Explaining blockiness!

[lnetzel]

What does the blockiness really come from.. sometimes it's fine and sometimes it's to much of it... is there some special type of footage that usually get really blocky?

I usually Encode DivX movies to VCD..
"TimeCode" turned out REALLY GOOD... MUCH better than VHS..
but "Killing Zoe" was'nt very good and VERY blocky, although I used the same template in TMPGenc.. Of course the source needs to be good.. but I could'nt really se any differance on the DivX movies...

Can smebody please explain what all this depends on?

[aldus4]

Here the definition of blockiness/artifacts and where it originates from the Glossary of a DVD website:

"Artiface: Something evident in a picture that was not there to begin with. For example, when you watch an AVI or Quicktime movie on your computer, you often see blocky picture elements. Those blocks are artifacts because they were not a part of the original picture. On DVD, artifacts are usually a symptom of poor mastering, poor playback equipment, or improper adjustment of your television monitor. Make sure to calibrate your picture using a test disc like Video Essentials if you feel you are seeing artifacts."

[KoalaBear]

"Blockiness" in an MPEG occurs when you have more motion in a scene than the bitrate can reproduce.

On a DVD, even though the average bitrate might be 2 to 3 Mbps, it can shoot all the way up to 10 Mbps for brief periods to ensure that there are enough bits available to reproduce the motion. (Try turning on the bitrate display if your DVD player supports that feature and you'll see what I mean).

A VCD, on the other hand, has a fixed bitrate of about 1.44 Mbps. Above a certain point there just aren't enough bits to reproduce the motion accurately, so the excess motion gets thrown away.

MPEG uses something called "motion compensation" to compress data between frames. It searches for a match between a macroblock in the current picture with a macroblock in the next picture. If a match is found, it sends a command (called a 'motion vector' to the decoder to tell it to shift the macroblock around in its memory so it can save the bits that would otherwise be required to encode a new macroblock for the next picture.

But when the motion is very fast, there are few or no matches between macroblocks in successive pictures and the encoder has no choice but to create new macroblocks to describe the difference between frames.

When the encoder reaches the point of saturation, i.e., not enough matches to use motion vectors and not enough bits to create new macroblocks, the remaining differences between frames are ignored. This is the point that "blockiness" becomes visible, and the longer the motion is sustained the worse the picture looks.

Hope this answers your question.

<font size=-1>[ This Message was edited by: KoalaBear on 2001-08-01 07:20:40 ]</font>

[LighthouseJ]

I've also noticed artifacts if you had a segment of a movie with the video of a solid color but with slight hue or brightness changes, instead of representing millions of pixels as the same color, it will use square blocks, 32x32 pixels I think of the color to reproduce the effect and saving bitrate. It's especially evident in divx and it's noticeable in scenes with lots of black and dark colors, you can see the array of black and dark grey squares fluctuate.

[lnetzel]

Great answers... thank you