Greetings ...

At present, it seems (in my own experience, anyway) that one 6-mhz digital channel can hold at least 4 sub-channels, one offerring HDTV, the other three SDTV.

Looking to the future, will improvements in digital compression technology allow for more sub-channels to be included within the same 6-mhz space? If yes, what do you think will be the potential number of HDTV and SDTV sub-channels occupying the same 6-mhz space?

Keep answers simple and give examples, such as -- 2 HDTV & 4 SDTV -OR- 3 HDTV & 1 SDTV

I do realize that a key element to sub-channels is the compression stream, which for a 6-mhz OTA digital TV channel is 19.39 Mbps, and that at the very least 11.00 Mbps is needed for an HDTV Channel.

Using Star Trek speak, I would suspect that current digital compression technology is perhaps somewhere at "impulse engine" level, and that maybe we have not yet reached "warp" or "transwarp" level - meaning, a way to signiifcantly increase the compression stream rate without compromising resolution quality

Thanx-A-Lot, Frank-0-Video

Without upgrading compression standards, such as going from MPEG-2 to MPEG-4, there's a limit to what we can do.

Right now, you can show about 7 SDs on a channel without them all looking crappy. I've seen two HDs look decent but not great. The gear is getting better all the time, but I don't think anyone really knows what the limit is at this point.

- Trip

Los Angeles already has a station with 18-1, 18-3, 18-5, 18-6, 18-7, and 18-8 right now.

http://thumbnails7.imagebam.com/2571/c659d925703881.gif (http://www.imagebam.com/image/c659d925703881)

Could they get 10 channels in there?

...at the very least 11.00 Mbps is needed for an HDTV Channel.

Hi there

That is not written in stone, and the broadcasters seem to be all over the map in setting the compression levels and the bitrates. KCBS-DT consistently has the highest bitrates, and KABC-DT the lowest. The average bitrate for "Lost" on KABC-DT is only around 10 Mbps.

Regards

Los Angeles already has a station with 18-1, 18-3, 18-5, 18-6, 18-7, and 18-8 right now.

All SD? 6 SD's is the limit per channel right now and it appears they are maxed out.

Los Angeles already has a station with 18-1, 18-3, 18-5, 18-6, 18-7, and 18-8 right now.

http://thumbnails7.imagebam.com/2571/c659d925703881.gif (http://www.imagebam.com/image/c659d925703881)

Could they get 10 channels in there?

All SD? 6 SD's is the limit per channel right now and it appears they are maxed out.

There's no limit on subchannels, only on bandwidth.

Could they do 10 subs? Yes. Would they all look like crap? Almost certainly, unless they scaled all of them down to 352x480i.

I find SD subs can look watchable at about 2.5 Mbps assuming a resolution of 704x480i. At that bitrate, you can fit 7 SD subs safely. Going up to 8 is doable but pushing it.

Going below 2 Mbps turns it into a sea of pixelation. First-hand observation on that one.

- Trip

All six channels are 704x480i video and 192 kbps audio

MPEG Video: Bitrate 3.044 Mbps Resolution 704 x 480i
AC3: Bitrate 192 Kbps Sample Rate 48 KHz

Going below 2 Mbps turns it into a sea of pixelation. First-hand observation on that one.

- Trip

Agreed. Even approaching 2 mbps is pushing it. Around 3.5 SD looks fine.

In my opinion with an HD, there is only room left for one SD. Possible to run slides on another channel for PSA stuff, but little to no moving video.

WUFT, just the Usage PID from HTML output of TSReader:

0x0000 (0.08% ~ 0.02 Mbps) 0x0000 MPEG-2 Program Assocation Table
0x0030 (0.02% ~ 0.00 Mbps) 0x0030 MPEG-2 PMT for program 1
0x0031 (73.86% - 14.33 Mbps) 0x0031 MPEG-2 Video for program 1
0x0034 (2.04% ~ 0.39 Mbps) 0x0034 AC-3 Audio for program 1
0x0040 (0.02% ~ 0.00 Mbps) 0x0040 MPEG-2 PMT for program 2
0x0041 (10.58% - 2.05 Mbps) 0x0041 MPEG-2 Video for program 2
0x0044 (2.04% ~ 0.39 Mbps) 0x0044 AC-3 Audio for program 2
0x0050 (0.02% ~ 0.00 Mbps) 0x0050 MPEG-2 PMT for program 3
0x0051 (6.57% - 1.26 Mbps) 0x0051 MPEG-2 Video for program 3
0x0054 (2.04% ~ 0.39 Mbps) 0x0054 AC-3 Audio for program 3


5.1 is using 14.33 and looks picture perfect.
5.2 is using 2.04 and could be better. Like a bad tube set from the 1970s
5.3 is 1.26. Macro Block City. Yuck. Zoom it to full screen if they are showing letter box and you think you are in the 1960's (enough said).

