour and a 1/2 to 2 hours.

Just experimenting now, and will have to see what some of the other filters do.

To be honest I'm so jazzed about what I can do already in Vegas with color-correction and things like quick-blur, I may never find real use for VDub, but after playing around for a while, may find something worthwhile.