View Full Version : Avi intermediate
March 23rd, 2009, 10:20 PM
I had a real hard time on one clip getting to play on exposure room without alot of artifacting, and blur. I did alot of panning. I tried rendering out the veg. file in all sorts of ways, mpeg4, wmv, quicktime,, avchd was best, but not good enough. tried an avi mpeg4 h.264, still not good. flash the same, either too big, or had issues, finally tried, just for kicks, a avi sony yuv, it produced a huge file, 40 gigs, but clean, so I rendered it as sony avc hd at 4mps, 720p and when exposure room cut it from 163 mb down to 40mbs, it still remained pretty clean with all the camera movement, anyone else ever try the avi codec before
LA Arboretum By Hugh Mobley On ExposureRoom (http://exposureroom.com/members/hmobley.aspx/assets/0cbe12b813384945a81e39092fa98d3a/)
March 23rd, 2009, 11:01 PM
I will be watching this thread with great interest. I have nothing to contribute unfortunately.
March 24th, 2009, 10:38 AM
Back in ancient tymes, before the Cineform codec was included with Vegas, Sony YUV was popular as a digital intermediate format. Not quite as big as uncompressed, but much less compressed than anything else.
You might try using an AVI-Cineform the same way as you did AVI-SonyYUV. The older codec included with Vegas is still an outstanding DI codec.
I just looked at your piece on XR - yes, you have some camera work there that's going to challenge codec performance! Lots of detail, all moving camera, every pixel changing on every frame makes it a great test piece, though I suspect that wasn't your goal...
So, yes, a workflow that ends with AVCHD or MP4 would seem to be best, although if you had access to the On2 VP6 flash codec it would be worth trying. Spendy, though.
March 24th, 2009, 02:39 PM
I already use cineform, I capture with cineform using high into an m2t, then convert to avi, would like to use mov. in cineform but vegas only likes it sometimes. I have the latest cineform codec in the dropdown avi but it was captured with cineform so I am assuming that would be redundant to use that one again, I was just messing around trying to get that one clip to look good in exposure room after they compress down to small, normally I just render the cineform avi intermediate into an avchd, or into a flash file. it didn't seem to make this one clip look any better. too much camera movement when panning, and I am using the V1, could be the cmos. I might try a higher shutter speed next time
March 24th, 2009, 05:02 PM
...I capture with cineform using high into an m2t, then convert to avi, would like to use mov. in cineform but vegas only likes it sometimes. I have the latest cineform codec in the dropdown avi but it was captured with cineform so I am assuming that would be redundant to use that one again...
I'm not understanding these statements. Given cineform capture, here's what you might try:
1) Capture with a Cineform application (Could be HDLink, NeoScene, whatever version you have). Capture into medium or high. You now have an avi using the cineform codec.
2) Edit with that capture on an HDV timeline. You could render directly to MP4, but,...
3) To follow your previous success, render to an AVI-Cineform intermediate file.
4) Convert that avi-cineform file to an AVC/h.264 format, such as MP4.
M2T shouldn't enter into it at all if you're doing your original cap with a cineform app. The free Spark flash codec isn't going to be much help, but the VP6 codec might be good if you have access to it (comes with the full CS3/CS4 or Flash CS3/4 applications, and some versions of Sorenson Squeeze).
March 25th, 2009, 07:26 PM
simple, I capture usiing cineform, only, I capture as an m2t file thru cineform and not directly to avi, I started doing this for two reasons, first its faster, and second cineform was having a problem when capturing and converting lets say 1 full tape. so now I do it this way. now I have the m2t file which I can have cineform convert it to either qt or avi. I used to capture and convert, but it takes too long, thats why we are going to switch to digital as soon as we can.
March 25th, 2009, 09:01 PM
There's nothing wrong with the workflow of starting from m2t captured from the tape, as with firewire that is a direct digital transfer with no generation loss - the file on the PC is now the same 1s and 0s from the file on the tape.
Generally speaking, you'd never want to convert to m2t, since it is a lossy format. It makes a poor digital intermediate.
Most people using the Cineform applications for capture go directly to the cineform codec, but it isn't the only way to work. The Cineform capture apps offer some nice tricks like pulldown removal, flipping, etc. However, those who have spent the money on the Cineform applications usually have done so because they like working with the Cineform codec, and the capture apps streamlines workflow by allowing direct capture to that codec.
Later conversion from m2t to avi-cineform is also common, since there's a cineform codec included with Vegas, and it really is a great codec for digital intermediate work, stands up well over multiple generations, good for color correction, etc.
Well, enough cine-babbling from me! How are things going on your project?
Rendering out of Vegas to avi-cineform is not redundant use of the codec, it takes advantage of one of the things the codec does very well - it's a great hi-def codec out of vegas when you want to do further work in vegas or another application, eg. to transcode to h.264/AVC/MP4.
I'd guess that the compression challenges with that clip have nothing to do with the V1's CMOS sensor, nor with shutter speed. The CMOS "rolling shutter" effects are seen at *much* higher subject/camera speeds.
I'm thinking it's all about the picture complexity of the constantly moving camera over highly detailed subjects.