[opendtv] Half Truths - Was More 1080p@60

At 3:18 AM -0800 12/4/07, Ron Economos wrote:

http://www.ambarella.com/news/press_releases/pr_09102007.htm

http://www.ambarella.com/docs/1080p60.pdf

Ron
Here we go again, and once gain it is the usual players in a race to the bottom.
I guess we cannot blame encoder manufacturers for misuse of their products, but the main thrust of this announcement is that it is now possible to abuse HDTV as much as SDTV has been abused for the past decade.
Harmonic has supplied several generations of encoders to the DBS systems among other customers. Each new generation makes it possible to deliver significantly higher quality at the same bit rate as the generation that it is replacing. Unfortunately with each new generation, the distribution systems choose to maintain the same poor level of quality at a lower bit rate. A significant factor in all of this is the se of pre-processors to filter out the stuff that is hard to compress...like accurate picture details.
There is SOME good news in this announcement. At least they got the part about the problems with interlace right. Unfortunately, they are using these logical arguments to justify the emission of 1080@60P in place of 1080@30i, using the dubious logic that program distributors (including broadcasters) need to catch up with the display manufacturers who are now pushing 1080@60P displays, often in screen sizes that are incapable of using this extra detail.
Consider the conclusions in the Ambarella 1080p60.pdf file for which Ron provided the link: 



Conclusions
Moving to a progressive 1080p format at frame rates of 50 fps in Europe and 60 fps in North America and Japan simply means that the broadcast industry is catching up to the advances in consumer displays and is ready to offer consumers increased sharpness and smoothness of motion that only a truly progressive format can deliver. Practically, the 1080p format is well suited to the capture devices and flat-screen TVs available today. The current interfaces between DVD-players, set-top boxes and TVs already support the format.
Since compressed versions consume only 20% more bits, 1080p is perfectly suited to the existing distribution broadcast infrastructure, whether the delivery medium uses cable, satellite, DSL or fiber to the home. The 1080p60 HD AVC decoders are already running in labs, and 3G HD-SDI and single-chip encoders are available now, so progressive-format technology can be deployed today.
The first statement is completely WRONG. The advances in consumer display technology have NOTHING to do with acquisition or emission. These advance, when properly applied to larger screen sizes simply take advantage of the very beneficial aspects of display oversampling. Just because it is possible to display 1080@50/60P DOES NOT MEAN that it is possible to acquire and compress this format to the level needed for emission. We've been through all of this before with the EBU H.264 encoding and display tests. The conclusion above DOES properly state the advantages of "increased sharpness and smoothness of motion that only a truly progressive format can deliver." But it then goes on to suggest that these benefits can be maintained for a 1080@50/60P format. It is a flat out falsehood that capture devices can properly deal with 1080@50/60P acquisition today; ALL 1920 x 1080 capture devices - both interlaced and progressive fall far short of the ability to deliver high frequency details WITHOUT high levels of debilitating noise. Most do not even sample the full raster, and very few attempt to compress it at this resolution. It is not surprising that one of the main features of the Ambarella chip is motion adaptive noise reduction; the question is, how much detail does this process remove? To get down to the bit rates they suggest for both 720P and 1080P requires very heavy pre-filtering - a good 1024 x 576 encoding could deliver equal results.
The reality is that 1080@30i has NEVER been distributed at bit rates below 15-18 Mbps without severe degradation and loss of the high frequency details. Much the same can be said about 720P. These formats BOTH need peak bit rates above 25Mbps to maintain the quality of the HD samples when the encoder is stressed. Saying that it only costs 20% more bits is ludicrous, when 100% of the bits available was not doing the job properly in the first place. The only conclusion we can draw is that these vendors are encouraging program distributors to keep delivering crappy pictures, as long as they can claim bigger numbers.
They are correct that progressive-format technology can be deployed today. The EBU studies provide the correct technical basis to justify this, and their conclusion is that we should use 720P for emission. If there is any good news here it is that program distributors can achieve the lowest HDTV bit rates with 720P. Perhaps if we can get 720P adopted as the "baseline" we may be able to use marketplace forces to get program distributors to use higher bit rates to deliver REAL HDTV. 

Regards
Craig




----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at FreeLists.org 
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word 
unsubscribe in the subject line.

Other related posts: