new 8600 software offering SSB stereo

[FFoti1]

It may be end up being passing fad, but you cannot argue with the logic of the increased L-R level being less susceptible to multipath effects. There is no harm with testing this technology. The things you can do with DSP now compared to ten years ago is amazing. We should all embrace these manufactures since they are working to improve the medium that 99.9% of radio listeners are using, analog radio.

If not for my public discussions on the topic, along with presentations at NRSC meetings, the other two guys wouldn't have a clue on this.




-Frank Foti

I don't think there has been a day where the likes of Bob Orban doesn't have a clue. I really think your comments here are inappropriate or do anything positive for your position. I don't really see anything innovative here and, just like surround FM, HD, FMX, Dolby, Quad and everything else, this will be a passing fad.

I still wonder why it took 12 years for someone to dust off this paper?

Well, until recently, the amount of processing power wasn't available, or it would have seen the light of day sooner. This is not new information. Also, this is the reason why it won't be possible to insert this into the older Omnia processors, with the *possible* exception of Omnia.One.

The goal here, would be a hopeful reduction of multipath due to the reduced bandwidth of the transmitted composite signal.

If we don't test these ideas, we'll never know.

-Frank Foti

 

[FFoti1]

Frank, I measured the composite response by feeding the tuner a monophonic FM signal either from an HP 8640B or a Sound Technology 1000A signal generator (I can't remember which one I used). I generated the audio test tone with an HP 3336C level generator. This equipment is flat over the 53-kHz composite passband, as verified with a Marconi 2305 modulation meter. I picked off the tuner composite response with a Tektronix P6202A active probe to avoid any circuit loading. I read the composite amplitude with an HP 3456A digital voltmeter. I read the waveform phase directly off the screen of my Tektronix 465A scope.

If you can see anything wrong with this procedure, please tell me.

Brian

Brian,

You've told me this in our prior private discussions in November. This is not new news. Further dialog on this topic needs to be reserved until your method is compared to known industry test standards, which account for the transmit portion. Your method does not do this. Until then, let's move on...

-Frank Foti

 

[k6sti]

Frank, perhaps you could explain what "known industry test standards" you're referring to. I don't have an LSB stereo generator so I can't directly measure separation. The best I can do is to measure the receiver composite response and calculate what the separation would be for a perfect transmit signal. If you can spot an error in my method or calculations, I'd appreciate knowing about it.

The other thing I've wondered about is the possible benefit. It sounds reasonable that a narrower L-R passband would reduce multipath distortion, but by how much? A little or a lot? I was going to write a simulation to see if I could get some idea, but after I read Bob Tarsio's paper on the negative results of his 1987 broadcast experiments using the SSB method in New York City I lost interest. Have you simulated the SSB method or run tests to measure the effect on multipath distortion? Do you have a reason to think your results might differ from Bob's?

Brian

 

[rorban]

Looks like you can experiment with SSB on the new Orban 8600 software. I wonder if Bob will implement this in the other current models of Optimod? I have a few 8300's in some multi-path prone terrain, but no 8600's yet. If anyone updates the 8600 and does some real world testing, please report back the findings on SSB stereo.

We might implement SSB on the 5500 in its stand-alone stereo generator mode. We have not evaluated whether SSB operation could be extended to other Orban processors. We hope that the new NRSC working group that is examining SSB operation will eventually obtain experimental results that will allow a definitive conclusion to be drawn about the relative merits of SSB and DSB operation. Once we have those conclusions and if they favor SSB, it will become appropriate for us to apply engineering resources to see if Orban's older designs can accommodate SSB.

There are a lot of variations on the SSB that could be tried. The first experimental release in the 8600 definitely needs improvement before SSB operation could complete in a major market with the conventional DSB processing. We are already looking at other ways of dealing with the peak overshoots in SSB mode caused by the loss of the interleaving that we have become accustomed to exploiting for modulation control in conventional FM stereo.

