|
|||||||||
|
Thread Tools | Search this Thread |
January 15th, 2010, 06:06 PM | #16 |
Inner Circle
Join Date: Jan 2006
Location: Minnesota (USA)
Posts: 2,171
|
Edius - arguably the most AVC friendly NLE out there right now, especially Edius Neo 2.5 (with the booster thingy that makes dropping AVCHD right onto the timeline work pretty smoothly, so long as you have a least a modest quad core CPU).
|
January 16th, 2010, 12:26 PM | #17 | |
New Boot
Join Date: Dec 2009
Location: Hagerstown, MD
Posts: 8
|
Quote:
|
|
January 16th, 2010, 02:05 PM | #18 | |||
Regular Crew
Join Date: Mar 2004
Location: Upper Pittsgrove, NJ
Posts: 95
|
Quote:
Cineform is, in fact, is based on a set of mathematics called wavelets, which is a fundamentally different compression technology than DV, JPEG, MPEG-2, and AVC, all of which are based on use of the discrete cosine transform. Not too important to know the gory details (I do, but I'm not trying to write a book here), just note that because wavelet is different than DCT, the kind of compression artifacts you'll see (or not see... these are pretty high quality CODECs) are different. There are some free technologies that do the same kind of thing as Cineform. The first, available today, is the Avid DNxHD CODEC. You can get this free for Windows (under Quicktime) on the Avid website. But the good news here is that this has been accepted as a standard, it's now SMPTE VC-3. Like Cineform, this has been designed to much lower in CPU use than MPEG-2 or AVC, but also to withstand repeated encoding and decoding without much damage, which is critical when you're doing heavy editing on video. I believe DNxHD is based on DCT, but it's new, and from a high end company like Avid, fairly trustworthy. Another standard in this space, though not quite ready yet, is Dirac Pro, which was invented by the BBC in England. This is very much like Cineform, based on wavelets. The Dirac format was created by the BBC with the idea that they, as well as the world, needed a video format for archival that was very high quality and not dependent on any patents or proprietary technology. It's too early to know just how fast this will be for editing, as it's kind of experimental at the moment. But it's coming... and the family of Dirac technologies was accepted by SMPTE as VC-2, so this is another open source industry standard. Quote:
Quote:
If it's just editing speed, try DNxHD if you're doing pro-level stuff, it may solve the problem. Otherwise, you can render out AVC in one of the higher end Sony MXF formats with virtually no loss. These use MPEG-2 at very high quality (better than HDV), and that's actually included with Vegas. The main issue there might be repeated layered edits.. I don't know if MXF/MPEG-2 would hold up as well as Cineform or DNxHD. But if it's just to get faster edits, try the MXF... it's built-in.
__________________
--Dave |
|||
January 16th, 2010, 04:09 PM | #19 | |
Major Player
Join Date: Nov 2009
Location: Melrose Park, Illinois, USA
Posts: 936
|
Quote:
|
|
January 16th, 2010, 05:49 PM | #20 |
Inner Circle
Join Date: Jan 2006
Location: Minnesota (USA)
Posts: 2,171
|
Heck, if you want a free intermediate codec, XVID can be configured to actually work quite well (limited to 8 bit 4:2:0 though). I did some testing with XVID (configured to use intraframe only compression) vs Canopus HQ and while Canopus HQ was a little faster, XVID was not exactly what I would call slow and yielded notably better SSIM scores and at slightly lower bitrates to boot (with some footage I shot of my puppy).
|
January 16th, 2010, 11:17 PM | #21 | |
Inner Circle
Join Date: Sep 2003
Location: Portland, Oregon
Posts: 3,420
|
Quote:
I tried DNxHD several months ago, with poor timeline performance, which I attributed to the QT wrapper. Not sure if this was a fair assessment. This was on something like a 2.6GHz Core2 Duo. Right or wrong, I'm thinking we need Direct-X or AVI or something other than a QT wrapper to get good performance on the timeline. Comments?
__________________
30 years of pro media production. Vegas user since 1.0. Webcaster since 1997. Freelancer since 2000. College instructor since 2001. |
|
January 16th, 2010, 11:55 PM | #22 | |||
Inner Circle
Join Date: Nov 2006
Location: Tallahassee, FL
Posts: 4,100
|
Quote:
Quote:
Quote:
__________________
DVX100, PMW-EX1, Canon 550D, FigRig, Dell Octocore, Avid MC4/5, MB Looks, RedCineX, Matrox MX02 mini, GTech RAID, Edirol R-4, Senn. G2 Evo, Countryman, Moles and Lowels. |
|||
January 16th, 2010, 11:56 PM | #23 | |
Inner Circle
Join Date: Nov 2006
Location: Tallahassee, FL
Posts: 4,100
|
Quote:
One of the reasons I wanted to move to MJ2K was to get an AVI container with a wavelet inside. Much like Cineform, but without the cost or having to have proprietary software to read and write the codec.
__________________
DVX100, PMW-EX1, Canon 550D, FigRig, Dell Octocore, Avid MC4/5, MB Looks, RedCineX, Matrox MX02 mini, GTech RAID, Edirol R-4, Senn. G2 Evo, Countryman, Moles and Lowels. |
|
January 17th, 2010, 03:48 AM | #24 |
Inner Circle
Join Date: Jan 2006
Posts: 2,290
|
I'm doing okay in 9c, no stuttering just yet. Still playing with it though. i7 920/Vista, 1080p footage in a 720p timeline. Perrone has reported some stuttering in the past though.
|
January 18th, 2010, 03:14 AM | #25 |
Trustee
Join Date: Oct 2009
Location: Central Coast Australia
Posts: 1,046
|
Its amazing to see the knowledge floating around in here, I love it.
So, uhm, could one of you rocket scientists answer me this? LOL. Would any of these codecs you speak of give me any better "bang for my buck" (quality for file size) than Lagorith? And is there anything wrong with Lagorith? A bit surprised it hasnt been mentioned. |
January 18th, 2010, 04:10 PM | #26 |
Inner Circle
Join Date: Jan 2006
Location: Minnesota (USA)
Posts: 2,171
|
I've never used Lagorith, but unless I'm mistaken it's a lossless codec, which means there there's no loss of quality at all from encoding, but file sizes will be larger than can be achieved with "lossy method" compression.
For "visually lossless" encoding of 4:2:0 source footage, that is really quite flexible (either optimizing for speed or for file size*), it would be difficult to beat XVID for the price (free). *XVID can be configured so that quantization never exceeds "1" even if using interframe compression with P and B frames, so in addition to being capable of configuration for very high quality (and quite fast) I frame only compression, the same image quality can be achieved with significantly smaller file sizes too (but performance is slower). MSU has a lossless codec (also free), which I believe is the best available for achieving the smallest file sizes without any loss of image quality. It's dang slow though. |
January 19th, 2010, 04:52 PM | #27 |
Regular Crew
Join Date: Dec 2009
Location: Encinitas California
Posts: 121
|
several years ago I used MSU often for some SD projects. Indeed it is slow - I rendered overnight. However, I never use it now. As I recall, I started having trouble with crashes - perhaps because I was moving from SD to HD
|
January 19th, 2010, 05:22 PM | #28 | |
Inner Circle
Join Date: Nov 2006
Location: Tallahassee, FL
Posts: 4,100
|
Quote:
__________________
DVX100, PMW-EX1, Canon 550D, FigRig, Dell Octocore, Avid MC4/5, MB Looks, RedCineX, Matrox MX02 mini, GTech RAID, Edirol R-4, Senn. G2 Evo, Countryman, Moles and Lowels. |
|
| ||||||
|
|