In the Proxy Media window I have NTSC DV Widescreen 24p (2-3 pulldown) and in the HD Media is 1080 60i intermediate. I see no "broadcast-quality" to turn off in Gearshift. In the target folder are two .avi files: the uncompressed file at 1.1gb (1440x1080) and the proxy file at 384mb but 720x480. Is there a way to compress yet keep the 1440x1080 ratio with Gearshift?

Kevin

There is something else wrong here. Try creating a different proxy.

When I used it, it worked just fine. On a 1,4Ghz laptop from Dell it worked smooth.

Maybe I have your workflow wrong here.... but this is what I think is wrong.

You generally use gearshift to create/work with low-res proxies of your master high-res HD files. It sounds to me like you are using Gearshift to generate high-res HD format proxies. Which rather defeats the purpose I think.

The correct workflow is as follows (I think - note:I don't use Gearshift (yet))

Create the low-res proxies (Gearshift from native to low-res proxies)
Work on your project.
Gearshift the project back to the native files
Render.

In the lower section of GearShift, leave it set to "none" so you're only creating the DV proxy. If you don't have a BMD card, or don't have a fasst RAID, then the YUV file is of no use for you. Just use the DV Proxy, and leave the other options set to "None"

Hey Spot,

No, I don't have a BMD card or RAID. In using the DV Proxy only it rendered in about 13 minutes. However, the 720x480 ratio disallows zoom/pan, doesn't it? With 1440x1080 I can zoom into part of the clip as long as I don't go under the 720x480 ratio and lose no resolution. I guess what I'm asking is, is there a way to get an editable 1440x1080 compressed file? Or will the zoom pixelate anyway in final render to mpeg dvd as it downsamples?

I am a novice to this as can be ascertained by these questions so be patient if I seem a little slow on the uptake.

Kevin
Fairbanks, Alaska

The DV Proxy is only to make the editing process smooth, so zoom in even though at DV resolution it may become pixellated. Before rendering the output version you will "Shift Gears" and replace the DV with the original m2t, so the same zoom will now take advantage of the HDV resolution and look great.

If you want to check how good it looks, do a quick "temporary" Shift Gears, and use Shift B to do a RAM render of that bit of the timeline. Enjoy, then Shift Gears back to DV to continue editing.

Kevin,

You may not have the same confusion I did when I first started HDV editing a few weeks back, but in case you do, this may help.

Since you cannot effectively directly edit the m2t files captured from an HDV camera, you must first create either a proxy or an intermediate.

The word "proxy" means stand-in or substitute. This is the way you want to go if you don't have a lot of disk space or if you are in a hurry. Vegas can generate these fairly quickly. When you are finished editing, but before you render, you substitute the original m2t files for your proxy and then render. Gearshift automates all of this for you.

By contrast, the intermediate files are high resolution, editable files. You can render directly from these and they are supposed to be in every way as good as the original m2t files, except that each frame is self-contained and does not rely on the frames that preceded it (the m2t files requires a dozen or more frames before the software can show you what the single frame under the cursor looks like, which is why it is so slow to edit). If you have the disk space and the horsepower, intermediate files are the way to go. I think Spot has some good introductory material at the VASST site that explains this, and he also has a book on the subject.

You can zoom in on either the proxy or the intermediate, although you'll be able to see more while editing using the intermediate, since you have all the pixels in that file. However, either approach should give you virtually the identical sharpness in the zoomed-in image (because, remember, with the proxy you substitute the original high-res m2t files just before rendering).

Here are a couple of threads I started on related issues about a month ago. Perhaps you will find them useful:

HDV Questions

My workflow for HDV to SD projects

"When you are finished editing, but before you render, you substitute the original m2t files for your proxy and then render. Gearshift automates all of this for you"

I've always editied HDV using Cineform intermediates so I haven't got involved in Gearshift magic. Discussions here have explained very well why we don't edit m2t files but this particular discussion (or the above, in particular) has caused me to wonder what happens in the automation process. I'd casually thought that Gearshift just cut the m2t files and applied effects as an automated copy of the editing that had been done on the proxies. Obviously not, for this would run into all the same problems of cutting long GOPs. VASST seems unavailable at the moment -- does DSE give a full explanation on that site. Just curious about the detail.