I wrote a white paper about the peak modulation of an SSB generator when fed by the same audio processing that is correct for controlling the peak modulation of a conventional DSB generator. The conclusion is that extra audio processing is necessary. The paper is here:

http://www.orban.com/support/orban/techtopics/Composite%20Peak%20Modulation%20of%20a%20Compatible%20SSB%20Stereo%20Encoder%20(rev%202).pdf

 

[rorban]

ftp://ftp.orban.com/8600/Documentation/8600_1.1.0_Readme.htm

One point that I left out of the readme is significant in countries that require compliance with the ITU-R BS.412 multiplex power limit. In the stereo subchannel, compatible SSB has 3 dB higher multiplex power than DSB. This is because the magnitude of the single sideband is twice that of each double sideband. Twice the magnitude is +6.02 dB, or four times the power present in the lower sideband in conventional DSB transmission. Meanwhile, the upper sideband in conventional DSB only adds 3 dB to the MPX power.

 

[FFoti1]

Frank, perhaps you could explain what "known industry test standards" you're referring to. I don't have an LSB stereo generator so I can't directly measure separation. The best I can do is to measure the receiver composite response and calculate what the separation would be for a perfect transmit signal. If you can spot an error in my method or calculations, I'd appreciate knowing about it.

The other thing I've wondered about is the possible benefit. It sounds reasonable that a narrower L-R passband would reduce multipath distortion, but by how much? A little or a lot? I was going to write a simulation to see if I could get some idea, but after I read Bob Tarsio's paper on the negative results of his 1987 broadcast experiments using the SSB method in New York City I lost interest. Have you simulated the SSB method or run tests to measure the effect on multipath distortion? Do you have a reason to think your results might differ from Bob's?

Brian

Brian,

The broadcast industry has always performed stereo separation tests using the entire transmission chain. This includes the stereo generator, STL (if used), exciter, transmitter, antenna, and then the modulation monitor and/or reference receiver for demod purposes.

Through testing, we'll learn if SSB transmission offers a reduction in multipath, and if so, how much. I've talked to, and read Bob Tarsio's papers. He offers good information to be considered. He also told me his testing was done prior to what we can do today, with digital based modulators.

The whole process of review and testing, via the NRSC working group is to be able to determine if there's a benefit here for the broadcast industry. Your input may be of value to the group, and I welcome you to consider getting involved.

If you wish to do so, you can get the information from the NRSC website.

-Frank Foti

 

[k6sti]

Frank, I checked the NRSC web site (everything available to a nonmember) and I can find no mention of SSB stereo.

When I measured the composite response of my tuner, I fed it a standard monophonic FM test signal. I used a single modulating tone with constant deviation over the 53-kHz composite passband. Isn't this is what "the entire transmit chain" will deliver at the receiver antenna terminals under similar test conditions? I don't see why it's necessary to include every piece of transmission equipment to validate a measurement if I reproduce the signal that equipment will generate. When a laboratory tests an FM receiver or a manufacturer aligns one, they don't use a complete broadcast station that radiates a signal on the air. They use local test equipment that generates an equivalent RF signal at the antenna terminals. I do the same thing. I'm not attempting to test the transmit equipment. I assume it's working normally. I'm trying to determine how a typical FM stereo receiver will respond. I'm familiar with stereo FM transmission standards, but if there's some subtle aspect I've overlooked and should include in my test signal, please tell me.

Brian

 

[FFoti1]

If not for my public discussions on the topic, along with presentations at NRSC meetings, the other two guys wouldn't have a clue on this.

-Frank Foti

I don't think there has been a day where the likes of Bob Orban doesn't have a clue. I really think your comments here are inappropriate or do anything positive for your position. I don't really see anything innovative here and, just like surround FM, HD, FMX, Dolby, Quad and everything else, this will be a passing fad.

I still wonder why it took 12 years for someone to dust off this paper?

Actually, my comments a quite appropriate, as they are accurate.

-Frank Foti

 

[FFoti1]

Frank, I checked the NRSC web site (everything available to a nonmember) and I can find no mention of SSB stereo.

