Cineform intermediary or Gearshift?

Laurence wrote on 9/23/2005, 7:41 AM

I just bought the Gearshift script yesterday to use with my HVR-A1 footage. I had been planning on using the A1 in SD mode until I saw what it looks like in HD! Anyway, having seen it, I'm well aware that HD is worth the extra effort.

Which is the better approach to HD: the Cineform codec or just using Gearshift? Right now I'm leaning towards Gearshift, but I'd like to know what people who've used both approaches think.

Comments

VOGuy wrote on 9/23/2005, 8:29 AM

Like most issues of this sort, the correct answer is "It depends".

Using the Cineform intermediates you have the advantage of seeing your video in "high-def" as you edit. You need a pretty fast machine, and the Cineform intermediate files are quite a bit larger then the original m2t files. Cineform claims minimal generation loss, but you are going down one additional generation in the process. If you have a long project, you should also consider that previewing your work will also take longer, because you are working with much larger streams of data.

Using Gearshift can give you several advantages, because you are using "proxy" files, essentially as workprints, to do your work. You won't see your project in full HD while you work, but you won't be waiting around for your machine to finish up operations either. Also, even with the proxy files, video data storage is much more manageable. Also, when you render your project, you will be saving a generation, because you're rendering directly from the original m2t files.

Because I like to see the pretty pictures as I work, I use the Cineform intermediate system. However, on my one HD project this year (
) , I'm the customer, and I'm not pressuring myself to keep my working time to a minimum - I can fool around all I want. If I was in the Video Production business, I would definitely use Gearshift.

JohnnyRoy wrote on 9/23/2005, 8:38 AM

I guess it really depends on your processor speed, hard drive capacity, and final format. He is what I have seen in the tests that I’ve done:

Original M2T
Length: 00:00:14;08
Size: 46.60 MB
Playback FPS: 7

HDV Cineform
Render Time: 00:00:55
Size: 190.06 MB
Playback FPS: 14

DV Widescreen
Render Time: 00:00:38
Size: 53.12 MB
Playback FPS: 29.97

The FPS was measured with the preview on my Secondary Display set to Good (Auto). As you can see the DV Widescreen proxy from Gearshift renders 35% faster, is 3.5 times smaller and plays back twice as fast as the Cineform codec on my P4 3.0GHz PC.

These are no where near scientific. I just render to the three formats, placed them on the timeline, pressed play and reported the frame rate that Vegas showed at the bottom of the preview window. There are three data points none the less.

So it’s up to you to decide which format to use. We wrote GearShift because most peoples PC’s can handle DV. If your PC can handle Cineform (and have the 4x the hard drive space) go for it. I use DV Widescreen proxies to edit and use GearShift to render from the original M2T files to DVD MPEG2. No quality loss, great performance, small filesize. Your mileage might vary, void where prohibited by law, batteries not include. ;-)

~jr

David Newman wrote on 9/23/2005, 9:38 AM

In addition the CineForm codec scales very well based on preview size. So if you need fast preview, a full rate timeline playback is available (I get 29.97fps with 720x540 preview -- 3.2Ghz P4) but at anytime you can work in the full HD mode. This is why the CineForm codec is better than other intermediary codecs. The choice of using a proxy vs an intermediate, is an offline vs online workflow, there are advantages to both.

David Newman
CTO, CineForm

Wolfgang S. wrote on 9/23/2005, 9:46 AM

It depends also on the workflow you prefer:

Do you wish to render your final video with m2t -> m2t, or do you wish to work in the sequenze m2t-> Intermediate->m2t?

The first one is Gearshift - so you render from your original footage your final video, and switch to an proxy (DV-avi, mjpeg-avi); the second one is if you apply the Intermediate, where you render from the intermediate to your final video.

There is a quality drop in both workflows. Tests have shown, that the quality drop is also significant if you render directly m2t->m2t. However, with the older Cineform Intermediate codec up to Vegas 6b, the quality drop was higher if you use the Intermediate. With Vegsa 6c, there is a new version of the Cineform codec implemented. I have not been able to measure the quality drop up to now.

Spot|DSE wrote on 9/23/2005, 9:47 AM

Laurence, this is why we offer the ability to do both Cineform AND DV proxy in GearShift.
If your machine can manage CineForm's codec, then it's the best option. FWIW, compare CineForm to ANY intermediary on Apple or PC platform, it's superior. But...slower than 3GHz and you'll struggle a bit. This is where the proxies come in.
You can also do 4:2:2 YUV which is superior to all. But it requires some serious hard drive horsepower to manage 4:2:2 YUV.

Wolfgang S. wrote on 9/23/2005, 12:38 PM

Douglas,

in the tests that I have done up to now, the Cineform did not show up so bright. However, as said, I have not measured the new codec version, as delivered with Vegas 6c.