Hi Serena,

GearShift swaps the media source in the media pool. It never touches the timeline or FX or anything. It makes no changes to your project. All it does is redirect the media pool to a different source file when you Shift Gears. (or course it helps with all the rendering stuff up front too)

~jr

Yes, the m2t gets cut. I don't think the problem lies in cutting the m2t during the render. The problem is that Vegas isn't well-equipped to display each frame accurately during the editing process as you scrub back and forth. It COULD be designed that way, it just isn't. Thus, the proxy technique works to overcome this limitation, and then during the render time, Vegas does the right thing because the render engine is designed differently than the editing engine.

At least that's my understanding. I could easily be proved wrong on this one.

Thanks JR. Guess my query was rather poorly framed. Of course if we put m2t on the timeline Vegas only inserts pointers to the source files (not actual cuts) and the same for intermediates and proxies. This gave rise to my "how does that work?" thought bubble. If Gearshift now redirects the pointers from the proxies to the source material, why does that produce better results than editing the m2t GOPS directly (putting aside the issues of why one uses proxies). I think this is probably really something I don't need to understand!

Serena

It's not that it produces a better quality render, it's that it makes editing easier. It's pretty much impossible to edit with m2t's because Vegas can't play them smoothly directly. How can you make fine edits when your preview is lurching 4 or 5 frames at a time with your CPU maxed? The proxies are just there to make your job easier during editing, not to improve final render quality.

Laurence, there is rather more to it than that. In m2t files not every frame carries all of the information needed (sony HDV employs 15 frame GOPS). So that it runs at 4fps isn't the reason for not cutting m2t (although it's a good one). Of course, given enough computing power an editor could decode and recode on the fly -- money fixes all these things.
Serena

Vegas does the extra buffering during rendering so that that is not an issue. There's only a fully drawn frame something like every 12th or 15th frame, but in actuallity you can cut anywhere. Even native editors like Womble or Premier Pro can cut anywhere. They screw up the GOP sequence in the process and this affects things like copying the edited M2T file back to tape but it does work.

In the case of Vegas using Gearshift and proxies, the full project is rerendered to an M2T file with a proper GOP sequence that can be written back to tape. Cuts can be anywhere, even on frames that aren't fully defined thanks to some kind of frame buffering that seems to work well. It looks fine. I wouldn't go multiple generations this way, but you can do it all in one generation. Rendering to Cineform is also an option if you're going to do further assembly.

Now I'm ending up with narrow side bars after final render from two different workflows.

1st: Projects setting to 1440x1080 24f; 1:46 sec. m2t to Cineform HD .avi (1440x1080, no Gearshift). Video fills preview window until I click to final render to mpeg2 NTSC Wide (24f 2-3 pulldown) and the picture adjusts and side bars appear. Render it out, bring it up in ADVD3, burn it, view it, and they're there.

2nd: Projects setting to 1440x1080 24f; 1:46 sec. m2t to NTSC DV Wide (24f 2-3 pulldown, via Gearshift). The bars appear so I p/c to 700 width on all events and problem solved, no bars on preview. I Gearshift back to m2t and click to final render to mpeg2 NTSC Wide (24f 2-3 pulldown), the picture adjusts and side bars appear again. Render it out, bring it up in ADVD3, burn it, view it, and they're still there.

John, I've read and reread your posts about the 1404.6 width on HDV events but I haven't been able to remove the bars. I'm not sure when in your workflow you're applying this and I'm obviously not doing something right. This has been so frustrating. And to top it off, when I played the DVD of a raven sitting on top of a dumpster in -45 ice fog, there was deinterlacing (I think that's the term) everytime the camera moved. Narrow horizontal lines jittered across the screen. In fact, the zoomed portion actually looked the best.

Any help is appreciated,
Kevin