1 person found this helpful
What you are looking for is called "smart rendering," and is available in some NLE's (Non Linear Editors). I think that Sony Vegas offers this, but better check first. Also, look at both CyberLink's PowerDirector and Magix MovieEdit Pro. These programs will take in your MPEG material, allow simple cuts and then Export to the MPEG CODEC, with no re-compression or Transcoding. Note: if you add overlay Titles, Transitions or Effects, they will have to Transcode, with re-compression.
The quest for quality vs file size is a tightrope walk, as you have found out. The high-quality and tiny file sizes is a "holy grail," and only so much can be done to attain it. Also, the judgement of "quality" can depend on the delivery scheme employed. WMV and DivX are good Export formats (do not edit worth a hoot!) for streaming media to be played on a computer. H.264 does a pretty good job too. For material to be viewed on a Web site, FLV is a good choice. All compress heavily, but depending on the viewing, look pretty good.
PrE is designed around a DV-AVI (a specific type of AVI) workflow. Basically, everything is internally converted to that. Aspects, like GOP are converted to full I-frame, Audio is Conformed to 48KHz 16-bit PCM/WAV, and then converted to 32-bit floating-point for precise editing.
The best workflow is to start with the best, least-compressed source material, and then let the delivery dictate how you Export/Transcode.
Good luck, and sorry that there is no definitive answer for you.
The more I learn of video the more confusing it is. Too many "great ideas" and not much coordination on formats.
When I look at reviews of video editing programs they ususally say for direct stream copy look at VirtualDub. Some college kid built this to work on games, and abondned project in 2005. A number of modules have been written as add ons, but a shame someone has not taken it and made it a full blown package for todays market.
But thanks for your help, you always offer good discussions.
1 person found this helpful
The latest release of VirtualDub is v1.9.9 released on 9th April 2010. VirtualDub is distributed under the GNU General Public License. I think (never having read it all the way through!) that would prevent anyone taking the code to make a commercial product out of it.
Insanity is hereditary, you get it from your children
You are correct. With Photoshop, it's basically PSD, TIFF, JPEG, PNG and maybe TGA for good measure. That does not even scratch the surface of AV formats, especially when you start with the variations of CODEC's, inside those formats. With PS, you just scan, or shoot, at a high resolution, and then work with the pixels. There is some compression, as with JPEG and LZW for TIFF, but that's really pretty simple. The biggest thing to learn in PS is Aspect Ratio, and then "don't JPEG compress a JPEG." After that, you're pretty much home free.
AV is a different wall of wax. Some of the image lessons play, like pixel x pixel dimensions, but things like DPI and PPI are thrown out the window.
In images, you can do all sorts of Aspect Ratios. In Video, you are looking at everything as though through a picture fame with a matte. This was easy for me, as I came from film, and was very familiar with the aperture of the gate on the projector. For others, it's a difficult concept to grasp.
I have had VirtualDub around for many, many years, but have never used it. Many do, and love it. I work with fairly common source footage, and almost always am producing to DVD-Video, so I am locked into the specs. of DVD. Those are strict rules, but take a lot of the choices out of the mix. I'm always shooting and editing for 720 x 480 NTSC @ 29.97 FPS, and only have to worry about my PAR (Pixel Aspect Ratio) being 1.21, or 0.91 (BBC).
When one begins working with different source formats/CODEC's, the playing field tilts quickly. Then, if they have varied Export needs, it tilts on another axis.
Wish that there were simple answers.