Data rate measuring Doo-Hicky?

[andrzej]

A long time ago there was a post here with some pertinent info http://www.satelliteguys.us/showthread.php?t=2528&highlight=technical. Here is an excerpt:

"Voom's carriers use 8PSK modulation with a symbol rate of 22.000 Mega-Symbols per second and three-bits per symbol giving a raw rate of 66 Mbps per transponder. Inner FEC is turbo-code at a 5/6 and outer FEC is Reed-Solomon. This results in a usable bandwidth of around 51 Mbps per transponder.

In comparison, Dish Network's QPSK carrier used for HBO-HD at 61.5 has a rate of around 27 Mbps and their 8PSK carrier at 61.5 with 2/3 code-rate gets around 41 Mbps.

The transport stream carried on their transponders is MPEG-2 standard. Voom uses Motorola's Digicipher II system for both conditional access and systems (things like channel names, transponder locations, program guide etc). DCII was developed many years ago by General Instrument before they were aquired by Motorola. They were heavily involved in the Grand Alliance that came up with the ATSC specifications. Needless to say, DCII and ATSC are very similar systems.

All video in the system is encoded in variable bit rate MPEG-2. From observing bandwidth used by each PID (shown by TSReader) I see an average of around 2-3 Mbps for SD channels and around 12 Mbps for HD. The highest I've seen is around 16 Mbps. This compares with about 13-14 Mbps for Showtime and HBO HD and 17+ Mbps for HDNet on Dish Network.

All audio in the DCII system uses Dolby's AC3 system. SD channels and Music Choice run 192Kbps and HD channels 384Kbps.

Virtually all the programming on Voom is scrambled. The exception being their barker channel (Inside Voom) and the Music Choice channels. The barker channel is encoded at 1440x1080 (2/3 horizontal resolution) and then interpolated out by the receiver. This is very similar to the way in which most DBS service providers encode SD material at 2/3 resolution (see http://www.coolstf.com/mpeg for more info).

The Music Choice channels also have associated on-screen graphics while the channel is tuned. These are delivered as MPEG-2 still pictures at very low rate and refresh every few seconds. These are encoded at 544x480 progressive."

I hope this helps.

 

[ChetK]

Maybe that's why, to me the compression is visible.

I don't know, but Wilt has agreed that a service call is in order for me. If I get a new receiver (swap) out of it, and the compression is visible, I'll know it's not the receiver, but low data rates (unless I would, by some chance, get two bad receivers in a row.)

 

[Scubasoul]

Maybe that's why, to me the compression is visible.

I don't know, but Wilt has agreed that a service call is in order for me. If I get a new receiver (swap) out of it, and the compression is visible, I'll know it's not the receiver, but low data rates (unless I would, by some chance, get two bad receivers in a row.)

Chet,

I really hope you have luck with the new receiver.

I just got a brand new one (specifically request it!) plus a 30" dish and still see the same PQ problems.

All those "bit" numbers mean very little (relatively speaking), I mean, either you like the PQ, or you don't, no matter what they tell you what the numbers are.

I still believe it's the receiver (how to explain OTA PQ), or the software.

 

[TheForce]

I have used the 169Time system with a computer monitor attached to measure the dynamic bit rates for audio and video feeds separately on Dish and Direct. I also had the opportunity to verify my readings with an RF engineer at one of those two facilities. (For security reasons, I won't say which one) The engineer was most impressed that I was able to read his bit rate so accurately. FYI- the bit rate is constantly changing and sometimes will yield readings that vary as much as 7-8 Mbps within 2 minutes depending on the video content. What the RF engineer suggested was that the readings be logged and an average over time be plotted for a quotable rate. He also reminded me that just because I get a reading of one thing on a Tuesday, doesn't mean he has the same rate for that channel on another day or later the same day. Much of what he and his staff does is make constant adjustments to the compression based on actual program content. What is an interesting evaluation using this method is to analyze the PQ based on readings at a particular time. But one cannot use this data to make a statement that a service is better than another service due to a collection of data one afternoon. This is why I never suport the often incorrect statement that HD PQ on DirecTV is worse than Dish. Fact is, I can make a case for either claim. I support the above statements that PQ is dependent on more than just bit rate
.

 

[vurbano]

I believe Dish is using VBR (or self-adjusting CBR that has normal and maximum allowable bandwidth usage) on the SD content.

By lumping in the SD channels with the HD content and guessing 2-3Mbps, I'm not sure the calculation of 11Mbps is completely accurate. I'm willing to wager it's 14-16Mbps. With all of the high-quality, high-motion HD content on Voom, you'd be screwed with CBR or VBR and such little bandwidth. As in my original post taken from DSLReports, It has to be at least 14Mbps for those high-motion channels...there is no reason it wouldn't approach 16Mbps on a show like MotoPIG. Are any TP's HD and SD content shared? Bandwidth can only be borrowed within the transponder.

O after what Sean added about a variable allocation the 11 MBPS is certainly wrong. i think 11 stinks. but maybe 12 is ok for a slow movie and 15 is required for rush etc. So some kind of device is allocating bandwidth dynamically. That certainly convolutes the matter.

 

[vurbano]

Maybe that's why, to me the compression is visible.

I don't know, but Wilt has agreed that a service call is in order for me. If I get a new receiver (swap) out of it, and the compression is visible, I'll know it's not the receiver, but low data rates (unless I would, by some chance, get two bad receivers in a row.)

It could be the current software/firmware as well. The system was designed for WM9/Mpeg4, im gonna stick it out till I see the final product. The way the software works with WM9 could create a much higher quality product, who knows?

 

[Hall]

I realize that there are other factors to video quality, but a 3 mbps difference can be quite a bit. I've authored DVD's that were 3 mbps (constant bit rate) that look pretty darn good.

You can't take a blanket number, using "3" as an example, and say something like it will make "quite a bit" of difference. The difference between 3mb/s and 5mb/s is 40%. The difference between 15mb/s and 18mb/s -- still "3" between the two -- is 17%.

 

[vurbano]

I think the thing we all have to keep in mind is that we know what PQ voom is giving us now, what PQ will they settle for under Wm9? just how far will they push it? what is their standard of quality?

 

[Ken F]

hall,

You can't take a blanket number, using "3" as an example, and say something like it will make "quite a bit" of difference. The difference between 3mb/s and 5mb/s is 40%. The difference between 15mb/s and 18mb/s -- still "3" between the two -- is 17%.

Obviously, it depends on the content. Your percentages are misleading because a 10% bit rate difference on a certain HD program could have a far greater impact on the picture than a 40% difference in bit rate on a SD program. This is because the compression level on HD is already much higher than it is with SD; there is a point where any further bit rate reduction results in significant quality loss unless excess filtering (removal of shadow detail) and/or downconversion is applied.

With 1080i HD, you're compressing 1.5Gbps down to 16-19Mbps. With SD, you're recompressing <143Mbps down to 2.0 to 3.0Mbps.

 

[Ken F]

vurbano,

WM9 will give us better quality at comparable and slightly lower bit rates. It remains to be seen just how much savings can be realized without degrading quality.

 

[andrzej]

...
With 1080i HD, you're compressing 1.5Gbps down to 16-19Mbps. With SD, you're recompressing <143Mbps down to 2.0 to 3.0Mbps.

Isn't that one of the advantages of 720p over 1080i? 720p requires less compression than 1080i with the same bandwith allocation.

 

[Ken F]

andr,

Yes, I suspect they can fit a 720p Cinema10 channel plus at least one SD channel in the same space as one of the 1080i movie channels.