Can a station broadcast mobile DTV and not OTA ATSC?

TonyT · 08-06-2009, 09:01 AM

My local news ran a story last night about Mobile DTV offering a variety of programming in the future. I thought to myself, how is ATSC M/H going to be any different than what's on TV now and then a thought came to me, will there be mobile DTV broadcasters that aren't broadcasting OTA ATSC signals? Could it provide for a cheaper way for entry level or startup companies to enter the broadcast arena?

Trip · 08-06-2009, 10:38 AM

The ATSC-M/H signal is a part of an existing ATSC signal. It's impossible to transmit M/H without the original signal.

Plus, the FCC requires at least one standard SD video feed.

- Trip

Piggie · 08-07-2009, 10:41 AM

Quote:

Originally Posted by Trip

The ATSC-M/H signal is a part of an existing ATSC signal. It's impossible to transmit M/H without the original signal.

Plus, the FCC requires at least one standard SD video feed.

- Trip

One could do that and leave room for 5 mobile channels. And not have too bad of a local channel. Probably one of the mobile channels is a network which they could put on the main feed.

But more over, on what channel? Most places are close to running out of channels.

Trip · 08-07-2009, 06:16 PM

The ATSC-M/H spec only allows for 8 groups, which is 7.33 Mbps, dedicated to mobile. (Not sure why they're crippling it like that.) That's probably 3 or 4 video streams.

- Trip

jsmar · 08-22-2009, 09:04 PM

Quote:

Originally Posted by Trip

The ATSC-M/H spec only allows for 8 groups, which is 7.33 Mbps, dedicated to mobile. (Not sure why they're crippling it like that.) That's probably 3 or 4 video streams.
- Trip

I believe this is incorrect. It took me a while to figure out where you may have gone wrong. I believe you are looking at the max number of groups per parade, which is 8. But as far as I can tell, you can have up to 112 parades. Of course that would only be feasible if each parade's PRC (parade repetition cycle) was 7, and each of those 112 (16 * 7) parades only used one group per subframe.

If you are looking for the maximum M/H bandwidth, it would be about double the number above. There are 16 slots per M/H subframe, and each subframe contains 156 TS packets. If the slot contains an M/H group then the M/H group uses 118 of the 156 packets, or 75.64% of the available bandwidth. There is no limitation on the number of slots that can contain an M/H group, i.e. all 16 slots can contain M/H groups.

I would note also that the number of groups can change from frame to frame (note - not subframe to subframe, i.e. each subframe within a frame must contain the same group layout), so categorizing an entire ATSC stream as containing a "number of groups" is probably not correct, although I am not sure in actual use if there will be any parades with a PRC value other than 1 (perhaps not for video, but I can easily see low bandwidth data doing this). I'm not sure if for example they wanted to allocate an average of 3.5 groups per subframe to a parade if they could set the number of groups for the parade to 7, but set the PRC to 2 (i.e. the parade would consume 7 slots per subframe in every other frame), or if setting the PRC to a value other than 1 is only for low bandwidth services that want to consume less than 1 group per subframe on average. I guess that would probably depend on the buffering standards for M/H video, which I have not looked at.

TVTom51 · 08-22-2009, 09:09 PM

Quote:

Originally Posted by jsmar

I believe this is incorrect. It took me a while to figure out where you may have gone wrong. I believe you are looking at the max number of groups per parade, which is 8. But as far as I can tell, you can have up to 112 parades. Of course that would only be feasible if the each parade's PRC (parade repetition cycle) was 7, and each of those 112 (16 * 7) parades only used one group per subframe.

If you are looking for the maximum M/H bandwidth, it would be about double the number above. There are 16 slots per M/H subframe, and each subframe contains 156 TS packets. If the slot contains an M/H group then the M/H group uses 118 of the 156 packets, or 75.64% of the available bandwidth. There is no limitation on the number of slots that can contain an M/H group, i.e. all 16 slots can contain M/H groups.

