new 8600 software offering SSB stereo

[fm-engineer]

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.

 

[fm-engineer]

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

 

[BabyDJ]

I saw AirAura at a NYC SBE 3 weeks ago and their box also has SSB/SC option. The presenter said it was in their V 1.2 software.

Funny how Orban and Vorsis have this, but the 11 doesn't yet (then again I don't think there is a finished 11 anywhere, except maybe on Superior Ave in Cleveland)

The Orban notes are interesting since Bob's scientific remarks seem valid against any market hype.

 

[fm-engineer]

Well, SSB is SSB, but how the overshoots are dealt with seems to be what will make or break the concept. The constant delay filters make sense, but the magic will probably be in the SSB look-ahead limiter, which would not be possible in analog. It appears SSB limiter must be used with the composite limiter as well. The L-R at the same level as the L+R is where the benefits of this system should surface.

 

[FFoti1]

Version: 1.0 software, for Omnia.11 will contain the SSB implementation. We're flattered by the reactionary mode of the other guys as this concept was brought to light by our company.

-Frank Foti

 

[k6sti]

In your implementation, what do you plan to do about the poor high-frequency separation the SSB method yields with analog receivers, Frank?

Brian

 

[FFoti1]

In your implementation, what do you plan to do about the poor high-frequency separation the SSB method yields with analog receivers, Frank?

Brian

Brian,

I have spoken to a few receiver manufacturers who have tested the SSB mode, with analog receivers and they do not share the claim about high-frequency separation that you made to me, prior, in a private dialog a few months ago. Also, they do not employ the type of testing you base your claim on.

To enlighten the group, you have a method of testing separation, in a receiver, using a non-standard method which does not take into account the entire transmission system. IIRC: Your testing does not even include a stereo generator. You are measuring only the wideband composite response at the output of the demodulator in the receiver, and then plotting that against the theoretical restored stereo separation response in the receiver.

As I've already explained, your method needs to be compared to a known industry testing method, and see if the results are the same. If they are, then you have a relevant alternative test method, for the receiver. I'd venture a guess you have not done this exercise, and I recommended that to you awhile back.

You need to compare your method with those that are of standard practice, before making your claim. As explained to you prior, we will give thought to your method of testing in our work with the NRSC. You are welcome to be involved with that.

From what I gather, the other implementations of the SSB stereo generator will operate much the same as our method, with regards to stereo separation performance.

It was decided at the last meeting of the NRSC to give the SSB transmission method further examination as a collective industry. The results from that will enlighten all of us about this topic. As I have said all along, if we find 'issues' then we'll know once and for all, as compared to blog debates which contain info based on non-standard testing.

-Frank Foti

 

[BabyDJ]

In your implementation, what do you plan to do about the poor high-frequency separation the SSB method yields with analog receivers, Frank?

Brian. You should probably pose this question to Bob Orban. I think he better understands the science behind the cause and effect of this process.

 

[k6sti]

Frank, you do not seem to understand my argument. It is simply this: Receivers with analog IF filters have a nonflat composite response. It slopes downward with frequency. The receiver's stereo decoder sees this response, not the flat spectrum transmitted, and it will cause channel separation at high audio frequencies to suffer when the upper stereo sideband is missing.

A sloping composite does not affect DSB stereo much because the USB and LSB response errors tend to cancel. Typically receivers are adjusted for best separation for 1-kHz audio. This means that the sum of the 37- and 39-kHz L-R responses equals the 1-kHz L+R response. When the signals are added and subtracted, one channel cancels. The receiver's stereo separation trimpot attenuates L+R to make it equal to L-R near 38 kHz to permit cancellation. For 15-kHz audio, the 53-kHz USB component will be somewhat low due to the composite slope, while the 23-kHz response will be somewhat high. But their sum remains about right so high separation is still possible.

With SSB, there is no USB to compensate the LSB. This won't make much difference for 1-kHz audio since the sloping response changes little from 37 to 39 kHz. Twice the amplitude of 37 kHz will still be about the same as the sum of the 37- and 39-kHz amplitudes. But for 15-kHz audio, the 23-kHz LSB component will not be the right amplitude to cancel the 15-kHz L+R signal. There is no longer a 53-kHz USB component to compensate.

I don't have an SSB stereo generator so I can't make any measurements. I'm just offering an argument that suggests SSB stereo will have lower separation for higher audio frequencies than DSB stereo. I have measured the the composite amplitude and phase for one of my tuners to get an idea of the magnitude of the problem. Using the response, I calculate that 15-kHz separation for the SSB method will be less than 30 dB for the tuner's wide IF filter and less than 20 dB for its narrow filter. These are the separations I would expect to see if a local broadcast station transmitted a SSB stereo signal. They are much lower than the 15-kHz separations the tuner is capable of providing for DSB stereo. The numbers would differ for receivers with different IF filters.