Well, I've seen 2.5, 2.0, and 1.5, and my observation is that 2.5 is pretty watchable, 2.0 shows pixelation but is definitely usable, and 1.5 isn't worth watching.

What's interesting in your case, Piggie, is that you're getting these results despite the SD feeds using 352x480i rather than the standard 704x480i. I would have expected the 2.0 one to still look good, and the 1.25 one to be lousy but not horrible.

- Trip

Greetings ....

I appreciate the replies so far, and I understand the problems of having so many sub-channels in the same 6-mhz space. But that is with the compression technology used now.

How much HIGHER will compression streaming rates be in the near future? If the rate for a 6-mhz channel is 19.39 Mbps now, could we expect to one day see rates at the 40 Mbps level? 60 Mbps? Higher? Please let me know where things stand in that regard.

Thanx-A-Lot, Frank-0-Video

How much HIGHER will compression streaming rates be in the near future? If the rate for a 6-mhz channel is 19.39 Mbps now, could we expect to one day see rates at the 40 Mbps level? 60 Mbps? Higher? Please let me know where things stand in that regard.

No. That bitrate is exactly how much data plus error correction can fit in the 6 MHz. In Europe the bitrate is roughly the same, and they're using a different system altogether.

Any future advances will come in the form of superior video compression. You really can't pack more bits into the same amount of spectrum, you can only make each bit matter more overall...

- Trip

You really can't pack more bits into the same amount of spectrum, you can only make each bit matter more overall..


Hi there

But if you scrap the 8VSB, there could be a modulation in the future that would allow higher bitrates (more bits per symbol) in the same bandwidth. Of course that could mean new transmitters and tuners/receivers. We haven't even been able to complete 1 changeover after what, 6 years of planning?

> Any future advances will come in the form of superior video compression

That is the cheaper path for transmitting more video information


So for the "near future" there won't be any change in video bitrate capacity because we'll be standardized on MPEG-2 encoding and 8VSB modulation.

Regards

Hi there

But if you scrap the 8VSB, there could be a modulation in the future that would allow higher bitrates (more bits per symbol) in the same bandwidth. Of course that could mean new transmitters and tuners/receivers. We haven't even been able to complete 1 changeover after what, 6 years of planning?

More like 15 years.

There's technology like that now, but the tradeoff is that you have to have error correction in an OTA signal. Look at cable; the same 6 MHz channel has almost 40 Mbps in it, but then cable doesn't have to deal with atmospherics and whatever other factors OTA has to overcome. If you shove more bits in the stream, you then have to increase the error correction to compensate. Any advances in alternate modulation methods will likely be very slow to provide any major benefits for that reason.

- Trip

Greetings ...

Bit and Bit, my DTV 101 education continues. Appreciate the info, Trip.

Let's hope then that the art of video compression is indeed still very much in its infancy. I see where the storage capacity of future Optical Media discs is projected to be in the 1000's of gigabytes, compared to present-day DVD (4.7 gigs) and Blue-Ray (25 gigs). Maybe future video compression technologies won't leap quite so far comparatively, but if the day comes when 8 digital TV sub-channels within the same 6-mhz space can each deliver the same type of picture as one current-day HDTV sub-channel, I know we'll all be pleased with that.

Thanx-A-Lot, Frank-0-Video

DVB-T has 3 different modulations and variable guard interval. If the chart on Wiki is correct it can go up to 31Mbps using 8MHZ. No idea if anyone uses it or anywhere close.
For ATSC, I don't think we're going to be seeing any noticable improvements in MPEG2 encoders so what we see today will probably be the same in 2050.

Greetings ...

CRTDude - Just saw the DVB-T Wiki-Chart you mentioned, and noted that it was for 8-mhz channels. If I interperet both it and you correctly, and if applied to a 6-mhz channel, would not the maximum Mbps be about 23.75 ?? If yes, would that equate to about 11.875 in a 3-mhz channel?

