mpg quality

auggybendoggy wrote on 11/8/2003, 6:23 PM
is there a rendering enging out there that does a MUCH better job than Main concepts mpg rendering. My AVI's are smooth as a whistle but when compressed down to MPG then they loose their shine.

Im filming with a 3ccd dv953 pana. It's a great camera but AVI's are just too big to burn to a DVD.

Any suggestions on using a different enging.

P.S. I've already tweaked my settings and no matter what it doesn't compare.
also someone might say thats the nature of compression but I ask "why then do my hollywood made dvd's look so good" its not avi's is it? How do they compress?

Auggybendoggy

Comments

Spot|DSE wrote on 11/8/2003, 6:49 PM
First, Hollywood uses hardware compressors. Second, they are compressed by frame, region, or defined region. Third, they are extremely expensive to do.
Encoding is an art form to get it best. This was the first class I ever started teaching, was compression. It's not just about taking an avi and dumping it into an encoder. Max bitrate in reality is around 6500 average. Knowing the tricks of getting every pixel to count is the art. Color, motion, garbage in the frame (particularly DV garbage, I don't care how cheap or expensive your cam is) are all factors in this formula. You can get very close to the Hollywood level, but not exactly there with a software compressor. Main Concept is among the best there is. Which is why it's the most used by many NLE plugins or systems.
auggybendoggy wrote on 11/8/2003, 6:55 PM
Spot, help me out here...
Your saying thats the top for us Vegas users?

I find that maxing out all bit rates makes it a bit smoother however my mbytes goes way up.

Perhaps you can help me so I don't have to learn the hard way of trial and error...

How long of a good quality mpg2 file will fit on a 4.7 gb dvd dics?

thanks for the help too spot, I appreciate it

Auggy
burchis wrote on 11/8/2003, 7:54 PM
Auggy,

I had this same concern a few weeks ago. Here is the thread that helped me out.

http://mediasoftware.sonypictures.com/forums//ShowMessage.asp?MessageID=225028&Page=0
TheHappyFriar wrote on 11/8/2003, 8:11 PM
I've found the Vegas encoder wonderful! I've rendered out stuff at 5mbs constant and it looks alsmost as good as the origional DVCPro footage.

If you want ot get REAL picky with youe encoding, here's 2 things to try:

1) Buy a hardware encoder, like an Optibase. I've used an older one at work that only did Half-D1 res (320x480 mpeg 2) and those mpeg's look as good as the Vegas rendered ones.

2) Or... buy a copy of TMPGenc. Then you can render out small seg's of your video at the quality you want, them join them together to make a final video. I've done that with good results.

What's your footage commang from? That could be a problem. Also, are you comparing the AVI on the computer to the DVD on the TV?
Spot|DSE wrote on 11/8/2003, 9:00 PM
A few tricks:
1. Remove 8-10 pixels from right/left sides of the image. Pan/crop works well for this, it will also take a few pixels from the top and bottom, but you'll not notice the cut. This removes the fringing that all DV cams have on the edge due to reasons that are too lengthy to explain here.
2. Use HSL to reduce saturation by as much as 10%. Less color means less for the encoder to look at. You'll not notice the color loss. Look at most Holly wood pix, you won't notice that they aren't colorful unless you are looking for it.
3. Use Black restore. Varying shades of black are a bear for the encoder to deal with. Set it at .010-.025 for best results.
4. Add .001/.002 of Gaussian blur. This also helps the encoder.

All of the above can help reduce contrast from pixel to pixel, meaning the encoder has less to deal with, resulting in a better encode. Redundancy is always better in an encode.
those are just a few tips, altogether or taken in pieces, that will guarantee a better encode.
stormstereo wrote on 11/8/2003, 10:58 PM
I guess this is applicable to any compressed rendering. Can I use Debugmodes 3D-plugin on the master video to cut off some pixels at left/right/top/bottom instead of pan/crop on every single clip? Will I achieve the same desirable advantages as with pan/crop? I know, I get black borders around my frame but that should mean less to encode right?
Also, any particular plug-in order? I use 3D(crop)-black restore-HSL (desaturate)-Gaussian blur.

Best/Tommy