|
|||||||||
|
Thread Tools | Search this Thread |
January 21st, 2006, 10:54 AM | #1 |
Tourist
Join Date: Jan 2006
Posts: 3
|
My workflow in post with HDV: no intermediates, please :)
My workflow in post with HDV for 720p (or 1080i should I use it):
Goal: Achieve a workflow with HDV recordings that: - does not re-encode lossless at any point, not even to a "visually lossless" codec. - does not involve writing intermediate files. Only the captured file and the final output file are desired. - is performant enough to do online HD editing with workable previews. - uses only opensource or freeware tools: no sense in paying for software that fixes shortcomings in other software. 1. Hook up HD1 to PC with Firewire 2. Capture TS files with CAPDVHS 3. Extract the MPEG out of the TS with HDTVTOMPEG 4. De-multiplex the audio and write a D2V file using DGINDEX 5. Write a myfile.AVS AVISYNTH script to: 6. a) Open the D2V video file with DGDECODE.DLL 7. b) Open the demultiplexed MPA audio file with MPASOURCE.DLL and remultiplex with video 8. Install IM-AVISYNTH.PRM in Premiere Pro 9. Start up Premiere Pro, open up the AVS file(s) as source clips and do your NLE work 10. Install DEBUGMODE FRAMESERVER for Premiere Pro. 11. Render the timeline in Premiere towards the Debugmode Frameserver encoder and choose a myfile.avi as signpost filename. 12. Open up your favorite encoder (windows media encoder, TMPEGEnc, Quicktime) and encode in the desired output format. This acomplishes my goals stated above. Only "inefficiency is the demultiplexing of the audio, generating a separate file containing the audio data. All the other files are "signpost" files, i.e. they don't contain any data but are used as a placeholder to exchange information between the frameserver and the application opening the file. I leave in the demuxing since this is needed to use the dgdecode.dll (which only handles video). Alternatively I could replace step 4-5-6-7 by making a 1 line script for AVISYNTH that simply uses DirectShow codecs to open up the mpeg program stream with audio included, but I rather not trust on whatever happens to be the default MPEG2 decoder at that moment in time on my system. I experienced lock-ups using the DirectShow codecs, hence the more complex apprach from 4 -> 7. If you want to do format conversions (PAL-> NTSC, frame rate, HD->SD, etc) you can add between step 11 and 12 a Avisynth file that does the desired conversion for you, again lossless without reencoding. I'm just writing this up so people can benefit from it. I don't see why some people buy expensive transcoders like the Mainconcept MPEG Pro plugin. They lower the quality of video by transcoding, you need to wait for them to write their re-encoded incredably huge files to disk and they cost you money. Once my system is setup, I don't have to do anything, it's just capture, edit and output. |
January 22nd, 2006, 01:51 PM | #3 |
Tourist
Join Date: Jan 2006
Posts: 3
|
I used HuffJuv with VirtualDub in the early days of NLE with Hi8 and D1 resolution.
However, with HD resolutions file sizes really become too big to handle. I prefer letting the CPU do some work while I'm editing instead of buying a RAID array of 10 200GB disks just to edit 2 hours of rushes. |
January 22nd, 2006, 05:56 PM | #4 |
Inner Circle
Join Date: Apr 2003
Location: Aus
Posts: 3,884
|
err.. why not just use premPro2... it does everything without all the messing around...
|
January 22nd, 2006, 06:30 PM | #5 |
Trustee
Join Date: May 2004
Location: Knoxville, Tennessee
Posts: 1,669
|
Presumeably because an $849 programme isn't what's usually meant by 'freeware' - see Goal 4 ;)
|
| ||||||
|
|