Thanx-A-Lot, Frank-0-Video

Los Angeles already has a station with 18-1, 18-3, 18-5, 18-6, 18-7, and 18-8 right now.

http://thumbnails7.imagebam.com/2571/c659d925703881.gif (http://www.imagebam.com/image/c659d925703881)

Could they get 10 channels in there?

Remind me never to watch that. It must look terrible. Leave all the subs out and give me full bandwidth HDTV....Please...

As Trip said:
Without upgrading compression standards, such as going from MPEG-2 to MPEG-4, there's a limit to what we can do.The ATSC standards address the use of MPEG-4/H.264/AVC.

http://www.atsc.org/standards/a72.php

Also, this article discusses the adoption of H.264/AVC

http://www.engadgethd.com/2008/09/22/atsc-2-0-includes-support-for-h-264/

This could effectively double, maybe even triple the capacity of a DTV channel. With H.264/AVC the first symptom of over-compression seems to me to be a dull/soft picture and banding as opposed to a macro-blocked/mosquito noise picture (like MPEG-2 gets) so it depends on how soft/banded the broadcasters are willing to let their video get as to how much bandwidth efficiency would increase.

However, most if not all the STBs and TVs being sold now in the U.S.A would not be able to do H.264/AVC so that is probably way in the future, maybe never. But never say never....

dmatch

If all that is being transmitted is a static image, then you can have a whole lot of subchannels that will show the images just fine. Once you add constant changes to the displayed images, then the number of subchannels that can display a good image goes down rapidly.

Greetings ...

CRTDude - Just saw the DVB-T Wiki-Chart you mentioned, and noted that it was for 8-mhz channels. If I interperet both it and you correctly, and if applied to a 6-mhz channel, would not the maximum Mbps be about 23.75 ?? If yes, would that equate to about 11.875 in a 3-mhz channel?

Thanx-A-Lot, Frank-0-Video

In theory, yes, but the other thing to keep in mind is that I'm pretty certain DVB-T requires a higher signal-noise ratio, and thus the extra bits are coming by reducing error correction. (Someone who knows more about it might want to check into that, as I'm not 100% sure. I remember seeing something about that somewhere.) The signal becomes harder to receive when you do that.

- Trip

Greetings Trip ...

All right, suppose we had an ATSC/MPEG-2 3-mhz channel with one single output. Would the max Mbps rate be 9.685 (half of 19.39 for 6-mhz) ??

Thanx-A-Lot, Frank-0-Video

Greetings Trip ...

All right, suppose we had an ATSC/MPEG-2 3-mhz channel with one single output. Would the max Mbps rate be 9.685 (half of 19.39 for 6-mhz) ??

Thanx-A-Lot, Frank-0-Video

There's actually a bit of empty space at either end of the ATSC signal to provide a buffer so stations can be side-by-side, so it would be a little less than 9.685 Mbps. I doubt receivers would be able to handle it, though.

- Trip

I thought the whole mantra of this forum(and av enthusiasts) was to desire the elimination of subchannels, thus giving us better HD?
I must have repeated this 50 times to others that don't frequent A/V sites, that their HD would look better if locals didn't murder the bitrate?

Greetings Once More....

There's actually a bit of empty space at either end of the ATSC signal to provide a buffer so stations can be side-by-side, so it would be a little less than 9.685 Mbps. I doubt receivers would be able to handle it, though.

- Trip

Unless of course we were functioning in a reality where single-program 3-mhz channels were the norm, and we assume something less than 9.6 Mbps -AND- superior than present actual MPEG encoding plus other factors.

Thanx-A-Lot, Frank-0-Video

In theory, yes, but the other thing to keep in mind is that I'm pretty certain DVB-T requires a higher signal-noise ratio, and thus the extra bits are coming by reducing error correction. (Someone who knows more about it might want to check into that, as I'm not 100% sure. I remember seeing something about that somewhere.) The signal becomes harder to receive when you do that.

- TripIt appears that DVB-T allows a choice of error correction (FEC) rates (1/2, 2/3, 3/4, 5/6, 7/8),

http://en.wikipedia.org/wiki/DVB-T

whereas ATSC uses FEC=2/3. This would allow the DVB-T broadcaster to determine the bits assigned to error correction in a similar manner as DVB-S (satellite) does and seems to provide more flexibility than ATSC.