When I measured the composite response of my tuner, I fed it a standard monophonic FM test signal. I used a single modulating tone with constant deviation over the 53-kHz composite passband. Isn't this is what "the entire transmit chain" will deliver at the receiver antenna terminals under similar test conditions? I don't see why it's necessary to include every piece of transmission equipment to validate a measurement if I reproduce the signal that equipment will generate. When a laboratory tests an FM receiver or a manufacturer aligns one, they don't use a complete broadcast station that radiates a signal on the air. They use local test equipment that generates an equivalent RF signal at the antenna terminals. I do the same thing. I'm not attempting to test the transmit equipment. I assume it's working normally. I'm trying to determine how a typical FM stereo receiver will respond. I'm familiar with stereo FM transmission standards, but if there's some subtle aspect I've overlooked and should include in my test signal, please tell me.

Brian

Brian,

I fully understand your test method. Theoretically, it makes sense, and I will perform this comparison test myself, when time permits. My point, in this discussion, is to see a comparison of your method up against the entire transmission chain, which includes the stereo generator, and modulator. The following links indicate the receiver separation test includes a function of the transmission portion:

http://www2.rohde-schwarz.com/file_6116/RAC-0502-0012.pdf
http://wireless.agilent.com/wireless/helpfiles/n7611b/fm_stereo_measurement.htm

I'd venture a guess, they do this to include any possible anomalies that occur in the stereo generation and modulation process. One might assume, it's all a linear process, but as the real world tells us, it doesn't always play out that way. Using this eliminates the assumptions you admit, that the transmit side is functioning as it would do so, in a theoretical manner.

Regarding NRSC, I'll look into what's going on there.

-Frank Foti

 

[BabyDJ]

Actually, my comments a quite appropriate, as they are accurate.

-Frank Foti

My point is validated after I spent time reading Bob's paper this weekend on DSB/SSB comparisons. While I still think the cure is worse than the disease (or, better yet, causes other diseases), I still think any benefit that may be presented here was done so more as a marketing pitch than a benefit to broadcasters, and that's what bothers me.

Your competition is better than ever as is your new product and I understand the need for some type of edge, but if this were as great as it seems, something would have been done over the past 12 years by Bob, yourself or someone else to make it happen. Christ, y'all added a second stage of processing for HD to the 6, ITU limiters etc.

 

[cgould]

Actually, my comments a quite appropriate, as they are accurate.

-Frank Foti

My point is validated after I spent time reading Bob's paper this weekend on DSB/SSB comparisons. While I still think the cure is worse than the disease (or, better yet, causes other diseases), I still think any benefit that may be presented here was done so more as a marketing pitch than a benefit to broadcasters, and that's what bothers me.

Your competition is better than ever as is your new product and I understand the need for some type of edge, but if this were as great as it seems, something would have been done over the past 12 years by Bob, yourself or someone else to make it happen. Christ, y'all added a second stage of processing for HD to the 6, ITU limiters etc.

I don't recall Frank ever saying the idea was perfect, and ready to go now. This is something he has put forth for consideration now that we processor manufacturers have the ability to easily implement this in DSP.

We're experimenting with various techniques, as are the other guys. Whichever one works best, we'll all evaluate and will determine what benefits there are. Is it something appropriate for an across the board implementation by every broadcaster? Probably not...but there may be cases where this *could" be a benefit for some broadcasters.

More testing is needed, so that's where the NRSC comes in. Brian and others have made some interesting observations that need to be looked into by the committee. That's what the NRSC is there for - to take in all this data, and make an informed choice amongst peers of all levels and interests in the field.

That's all we're trying to say.

What seems to irk you to no end was the fact that we started the noise by officially announcing the idea to the general public as a company. Nothing we can do about your personal feelings on this...this is an issue for you to work out.

-C

 

[FFoti1]

Actually, my comments a quite appropriate, as they are accurate.