If you can spot an error in my reasoning, please tell me. I don't see any way around the problem except to transmit a rising L-R SSB spectrum. But this would provide an exact correction only for a single composite response and it would lower high-frequency separation for receivers with digital IF filters that yield a flat composite. So I don't think it's a good solution.

Brian

 

[k6sti]

A plot of the wide-IF composite response for the tuner I mentioned appears midway in this article:

http://ham-radio.com/k6sti/ibocfltr.htm

I wrote a little utility to do calculate separation given the composite amplitude and phase response. The ZIP file includes responses for the wide and narrow IF filters. It's here:

http://ham-radio.com/k6sti/SEP.ZIP

The utility is a DOS program. After unzipping, type SEP from the command line for instructions. You can use the response values in WIDE.TXT and NAR.TXT to manually calculate separation values to check the program.

Brian

 

[FFoti1]

In your implementation, what do you plan to do about the poor high-frequency separation the SSB method yields with analog receivers, Frank?

Brian. You should probably pose this question to Bob Orban. I think he better understands the science behind the cause and effect of this process.

I fully understood the question. Brian and I have had this discussion, in the past. My rebuttal has been that Brian needs to test his method up against known industry standards. His method is non-standard. He even mentions, in a later post, that he does not have a stereo generator. Thus, his basis is on a single-ended measurement, and he has not taken the entire system into consideration.

IF, what Brian claims is true, then this issue will effect *any* SSB implementation...All of them, including those of other competitors.

-Frank Foti

 

[FFoti1]

Frank, you do not seem to understand my argument. It is simply this: Receivers with analog IF filters have a nonflat composite response. It slopes downward with frequency. The receiver's stereo decoder sees this response, not the flat spectrum transmitted, and it will cause channel separation at high audio frequencies to suffer when the upper stereo sideband is missing.

A sloping composite does not affect DSB stereo much because the USB and LSB response errors tend to cancel. Typically receivers are adjusted for best separation for 1-kHz audio. This means that the sum of the 37- and 39-kHz L-R responses equals the 1-kHz L+R response. When the signals are added and subtracted, one channel cancels. The receiver's stereo separation trimpot attenuates L+R to make it equal to L-R near 38 kHz to permit cancellation. For 15-kHz audio, the 53-kHz USB component will be somewhat low due to the composite slope, while the 23-kHz response will be somewhat high. But their sum remains about right so high separation is still possible.

With SSB, there is no USB to compensate the LSB. This won't make much difference for 1-kHz audio since the sloping response changes little from 37 to 39 kHz. Twice the amplitude of 37 kHz will still be about the same as the sum of the 37- and 39-kHz amplitudes. But for 15-kHz audio, the 23-kHz LSB component will not be the right amplitude to cancel the 15-kHz L+R signal. There is no longer a 53-kHz USB component to compensate.

I don't have an SSB stereo generator so I can't make any measurements. I'm just offering an argument that suggests SSB stereo will have lower separation for higher audio frequencies than DSB stereo. I have measured the the composite amplitude and phase for one of my tuners to get an idea of the magnitude of the problem. Using the response, I calculate that 15-kHz separation for the SSB method will be less than 30 dB for the tuner's wide IF filter and less than 20 dB for its narrow filter. These are the separations I would expect to see if a local broadcast station transmitted a SSB stereo signal. They are much lower than the 15-kHz separations the tuner is capable of providing for DSB stereo. The numbers would differ for receivers with different IF filters.

If you can spot an error in my reasoning, please tell me. I don't see any way around the problem except to transmit a rising L-R SSB spectrum. But this would provide an exact correction only for a single composite response and it would lower high-frequency separation for receivers with digital IF filters that yield a flat composite. So I don't think it's a good solution.

Brian

Brian,

I fully understand your argument, without question.

For the record, so the group will be properly informed, you and I had this debate back in November. At that time, I told you there would be further review, and probable testing of the subject. That is about to begin via the NRSC.

I also told you that we would evaluate your test method, as well as look into your question. I fully understand it, and am not doubting it. If it is valid, we'll learn this as an industry.

Of importance though, is your non-standard method. I, professionally, request that you evaluate and compare your method up against known industry standards, which requires the use of the entire transmission path in order to evaluate stereo separation. If the results of stereo separation measurements, using a full transmission system approach, mirrors your method, then your argument stands to reason. But until you perform that exercise, you cannot validate the data.