I would note also that the number of groups can change from frame to frame (note - not subframe to subframe, i.e. each subframe within a frame must contain the same group layout), so categorizing an entire ATSC stream as containing a "number of groups" is probably not correct, although I am not sure in actual use if there will be any parades with a PRC value other than 1 (perhaps not for video, but I can easily see low bandwidth data doing this). I'm not sure if for example they wanted to allocate an average of 3.5 groups per subframe to a parade if they could set the number of groups for the parade to 7, but set the PRC to 2 (i.e. the parade would consume 7 slots per subframe in every other frame), or if setting the PRC to a value other than 1 is only for low bandwidth services that want to consume less than 1 group per subframe on average. I guess that would probably depend on the buffering standards for M/H video, which I have not looked at.

When you say parade, what does that mean?

jsmar · 08-22-2009, 09:40 PM

Quote:

Originally Posted by TVTom51

When you say parade, what does that mean?

It's from the A/153 (ATSC M/H) standard, but I think it is terrible terminology. My guess is that the committee thought it was funny because a parade can contain one or two "ensembles" (think musically). Each ensemble encapsulates an IP datagram stream. What I am not sure of yet is whether or not there will be only one "service" (i.e. video) per IP stream, or multiple services per IP stream, or in the other direction, can a service span multiple streams from different parades. In the IP world packets can traverse multiple routes, so I would imagine that it is possible, although I don't know if that would be supported.

I've been concentrating on the very low levels of the standard right now, but I plan to try to fully understand the higher levels eventually.

TVTom51 · 08-22-2009, 09:55 PM

Quote:

Originally Posted by jsmar

It's from the A/153 (ATSC M/H) standard, but I think it is terrible terminology. My guess is that the committee thought it was funny because a parade can contain one or two "ensembles" (think musically). Each ensemble encapsulates an IP datagram stream. What I am not sure of yet is whether or not there will be only one "service" (i.e. video) per IP stream, or multiple services per IP stream, or in the other direction, can a service span multiple streams from different parades. In the IP world packets can traverse multiple routes, so I would imagine that it is possible, although I don't know if that would be supported.

I've been concentrating on the very low levels of the standard right now, but I plan to try to fully understand the higher levels eventually.

Appreciate the breakdown.

jsmar · 08-23-2009, 12:07 AM

Hmm, I'm used to the AVS Forum where I can edit my posts as appropriate. It appears that there is a fairly short time limit on editing posts in this forum.

Anyway, I wanted to add something to my above post regarding setting the number of groups for a parade to 7 and the PRC to 2. That was just an example. It could also be done by alternating the number of groups between 3 and 4 every other frame. I've seen some posts from Trip (Sorry Trip, I'm not trying to pick on you

) indicating that the number of groups is completely static. That may be true in current practice, but that is not what is in the standard. The number of groups for each parade can be changed dynamically from frame to frame (but not subframe to subframe). In fact, the standard specifically enables this by signaling the number of groups for the current frame during the first two subframes, and then switches to signaling in advance the number of groups for the next frame the parade will appear in during the transmission of the 3rd, 4th and 5th subframes.

Trip · 08-23-2009, 03:31 AM

I'd rather learn what's right than continue to be incorrect.

I've seen I think it was Harris (though it could have been someone else) specify that 8 NoG limit as well, or that seemed to be the implication. They talked about dedicating 7.33 Mbps to mobile and then the rest to "high quality HD" (a laugh, I know) but it's certainly possible that I misinterpreted.

Of course, in practice, I've seen no station go over 6 NoG so far, and I don't think I've seen a station change the NoG. I had assumed WHUT might test with 8, given they only had a single SD stream plus UpdateTV, but they didn't.

I'm glad someone understands that spec, because I sure don't. And really, I don't have the time to pick through it. My only question is, will you be writing software to decode it when you're done?

- Trip


Home \| DTV & HD News \| Today's Posts \| Interviews \| Join DTV USA Forum