-Frank Foti

My point is validated after I spent time reading Bob's paper this weekend on DSB/SSB comparisons. While I still think the cure is worse than the disease (or, better yet, causes other diseases), I still think any benefit that may be presented here was done so more as a marketing pitch than a benefit to broadcasters, and that's what bothers me.

Your competition is better than ever as is your new product and I understand the need for some type of edge, but if this were as great as it seems, something would have been done over the past 12 years by Bob, yourself or someone else to make it happen. Christ, y'all added a second stage of processing for HD to the 6, ITU limiters etc.

Nothing marketing related at all. If this was marketing driven, the approach would be much different, as it would be positioned as a product advantage by one over another. That's not the case here. I expected the other guys to implement the concept. They both must feel there's some value by adding it, or they would not have invested the man-hours to do this work.

I'll take your comment about marketing pitch to apply to our competition too, as they raced to implement this. I know Bob Orban well enough that if he didn't feel there was at least something to learn over the idea, he wouldn't invest the time in it. I know he has done this, and has told me so personally. If you feel this is marketing driven, then your lament must be towards all whom are offering this tech.

I sense you haven't read the initial paper by William Gillman, or what was published by Radio World last year. The intended goal is to reduce the perception of multipath. This is clearly stated in both papers. We'll know more, whether this works or not, once we field test the method, along with the upcoming work via the NRSC working group. There's value in finding out the answer based on actual real world testing, as compared to judging one concept versus another solely on white papers. I've clearly stated the SSB method needs to be tested before drawing any conclusions. I have not drawn any conclusion, and yes I hope something positive does come out of it. That's stated in the white paper and Radio World article.

Seems it would be a shame if we were to discount an idea, based on tech papers alone, and if the concept does offer benefit, then our industry would have suffered needlessly. As it is, broadcasting needs all the help it can get, and if this reduces real world sonic annoyance to the FM listener, then all of us win.

On account of adding the second stage of processing for HD, that used up the remaining DSP cycles that could've been used for the SSB modulator. Additionally, there were years spent by all of us (processing, STL, and transmitter companies) working to see that HD Radio transmission systems were implemented in the best manner possible. Modestly speaking, the idea of inserting diversity-delay into the audio processor was conceived by us. It wasn't done for marketing purposes. It was done to reduce the 'goofy' system architecture iBiquity designed into HD Radio installations. It also reduced points of failure in the overall system. It made sense that all processor manufactures offer this, as it helped create a common system integration method. This happened, and a lot of broadcast engineers were duly grateful.

I've had the SSB concept on my To-Do list since Bill Gillman gave me his paper in 1997. Had we the resources available, both in time and DSP cycles, it would have seen the light of day sooner. If I'm reading your comments correctly, you feel I'm wrong for bringing this to light now.

IMO: I sense you have a non-positive bias towards us. Not sure why. You're making strong claims against this topic, yet I noticed you posted positive comments regarding the competitions implementation of the SSB concept you feel is all marketing driven.

-Frank Foti

 

[FFoti1]

My point is validated after I spent time reading Bob's paper this weekend on DSB/SSB comparisons. While I still think the cure is worse than the disease (or, better yet, causes other diseases), I still think any benefit that may be presented here was done so more as a marketing pitch than a benefit to broadcasters, and that's what bothers me.

I forgot a comment regarding the above...

How do you support the notion of "the cure is worse than the disease (or, better yet, causes other diseases)?" You're making a claim based upon an opinion in a white paper, as compared to gathering information based on real world testing. Maybe this is true. But, we won't know if it is, or not, unless we test it.

-Frank Foti

 

[BabyDJ]

I'll take your comment about marketing pitch to apply to our competition too, as they raced to implement this. I know Bob Orban well enough that if he didn't feel there was at least something to learn over the idea, he wouldn't invest the time in it. I know he has done this, and has told me so personally. If you feel this is marketing driven, then your lament must be towards all whom are offering this tech.

As I stated I don't feel there will be any advantage to this system unless you make some other tradeoffs. Granted everyone makes tradeoffs in processing, but if one wants to reduce multipath, there are better ways of achieving it with the gear we currently have available to us.

