-
QT Movie, QT Conversion, Compressor compression differences
My goal is to understand how to produce the highest quality web ready videos.
My assumptions:
1) FCP is non-destructive and source files will be re-rendered if placed in new timelines.
Example: I can edit DV 25 footage and 8-Bit graphics in a timeline with DV NTSC compression, then copy this timeline into a 601 uncompressed 8-bit timeline to print to a Beta tape. My DV footage will always be 4:1:1 and will suffer half a generation loss when transcoded to uncompressed 8-bit, but the 4:2:2 colour quality of my graphics and visual effects will be maintained.
2) Exporting as a QT movie re-renders all the source files, producing the same results as my example above: a DV compressed timeline exported using the Apple Uncompressed 8-bit 4:2:2 codex will produce better quality graphics and DV footage which is negligibly worse when compared to the same DV compressed timeline exported using the DV/DVC Pro NTSC codec.
Assuming above is correct (or even mostly correct),
My question is, what does FCP do when I export a DV compressed timeline with 8-bit graphics and effects using QT Conversion or Compressor. Does it:
A) Transcode the entire timeline as if it were all DV compressed, reducing my graphics to 4:1:1 before they have a chance to be rendered by a codec which might better maintain the quality of 4:2:2?
B) Uncompress each frame to its source and then recompress with the new codec? Maintaining the DV 4:1:1 and the graphic
-
3 Replies
-
None of that really matters. You’re going to a terribly lossy and reduced frame size for “web” delivery. Tell us what your target movie format and specs are and we’ll try to guide you backwards from those.
It doesn’t pay to maintain super-high production values at every step of the production process if you’re eventually just compressing for Web formats. You can bail out of certain steps and save yourself tons of time and rendering space by knowing where the effort is misplaced.
bogiesan
-
I appreciate that none of my concerns really influence my final product when i’m compressing for the web (typically I use h.264 or Sorenson 3), so I probably shouldn’t have stated ‘high quality web video’ as my primary concern.
Mostly, i’m a detail freak, I am looking for conformation of my assumptions (only if they are correct, mind you), and an explanation of how QT Conversion and Compressor render/transcode and/or render/transcode differently from one-another.
As far as details go, for the immediate future I will be working with Standard def DV footage and 8-bit graphics in FCP 4.5; however, I’m hoping for answers which I can apply to my workflow regardless of my output destination.
A better example than web video might be: what happens when I use QT Conversion to create a 720×480 file with the Animation codec (at Best quality, millions of colours) from a 720×480, DV NTSC compressed timeline?
Thanks again.
-
You’re presuming that there is an extra step between your timeline and the output format, and there isn’t. When using compressor or quicktime conversion, your source sequence is the input, and your settings determine the output, there isn’t anything else going on.
Now, here’s where it gets interesting. When you export a timeline that has graphics, say from LiveType or Motion, and the final output is smaller than the source, it doesn’t take a direct conversion of your rendered timeline, it re-renders the entire timeline in the new format, so any overlays of graphics get redone correctly – but only from FCP. If you export a Quicktime movie and then do the conversion using that as source, the text will start at the settings that the quicktime pulled from FCP, and every piece of video is one track that gets re-rendered. It never looks bad, but it sounds to me like you’re pretty picky and would notice (I do, too, but I shut up about it, because no one else notices.)
So what you really wanted to know was how to avoid a generation gap from your LiveType, Motion or AE clips when outputting, and to do that, use the original timeline, no matter it’s settings, as it has the master (unrendered) clips from motion, ae and livetype.
Reply to this Discussion! Login or Sign Up