MxxCon
join:1999-11-19
Brooklyn, NY
clubs:  
| reply to Da Man
Re: Our way or the highway
18Mbps for 1080i != quality
The problem with such rule is there is no standard measurable unit of image "quality".
bitrate means nothing if codec doesn't know how to properly utilize it. while at the same time quality codec can encode superior image at smaller bitrate.
Just look at "megapixels" in digital cameras. |
|
sporkme
drop the crantini and move it, sister
Premium,MVM
join:2000-07-01
Morristown, NJ
 ·Optimum Online
| said by MxxCon  :The problem with such rule is there is no standard measurable unit of image "quality". There certainly good be though. Take the original, zoom in to some number of pixels. Do the same on the compressed version. Count how many pixels are different.
You can do this experiment at home. Take a high quality jpg. Save it with really crappy compression settings. Open up both. Zoom in to a complex area of the image. Count the pixels that are different. |
|
EPS
join:2008-02-13
Hingham, MA | But what is "different"? If a shade of red is slightly changed, is that equal to the shade of red getting blurred into gray? |
|
MxxCon
join:1999-11-19
Brooklyn, NY
clubs:
| reply to sporkme
said by sporkme :Take the original, zoom in to some number of pixels. Do the same on the compressed version. Count how many pixels are different. You can do this experiment at home. Take a high quality jpg. Save it with really crappy compression settings. Open up both. Zoom in to a complex area of the image. Count the pixels that are different. that kind of testing goes exactly against the way lossy video and audio compression works.
you try to throw away as much data as possible while trying to keep it perceptually identical TO HUMANS.
so at which point will you consider 2 pixels different? how would you know if some different pixels are important in an image or not?
would you want to waste video bandwidth trying to preserve details of a perfectly black sky with a few pixels being one shade brighter? or would you rather use it somewhere more obvious? |
|
PDXPLT
join:2003-12-04
Banks, OR
1 edit | reply to MxxCon
said by MxxCon :The problem with such rule is there is no standard measurable unit of image "quality" Not true.
There are ITU and MPEG standards for making objective measurements that correlate to perceived image quality.
And as for other "HDTV standards", there are whole volumes of ATSC and MPEG standards for modulation, coding, and transmission of HDTV signals.
The title of this article is the problem. There are plenty of HDTV standards, including those for objective quality measurement.
What there aren't any of are HDTV quality regulations. There's no law that that states what image quality providers must broadbast, other than an ambiguous one that says at least one prgram equivalent to NTSC quality must be provided.
Broadcasters, cable MSO's, and DBS providers believe most consumers don't know and don't care about pciture quality. They're probably right; how many TV's are horribly misadjusted? Most of them. The providers believe they'll be more profitable providing quantity over quality. So they pack in as many overcompressed channels as they can.
The one place you may get relief from this is Blu-Ray. There is no economic incentive to compress more than what is necessary to fit on the disk. And the folks controlling the compression are the same who produced the film, and have an interest in having its quality preserved and displayed. |
|
MxxCon
join:1999-11-19
Brooklyn, NY
clubs:
| said by PDXPLT :There are ITU and MPEG standards for making objective measurements that correlate to perceived image quality. show me
--
Check out my awesome city of MxxTopia »mxxtopia.myminicity.com/ind or »mxxtopia.myminicity.com (the more people visit, the bigger it is) |
|
Dogfather
Premium
join:2007-12-26
Laguna Hills, CA | reply to MxxCon
That is part of what a standard would determine...at what point are they considered different. |
|
Dogfather
Premium
join:2007-12-26
Laguna Hills, CA
·Cox HSI
 ·Verizon FIOS
 ·Cox VOIP
·ViaTalk
·RoadRunner Cable
 ·MegaPath
·Verizon west (ex G..
·Time Warner VOIP
3 edits | reply to MxxCon
ITU-R BT.500
»www.dii.unisi.it/~menegaz/Doctor···0-11.pdf
My guess is that the MPEG group would follow this recommended methodology as other engineers and researchers would. |
|
