Talk:Compression artifact

From Wikipedia, the free encyclopedia

For those searching using British spelling: Compression artefact. —Preceding unsigned comment added by 217.24.129.50 (talk • contribs)

Note that a redirect from the alternative spelling was added just minutes before the comment above was made. - dcljr (talk) 07:45, 17 February 2007 (UTC)

Contents 1 Audio section 2 Unclear 3 Merge 4 A 42 kB thumbnail

[edit] Audio section

Hi, the audio section looks a bit unprofessional; it seems to be a partial revert of a deletion I made a while back. I think it needs to be rewritten to become more precise and accurate. --Kjoonlee 03:37, 13 October 2006 (UTC)

Saying "aggressive" in the leading sentence sounds a bit NPOV, does it? - 88.73.236.209 00:50, 19 January 2007 (UTC)

[edit] Unclear

In its discussion of picture encoding, the article states:

Where predictive coding of motion pictures is used, as in MPEG-1, compression artifacts tend to remain on several generations of decompressed frames, leading to a "painting" effect being seen, as if the picture were being painted by an unseen artist's paint-brush.

Where motion prediction is used, as in MPEG-2 or MPEG-4, compression artifacts tend to move with the optic flow of the image, leading to a peculiar effect, part way between a painting effect and "grime" that moves with objects in the scene.

So, am I to understand that "predictive coding of motion pictures" and "motion prediction" are two different things? This should probably be explained a bit further. A quick glance through the linked articles didn't help. - dcljr (talk) 07:45, 17 February 2007 (UTC)

Yes, simple "predictive coding" of motion pictures and "motion prediction coding" of motion pictures are two slightly different things.

Simple "predictive coding" uses the last image to predict the next image, only painting the pixels that change, sending the new pixel colors. The "motion prediction coding" technique builds on top of it. It also paints only the pixels that change, but rather than sending the actual new pixel colors, it sends a message describing the "motion" of a block of pixels from the last image.

Say you have a movie that has scene with a white truck slowly moving left in front of a landscape.

If you transmit it using simple "predictive coding", and someone loses the signal in the middle of previous scene, then regains the signal in the middle of the picket fence scene, that person will see the white (at the leading edge of the truck) and the various landscape colors (at the trailing edge of the truck) slowly paint over the previous scene.

If you transmit it using "motion prediction", it will compress much smaller, but that person will see a truck-shaped cutout of the previous scene slowly moving left.

Please update the article to clarify the unclear parts. --76.209.28.72 21:02, 22 June 2007 (UTC)

[edit] Merge

I think JPEG artifacts should be merged into this article, as they are both practically the same thing, and this article is longer. --AAA! ^(AAAA) 07:45, 16 March 2007 (UTC)

[edit] A 42 kB thumbnail

The original image thumbnail is actually 7.95 kB now (I guess wikimedia's thumbnail generation has changed). That's smaller than the poor quality version below it, so can't really correct the size without making things confusing. —Preceding unsigned comment added by AbleRiver (talk • contribs) 16:11, 31 August 2007 (UTC)