Using HDLink to import 1080i/60 from my XH-A1s, I'm able to start with relatively good looking video on my CS4 time-line. (Gotta love FirstLight when recordings are less than stellar!)

While following various threads on the topic of HD-to-SD-DVD, I decided to experiment with the various combinations of TMPGenc Xpress (trial), VirtualDub, Debugmode Frameserver as described by Jon Geddes and a couple of others. (Environment is Vista64/i7 with ATI Radeon 4870, 6GB RAM and several TBs of disc)

A 70-minute CS4 timeline took nearly 16 hrs to render using the Frameserver/VirtualDub scenario and the net results are pretty good, but it seems that using VirtualDub to deinterlace and resize a Cineform AVI (using Cineform codec) out of AME is as good or better than the long process.

Also, if I use TMPGenc's Authoring Works 4 to create the SD-DVD without using TMPGenc Xpress to create the mpg vs. using the AVI rendered from VirtualDub with the Cineform 1440X1080 AVI as the source for VirtualDub, I cannot visually detect any difference between the end-products....such as authoring from mpg out of Xpress or straightaway from the Vdub AVI.

Can anyone explain if or why I should I move the 720X480 AVI to "mpg" via TMPGenc Xpress before authoring or is that unnecessary, perhaps because the TMPGenc Authoring program does that on the fly?

My senior-citizen brain doesn't seem to be able to keep all the nuances together, so thanks in advance to anyone who can/will help simply/clarify what I want to do: get the best possible SD-DVD when using CS4/Cineform to handle the capture/editing from the Canon 1080i/60 source.

Dave Nuttall
San Antonio, TX

Hi Dave,

I'm not familiar with TMPG's Authoring Works, but you'll find a lot of users of their Enc 4.0 Express around here.

I was excited to hear that Frameserver was now functional with CS4, but the render times pretty much make it a moot point for my workflow.

I render a master CFHD file, take that into TMPGEnc. Encoding aside.... it does a superior job of scaling and deinterlacing (if that's part of your formula) than the hapless AME.

You can use its presets to output a very nice MPEG2 and AC3 set. Data rates are easily set if you need anything less than 8000 kb/s.

Then take those files into your authoring program.

This workflow not only achieves the best looking DVD ingredients (to my eye), but it does so relatively quickly.... and you have a resultant CFHD master output file for your archive.

I was excited to hear that Frameserver was now functional with CS4, but the render times pretty much make it a moot point for my workflow.
Yes, at less than 3fps (avg 2.5 as I recall) it makes a big deal out of what seems to be not such a big problem when you throw VirtualDub or TMPGenc Xpress at the CFHD output from AME.

I render a master CFHD file, take that into TMPGEnc. Encoding aside.... it does a superior job of scaling and deinterlacing (if that's part of your formula) than the hapless AME.

I got similar results from the trial version of TMPGenc but was surprised by the visual equivalency of using VirtualDub with TMPGenc authoring. I guess I'll go ahead and purchase the TMPGenc Xpress instead of tweaking the VirtualDub parameters!

This workflow not only achieves the best looking DVD ingredients (to my eye), but it does so relatively quickly.... and you have a resultant CFHD master output file for your archive.

I'm seeing approximately 1GB/minute of run-time, whether raw input from the camera or AME output using the Cineform codec. Does that track with your experience, Marty?

double post removed.