If you don't have a stereo generator, I suggest you find one. If you would like to borrow one, I'll gladly loan it to you. We manufacture a DSP based stereo generator product.

I'll gladly debate the topic once you offer data that is comprehensive to what the industry utilizes. Thus far, you have not taken the time to validate your test method against industry standards.

-Frank Foti

BTW: I find it amusing that two of our competitors raced to get this idea running in their products. Thus, your 'argument' is with those who are employing the concept.

 

[oldiesstation]

Well,once we move past all the trade puffery,hype and spin,the indicator will be how it performs in the field .Vorsis Air Aura seems to be first to market with SSB innovation and engs that have heard this box say it amazing.Both of O's boxes are on par as well,so now it's just a matter of choice.But more field tests are needed to see how SSB plays out(Pardon the pun)..

 

[FFoti1]

Well,once we move past all the trade puffery,hype and spin,the indicator will be how it performs in the field .Vorsis Air Aura seems to be first to market with SSB innovation and engs that have heard this box say it amazing.Both of O's boxes are on par as well,so now it's just a matter of choice.But more field tests are needed to see how SSB plays out(Pardon the pun)..

Agreed. This needs to be tested, as compared to on-going blog debates. I've said this all along.

Actually, this concept is not innovative, as it's been discussed over the years. All I did was do some implementation work, and get some initial positive feedback to continue the effort.

If anything, we were the first to bring it to light recently, and the other competitors are now following in our footsteps. 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

 

[k6sti]

Frank, I'm not proposing a test method, standard or non. I'm just explaining why I think the SSB stereo method will result in lower high-frequency separation for tuners with an analog IF strip than does the present DSB method. The only assumption I'm making is that a perfect stereo signal is broadcast, SSB or DSB. In other words, I'm assuming the best case. Separation will be lower for imperfect exciters or propagation paths.

I have a good stereo generator, but of course it uses the DSB method, not SSB. I wrote a composite stereo generator that uses DSB or SSB, but I don't have the 192kb sound card required to generate an output signal. If anyone who does would like to try the program, it is here:

http://ham-radio.com/k6sti/stereo.exe

It is a DOS program. Type STEREO at the command prompt for instructions. The program creates a .WAV file. Play the file from a 192kb sound card connected to the modulation input of an FM signal generator. If you can measure it, the program will compensate for sound card or modulator roll-off.

Brian

 

[FFoti1]

Frank, I'm not proposing a test method, standard or non. I'm just explaining why I think the SSB stereo method will result in lower high-frequency separation for tuners with an analog IF strip than does the present DSB method. The only assumption I'm making is that a perfect stereo signal is broadcast, SSB or DSB. In other words, I'm assuming the best case. Separation will be lower for imperfect exciters or propagation paths.

I have a good stereo generator, but of course it uses the DSB method, not SSB. I wrote a composite stereo generator that uses DSB or SSB, but I don't have the 192kb sound card required to generate an output signal. If anyone who does would like to try the program, it is here:

http://ham-radio.com/k6sti/stereo.exe

It is a DOS program. Type STEREO at the command prompt for instructions. The program creates a .WAV file. Play the file from a 192kb sound card connected to the modulation input of an FM signal generator. If you can measure it, the program will compensate for sound card or modulator roll-off.

Brian

Brian,

To gather the data you mention, you had to the test the receiver. In essence your posted criteria is a test method. To validate this, it needs to be compared to a known standard.

We will evaluate this within the working group of the NRSC. I'll gladly share what we learn. Until then, might be best to move on...

-Frank Foti

 

[BabyDJ]

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?

 

[k6sti]

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

 

[fm-engineer]

Frank, Bob, et al.

I'm glad to see we are working on something to improve the actual medium that people are listening to. Too many years have passed where all of the focus has been on IBOC. We will probably not be shutting down the analog transmission in my lifetime. Thousands of analog radios are being built each day. The FM stereo system we use today remains unchanged for almost half a century. Anything to improve it should be a priority.

Now to my question: If these SSB generators are being written into code for experimental testing, will they be offered in the popular existing products? In other words, engineers like me could begin real world testing if we had the software. I have access to Orban and Omnia boxes all throughout some of the toughest multipath terrain you will find. If particular, there are fifteen or so Omnia 6's, Optimod 8300-8500's under my control. If these boxes are tapped out when it comes to DSP, will you offer SSB only software so testing can begin without buying new $15k boxes such as the Omnia 11? Just a thought you may want to entertain if you want some feedback from the trenches.

 

[fm-engineer]

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?