And as far as my experience with DVB-S goes you are right that reducing bits assigned to error correction makes it harder to receive a stable signal. Some FEC=7/8 DVB-S signals can be more difficult to receive than a FEC=3/4 signal, depending on the power output of the satellite transponder and receiving dish size.

dmatch

Los Angeles already has a station with 18-1, 18-3, 18-5, 18-6, 18-7, and 18-8 right now.

http://thumbnails7.imagebam.com/2571/c659d925703881.gif (http://www.imagebam.com/image/c659d925703881)

Could they get 10 channels in there?

LA also has KXLA - 44-1, 44-2, 44-3, 44-4, 44-5, 44-6 and 44-7. All SD, all Asian programming. It doesn't look quite as bad as you'd expect.

I thought the whole mantra of this forum(and av enthusiasts) was to desire the elimination of subchannels, thus giving us better HD?At least in this sub-forum it should be.

How far in the future? I'm sure the day will come when there is the possibility of of 2 HD subchannels of excellent quality in 6 MHz but I don't want to even think about how many delays might occur before that transition takes place.

I posted a thread in the CECB forum similar to what I'm going to ask here. CECB boxes were allowed by the NTIA to be upgradeable either through OTA or service ports. There are several boxes that have RS232 ports and some users have reported that their boxes have menus for downloading upgrades OTA.

Would it be possible for broacasters and manufacturers to send upgraded firmwares that could decode MPGE4? I assume that this will only possible for boxes using SoC integrated decoders as hardware MPGE2 decoder might be hard to make them decode MPGE4. Following this same line of thought, are current CECB boxes powerfull enough to decode MPGE4 just be simply upgrading the firmware?

Why are Sub-Channels needed anyways...

Just wondering

Why are Sub-Channels needed anyways...

Just wondering

They're not needed per say. They're mostly used as another revenue stream for stations.

All depends.

If you retain ATSC 8VSB then 19.2Mbs is still the limit for a 6MHz channel.

In Europe (well the UK) we're about to convert one nationwide 8MHz multiplex from 18Mbs 16QAM 2k DVB-T COFDM (though it previously ran 24Mbs 64QAM 2k DVB-T COFDM) to 36Mbs 256QAM 32k DVB-T2 COFDM. This is possible because we don't run on the One Station per Frequency model of the US - so stations and muxes are not 1:1 mapped (i.e. multiple stations can share portions of the same frequency multiplex) - and up until now all of our OTA digital stuff has been 16:9 SD or 4:3 SD only. We also have widespread use of rooftop aerials (antennas/antennae) so can use modulation schemes that are less robust than areas where set-top or internal aerials are widespread. We're also shifting to H264 for our DVB-T2 stuff - meaning we should be able to get 3 x 12Mbs H264 1080/50i streams into a single 8MHz channel at rpretty good quality if the new H264 encoders are used.

If you retain MPEG2 compression AND retain ATSC 8VSB modulation then you have a limited number of sub-channels before you start hitting quality issues - though that will also depend on the quality of the MPEG2 encoders used for the main and sub-channels, the quality of the source material (and the route it has taken to the final encoder) - and the main channel format.

In a 19.2Mbs stream - if you only ran SD content - then you could probably get 4-6 SD streams in at watchable quality (The UK gets this number into its 18Mbs muxes with a mix of 720x576 and 544x576 resolutions). If you are happy with NTSC-analogue quality you could drop the resolution from 704x480 to 528x480 and probably sqeeze 8 in.

HOWEVER - if you want an HD stream as well - then it is probably marginal whether you can squeeze any sub-channels in if you want a decent MPEG2 1080i signal to be carried at full resolution, and probably only one with a 720/60p stream (possibly two)

It also depends on the content of the sub-channel. If it is slow moving weather forecast imagery it will require far less data than a sports sub-channel or a music sub-channel.

So many variables - difficult to give a definitive answer.

Where I live I get channel 29 WFME. They have 9 subchannels which are all audio in mono. Also there main channel looks like crud.

I find subchannels in general a waste. If channel 7 WABC would get rid of its 7.2 subchannel which 90% of the time just shows infomercials they can bump up the bitrate of there main channel and make the 720p station look better.