I just found it ironic that I have not seen it in the 11 and that the other brands are out with it. I have a hunch that since it took some time to develop it that the other guys were working on this before your article last September IIRC.

I sense you haven't read the initial paper by William Gillman, or what was published by Radio World last year. The intended goal is to reduce the perception of multipath.

After much searching I did locate Gillman's paper, which basically applied to television.

IMO: I sense you have a non-positive bias towards us. Not sure why. You're making strong claims against this topic, yet I noticed you posted positive comments regarding the competitions implementation of the SSB concept you feel is all marketing driven.

-Frank Foti

Honestly, I'm not a big fan of your processors and have been hounded into trying to get an 11 by my management because of SSB. I have a 6, an 8400 and two Vorsis AM processors and wanted my own choices to replace my older gear. I'm actually happy that there are others that have come out with it.

I've spun the Telos wheel alot in my career and think it's, right now, the RCA of this generation... everyone seems to settle on it. But I'm always looking for something that takes it to the next level. If it's Orban or Breakaway or Vorsis or whatever.

 

[rorban]

I forgot a comment regarding the above...

How do you support the notion of "the cure is worse than the disease (or, better yet, causes other diseases)?" You're making a claim based upon an opinion in a white paper, as compared to gathering information based on real world testing. Maybe this is true. But, we won't know if it is, or not, unless we test it.

-Frank Foti

The white paper I wrote contains a calculation proving that a certain input signal (two sinewaves, identical expect for a 90 degree phase shift, applied to the L and R channels) when peak-limited using techniques appropriate for DSB modulation, causes peak modulation of about 138% when using compatible SSB. The calculation agrees with the measurement shown in the white paper. I therefore concluded that the SSB modulator needs more peak limiting than the DSB case to accommodate cases like this. Moreover, I observed experimentally that real-world program material when peak-limited correctly for DSB operation also causes very substantial overmodulation in the SSB case. Our SSB modulator uses filters with no group delay distortion, so it is essentially "ideal." Because my conclusion is backed up by theory and measurements that verify the theory, I would call it way more than an "opinion."

It may turn out to be possible to modify the system and/or improve the extra audio processing so that the DSB and SSB transmissions sound essentially equal on-air. This is still something we are investigating. One possibility is to use DSB from 0 to around 2 kHz (a frequency band where a significant amount of program energy resides) and SSB above 2 kHz. This would extend the baseband spectrum to 40 kHz instead of 38 kHz, as pure SSB would. However, it would ease audio processing requirements for SSB overshoots and would also reduce the 3 dB multiplex power penalty that pure SSB introduces.

Bob Orban

 

[BabyDJ]

The white paper I wrote contains a calculation proving that a certain input signal (two sinewaves, identical expect for a 90 degree phase shift, applied to the L and R channels) when peak-limited using techniques appropriate for DSB modulation, causes peak modulation of about 138% when using compatible SSB. The calculation agrees with the measurement shown in the white paper. I therefore concluded that the SSB modulator needs more peak limiting than the DSB case to accommodate cases like this. Moreover, I observed experimentally that real-world program material when peak-limited correctly for DSB operation also causes very substantial overmodulation in the SSB case. Our SSB modulator uses filters with no group delay distortion, so it is essentially "ideal." Because my conclusion is backed up by theory and measurements that verify the theory, I would call it way more than an "opinion."

When you add the already dynamic challenged audio we are putting into these boxes, I dare say the last thing I would want is added limiting. No matter how transparent or look ahead you could make it, the effects will be noticed.

I found your paper very enlightening and thank you for expanding on this idea, the good, bad and ugly.

 

[BabyDJ]

What seems to irk you to no end was the fact that we started the noise by officially announcing the idea to the general public as a company. Nothing we can do about your personal feelings on this...this is an issue for you to work out.

