Dish should investigate VC-1

[mike123abc]

Dish is having all these problems with MPEG-4 (see all the HD light and complaints in other thread), so one wonders if they would not be better off experimenting some with VC-1. If you look at the Blu-Ray/HD-DVD discussions VC-1 seems to produce a fantastic picture on reduced bit rates (way better than MPEG-2). The chips in the 622/211 are VC-1 ready already (L3 1080i not L4 1080p).

One has to wonder why all the HD-DVD movies are coming out in VC-1 and not MPEG-4. You would think they would have tried both and decided which one had the better picture.

 

[John Kotches]

So does MPEG-4.

The reason the movies are coming out in VC-1 is that MS is supporting the heck out of it. I know one of the people at MS that is working closely with the studios on this.

The difference here, is that you have the luxury of multiple passes of encoding to optimize everything when you are doing an offline encode as is the case with HD-DVD.

VC-1 would in all likelihood be just as problematic if it was forced to do realtime encodes as is required for Dish.

I'm not down on either codec; the reality is that offline (non realtime) and online (realtime) encoding are completely different ball games.

Cheers,

 

[shanewalker]

Damn, John, you beat me to the punch. Anyways, here is what I was going to say--

I've no doubts that proper VC-1 produces a fine picture. I also have no doubt that proper MPEG-4 encoding produces excellent PQ. I think the proprietary nature of VC-1 (versus MPEG-4 being an open standard w/ many spec partners) means that its been a bit more 'efficient' in getting the kinks worked out for some OEMs.

Or that they are heavily involved with/subsidized by corporate realtionships between Toshiba/HD-DVD partners and Microsoft. That can sway many companies to proclaim the superiority of a certain format (or OS ), regardless of any real qualitative benefits to the customer.

Read the article on HD-DVD featuring an interview w/ Joe Kane and a Microsoft rep in this months "Widescreen Review" for a peek into this sort of thing.

 

[mike123abc]

Yeah I realize that the real time encoding is a separate beast from the HDDVD world. But, Microsoft is really pushing VC-1, also in real time hardware encoders. I wonder which will get improved faster? VC-1 is now an open standard just like MPEG2/4. It is licensed through a patent pool. VC-1 just has a huge investment by MS in it to get it working better, one wonders if it would not improve at a faster rate than MPEG-4.

 

[HokieEngineer]

I noticed VC-1 mentioned in the overview for the encoders Dish uses:

DiviCom® MV 3500 Multi-Codec HD Encoder

Multiformat, multi-codec high definition (HD) encoder for MPEG-4 AVC with capability to support MPEG-2 and SMPTE VC-1at 1080i and 720p resolutions.
Motion Compensated Temporal Filtering (MCTF)
MotionTrack technology for full-field motion estimation
Multi-pass LookAhead CBR/VBR encoding
DiviTrackIP statistical multiplexing for HD and HD/SD pools
MPEG, AC-3, AAC and aacPlus audio encoding
ASI and IP outputs


http://www.harmonicinc.com/view_csd_product_group.cfm?classID=2374

 

[mike123abc]

Another thing is why is VOOM not using VC-1?

They should be maximizing their fiber feed to Dish with either VC-1 or MPEG-4, most (and all on some channels) are prerecorded, plenty of time to encode using state of the art VC-1 encoders. They could be done a week ahead and have a couple parallel tracks running to get them done. If good VC-1 takes 3x real time they could be doing 3days at the same time. Besides since they run the same shows a bunch they really have plenty of time. They could be delivering pristine VC-1 like the HDDVD player to dish for direct uplink.

 

[John Kotches]

And unless the new Broadcom chip can handle VC-1, that's even more receivers that will need to be put on the technological scrap heap.

All of the encoders will improve over time, this is a given. Just as MPEG-2 encoders have.

Which would improve faster? Well, experts in the field aren't hatched overnight, and unless you want to kill the people working on VC-1 support, it's a matter of picking the areas to support for Microsoft. So I would expect that improvements to realtime encoding might be a little slower than offline encoding.

Cheers,