Sony Vs Main Concept render times

im.away wrote on 11/7/2010, 6:29 PM

G'day,

I'm currently rendering a 43 minute project shot natively at 1440*1080 50i. I've rendered it using the "PAL DVD Seperate Streams" template and produced a DVD with a 8 Mb/s sample rate and it looks pretty good when viewed on an average monitor. It took just under 2 hours to render. No problem. (I have to use a laptop PC by necessity, so I'm used to slow renders.)

Next I decided to render a version for posting on Vimeo. I chose the Sony AVC (MP4) template and selected 1280*720, 25p, 1.00 pixel ratio, 5Mb/s with audio set to 44.1Khz, 320 Kb/s stereo. This took a tad over two and a half hours to render and ended up being a 1.76 Gb file. Looks real nice when viewed on my PC. No problem.

Because the file size is pretty big and my upload speed isn't that great, I decide to try the Main Concept template (to see if it could make a smaller file) which I duly altered to match the settings I used with the Sony codec. Well, one and a half hours in, it is just 25% rendered and the "guess-o-meter" says it will take five and three quarter hours to complete!

I really don't know much about codecs and their internal workings but I'm guessing that the Sony AVC codec, while saying that it is sampling at 5Mb/s, is possibly a variable rate that is 5 Mb/s MAXIMUM - but not necessarily averaging that rate?

I set the Main Concept template to variable rate with 5Mb/s maximum and 4 Mb/s average. Because the job is still a long way from finished, I don't know what the final file size will be.

Anyone out there know what is really going on here?

Cheers

Russ

Comments

musicvid10 wrote on 11/7/2010, 7:09 PM

Sony AVC is CBR as far as I know.
Mainconcept VBR will take longer since it must make one or two separate analysis passes.

But keep experimenting, and when you have come up with your best solution, run the same test with Handbrake using single-pass CQ and decomb option at default. Compare sharpness and motion quality as well as rendering times.

Here is a set of tests I ran a while back for Randy Brown, a Vegas forum member.
http://vimeo.com/14514702
http://vimeo.com/14515027
http://vimeo.com/14515826

John_Cline wrote on 11/7/2010, 7:21 PM

File size is entirely a function of bitrate. If you render at 5 mbps, the file will essentially be the same size regardless of the encoder used. Since you set a 4 mbps average, by definition that file will be smaller than one encoded at 5 mbps.

amendegw wrote on 11/8/2010, 2:26 AM

@musicvid

First, much thanks for the postings on Handbrake. As a result of your research, I've started using Handbrake as my mp4 encoder of choice (for final/quality web postings).

Next, were you aware that there is a border problem in the Vimeo videos you posted? Most noticable at 48-49 secs. Obviously, this has nothing to do with the test you did, but the video is very well done and I didn't know whether you were using the clip for other purposes.

...Jerry

musicvid10 wrote on 11/8/2010, 9:05 AM

Yes, the border problem was in the original MPEG that Randy supplied me for testing. I considered cropping it in HB, which is easy enough to do, but decided to leave it the way it came for purposes of the test.

The original discussion is here. I also put up a Flash tutorial there that shows my "best" way to crop and encode that particular video in Handbrake for upload to Vimeo. The original files may still be available too:
http://www.sonycreativesoftware.com/forums/ShowMessage.asp?ForumID=4&MessageID=725579

musicvid10 wrote on 11/8/2010, 1:53 PM

This is the cropped upload version following my tutorial at mediafire.com/?l8i6leq5j0tnhhc
Only difference is this was encoded at CQ RF 17 resulting in higher finished bitrate.
Note the slight softness from cropping and resizing back to 640x480 at 1.0 PAR.

http://vimeo.com/16629896

amendegw wrote on 11/8/2010, 3:25 PM

musicvid,

fwiw the border funkiness (a technical term) is still there at secs 45-50. This is the aerial still overlooking the rodeo facilities.

...Jerry

musicvid10 wrote on 11/8/2010, 4:05 PM

Right, I made no attempt to custom crop it any further. The video is just an example that was provided by another forum member, and not my product . . .