Hi there,

Im just getting lost at the quantize matrix settings at TMPGEnc. What is it for? Whats the best settings for a good quality VCD? Thanks for the help.

-- E M Cabana (emcdvm@mozcom.com), March 14, 2002

Answers

I recommend you leave those alone. To be honest with you, I'm not sure what it's for and I would say that 99% of people who use TMPGenc don't understand it either. To get the best possible results, you should go to Advanced under Settings and use Noise Reduction. Set Still Picture to 100, Range to 4, Time Axis to 100, check Enable Filter and check High Quality Mode. It will take you 3 or more days to encode if you do this, but it will result in superb quality. If you don't need superb quality and just want very good quality, you can skip Noise Reduction. TMPGenc does a pretty good job without it and it will take a lot less time to encode. Just take the default settings for VCD and you'll be fine.

-- Jason (Jason.Shumate@equant.com), March 14, 2002.

If you're really desperate to know what it does, why not list down the settings on paper, tweak them all u want, then compare the results? It may take daze and u might get accused of reinventing the wheel at that but it will fully satisfy your curiosity and u just might even do Hori proud.

-- Mehmet Tekdemir (turk690@yahoo.com), March 14, 2002.

hi! Quantize Matrix is easy to understand^^

-The MPEG compression algorithm first cuts the video into 8x8 pixel chucks, known as marcoblocks. -Each of these marcoblocks are then passed onto an equation, known as DCT(Discret Cosine Transform). -What the DCT does is simply sort out the higher frequency of the video from the lower frequency ones, much like an audio. -Imagine a picture of an old man, you can see all the wrinkles of age on his face. This are the higer frequencies which took up alot of data. -A picture with all the higer frequency taken out, will look like cartoons, and the old man's wrinkles will be gone. -Here is where Quantization Matrix came in, the lower frequncy data will be located at the top left hand coner of the matrix, while higher frequncy content will be on the bottom right. -The numbers you put in is the quantize number, or denominator. -The higher the quantize numbers, the more zeros will be oupputed, thus leading to more compression, less data. -The magic of the MPEG compression algorithm is too keep all the important elements of the video(the old man, low freq), and throws away all the un-important stuffs(the wrinkles, high freq). -The number you put on the top left should be small, thus keeping the low freq data. -The number on the bottom right should be kept large, this will discard all the high freq data.

In short, higher number will result in poorer quality and smaller mpeg files. Play around if you got the time! ^^

-- Peter Gwee Min Yun (straitjacketanddiaper@yahoo.com), October 26, 2003.