-C

I just seem to think you've taken it as your idea and not the original writer. It's funny, not only did Gillman think about this but US patent 7295628 by AKG discusses a way of depending on the aliasing in the upper sideband to fold it over. I have not seen the paper per se, but I've been told about it and waiting to get a copy.

 

[FFoti1]

I'll take your comment about marketing pitch to apply to our competition too, as they raced to implement this. I know Bob Orban well enough that if he didn't feel there was at least something to learn over the idea, he wouldn't invest the time in it. I know he has done this, and has told me so personally. If you feel this is marketing driven, then your lament must be towards all whom are offering this tech.

As I stated I don't feel there will be any advantage to this system unless you make some other tradeoffs. Granted everyone makes tradeoffs in processing, but if one wants to reduce multipath, there are better ways of achieving it with the gear we currently have available to us.

I just found it ironic that I have not seen it in the 11 and that the other brands are out with it. I have a hunch that since it took some time to develop it that the other guys were working on this before your article last September IIRC.

I sense you haven't read the initial paper by William Gillman, or what was published by Radio World last year. The intended goal is to reduce the perception of multipath.

After much searching I did locate Gillman's paper, which basically applied to television.

IMO: I sense you have a non-positive bias towards us. Not sure why. You're making strong claims against this topic, yet I noticed you posted positive comments regarding the competitions implementation of the SSB concept you feel is all marketing driven.

-Frank Foti

Honestly, I'm not a big fan of your processors and have been hounded into trying to get an 11 by my management because of SSB. I have a 6, an 8400 and two Vorsis AM processors and wanted my own choices to replace my older gear. I'm actually happy that there are others that have come out with it.

I've spun the Telos wheel alot in my career and think it's, right now, the RCA of this generation... everyone seems to settle on it. But I'm always looking for something that takes it to the next level. If it's Orban or Breakaway or Vorsis or whatever.

Glad to know your bias against us is public. We still send love your way BabyDJ.

If you don't feel SSB will make any difference, then don't use it. SSB will be an optional selection in the stereo generator, and no one is forced to use it. Please enlighten us how we can reduce multipath - with existing gear - that will also alter the modulation index of the transmission, which is a significant factor in the cause and effect of multipath?

I've said all along our SSB implementation will be in our version 1.0 software release which is coming quite soon. No need to cast false allusions.

Our paper was written based on our implementation that runs realtime, and all of our demonstrations were done using a hardware prototype. We had this going over two years ago. The irony is the other guys followed our lead once it got considerable attention via the article and overwhelming response at the NRSC meeting in Wash, DC last year. Also, I knew there would be a filtering out process on the topic from the industry. Therefore we pushed the release back to the version 1.0, as we're finalizing some other items first, then we will integrate SSB into the stereo generator.

The Gillman paper mentions how the method can also be used for television. It was written specifically for FM radio broadcasting. Bill told me this when he gave me the paper at NAB 1997.

If your management is hounding you about testing Omnia.11, please have them contact me directly, and I'll gladly take the heat they're sending your way. Regarding your view of our company as a whole, I'm willing to sit down and discuss this with you. Name the time and place. I enjoy visiting NYC and/or the tri-state area.

-Frank Foti

 

[oldiesstation]

Seems the Vorsis Air Aura had this feature over 2 months ago.Did not toot their horn over it.Maybe you could hound your management into taking one for a test drive.

 

[FFoti1]

What seems to irk you to no end was the fact that we started the noise by officially announcing the idea to the general public as a company. Nothing we can do about your personal feelings on this...this is an issue for you to work out.

-C

I just seem to think you've taken it as your idea and not the original writer. It's funny, not only did Gillman think about this but US patent 7295628 by AKG discusses a way of depending on the aliasing in the upper sideband to fold it over. I have not seen the paper per se, but I've been told about it and waiting to get a copy.

This is totally wrong! Now your comment is inappropriate. If anything, we went out of our way to credit Bill Gillman for his paper. Also, there are hundreds who've heard my public presentations on this topic, and each time I have given credit to where credit is due. What I have said, is we were the first to offer a public demonstration of the method.

-Frank Foti