NEW FM PROCESSOR

[gtomas]

I'd love to get some more MPX clips of DSPXtreme and DSPXmini, would be great to see if you can tweak away some of the distortion and pumping I hear in the current torture test clips -- you could really gain an edge on O&O then, beyond just price. Particularly the Evanescence cut in DSPXmini, something really strange is going on there across the board (listen at 00:14 for example).

There will always be little things that you can point out, with any processor. I could find at least a dozen for each of the processors, if I was to thin-slice those clips. But it's how noticeable they really and how often do these occur on long term listening, in a big picture, that make them a real problem or just an unnoticeable "fluff". Of course, that doesn't mean we are not working on further improving our processors.

But I strongly believe we should leave it to the listeners to judge those clips. We as designers are way to biased...


Regards,
Goran Tomas

 

[konbaasiang]

Hi Goran!

I'd really like the thin-sliced nit-picking list of the problems you hear in the BBP (Plutonium) clip! It's probably like you say, a case of me not seeing the forest for all the trees -- I was under the impression that I had taken care of all the nit-picking things already, both front-end and back-end.

If you could send this to me I would be most appreciative.

Best regards,
///Leif

 

[rrsounds]

I'd love to get some more MPX clips of DSPXtreme and DSPXmini, would be great to see if you can tweak away some of the distortion and pumping I hear in the current torture test clips -- you could really gain an edge on O&O then, beyond just price. Particularly the Evanescence cut in DSPXmini, something really strange is going on there across the board (listen at 00:14 for example).

There will always be little things that you can point out, with any processor. I could find at least a dozen for each of the processors, if I was to thin-slice those clips. But it's how noticeable they really and how often do these occur on long term listening, in a big picture, that make them a real problem or just an unnoticeable "fluff". Of course, that doesn't mean we are not working on further improving our processors.

But I strongly believe we should leave it to the listeners to judge those clips. We as designers are way to biased...


Regards,
Goran Tomas

Agreed. Some effects are so minimal, or random that they become insignificant in the larger scheme of things. On the other hand some effects are pretty obvious, but that doesn't necessarily mean a project should be scrapped because they are noticeable to 'insiders.'

For instance, quite a few people actually *like* the effect we call 'pumping.' So, while there are those of us who are doing everything we can to eliminate the effect altogther, it turns out a segment of the customer base has fallen in love with it...

As many of us have said time and time again, it all comes down to taste. My loathed 'artifact' can be someone else's favorite effect.

Few of us have the financial backup to do critical laboratory-grade testing of an algorithm with a large number of listeners. But, IMHO, it's not necessarily a bad thing if an industry-educated niche-product manufacturer designs to please him/herself individually or corporately, with the (hopefully) informed confidence that a significant portion of the audience will agree, and that financial success will follow. IOW, if you believe in it, why not go for it!

I like the abundance of smaller processor developers these days (I'm one of them). The design resources available to us widen every day. I believe the smaller design groups may be in a better position to take chances, because if they fail they don't risk taking down a large organization with them. And if they succeed, an astute larger group may see an opportunity to 'scoop up' the smaller one.

Kind Regards,
David

 

[cgould]

As many of us have said time and time again, it all comes down to taste. My loathed 'artifact' can be someone else's favorite effect.

Hehe. I was having the same discussion off-list.

-Cornelius

 

[jesseg]

I was playing around some with Jill FM in the GUI, and ended up getting a pretty darn good AC sound out of it, and I think I'll share it.

http://ictybtihky.com/bafm/Jill FM AC Tweak.png

Adjust Speed & Range as desired.

MPX FLAC:
Chicago - Look Away - BBP Jill FM AC Tweak 75uS MPX.flac

MP3 (for the lazy):
Chicago - Look Away - BBP Jill FM AC Tweak 75uS MPX - MPX Tool Output.mp3

 

[fl-lpfm]

Question about installation. I couldnt find it in this thread. My transmitter has a composite input and a mono input. will i install the output of the sound card to the mono xlr input to the transmitter or will a i need to create a composite cable to connect the sound card to the composite input of my transmitter?

Thanks

 

[konbaasiang]

Howdy!

What is the difference between a Composite Input and a Mono Input? Generally, they're the same. However, it could be that the Mono Input has built in pre-emphasis and the Composite Input doesn't.

If so, you'll have to use the composite input, but the cable is actually very simple. All it takes is a standard 1/8" plug to 2xRCA plug cable, and an RCA female to BNC adapter. Plug one of the RCA plugs into the adapter, wrap some electrical tape around the other one. That's actually all it takes!

I'm getting ready for the 1.0 release by the way. Stay tuned .

///Leif

 

[SUPERCASTER]

A suggestion, Leif.

Include an FMeXtra digital SCA composite stereo signal(s) soundcard output from BBP and you'll have a worldwide HD radio beating product. For you, this should be relatively simple.

FMeXtra is non proprietary, uses virtually the same aac plus codec as HD radio, does not require new FM transmision equipment (just the composite encoder), is simpler, more reliable, is totally compatible with existing FM standards, has much greater coverage then HD radio (about equal to FM stereo) and does not jam adjacent channels with digital hiss, as HD radio does.

The only disadvantage is that FMeXtra radios are not yet widely available.

For FMeXtra info:

www.dreinc.com

Thanks again for your great Breakaway Broadcast Processor.

 

[gtomas]

FMeXtra is non proprietary, uses virtually the same aac plus codec as HD radio, does not require new FM transmision equipment (just the composite encoder), is simpler, more reliable, is totally compatible with existing FM standards, has much greater coverage then HD radio (about equal to FM stereo) and does not jam adjacent channels with digital hiss, as HD radio does.

FMeXtra uses MPEG-4 standard HE-AAC codec (aka aacPlus). HD Radio doesn't. It uses proprietary HDC codec which also employs SBR, but these two codecs are not the same and they don't sound the same. According to the users, HDC seems not quite as good as HE-AAC.


Regards,
Goran Tomas

 

[konbaasiang]

Absolutely correct. The "HDC" codec is Lucent's PAC codec and Coding Technologies SBR process together. It's nowhere near aacPlus in quality. As far as I understand it, they went that route for political reasons (someone influential probably has a financial stake in the use of the PAC codec).

It's a real shame. FM Stereo has survived for how many years now? Over 40 for sure.

Did anyone anticipate HD radio surviving for even 10 years? Digital codecs may not evolve as rapidly as CPU speed, but at least they could have picked the best one that was available at the time!

There was talk about HD Radio coming to Thailand, but then the coup happened, and I haven't heard a thing about it since, so looks like I don't have to worry about it .

But, I digress...

I would LOVE to have FMeXtra support! It's a very promising standard.

However, there is one big stumbling block for FMeXtra on a PC which cannot easily be overcome:

While an analog MPX signal, even with RDS, does not extend beyond 60 kHz, the digital subcarrier required for FMeXtra extends all the way up 99 kHz.

This is simply too high for 192 kHz sound cards. 192 kHz sound cards can do 80 kHz *at best*, but they are often significantly rolled off even at this frequency. We’d need a sound card to run at 256 or 384 kHz to meet the basic requirements. After that, it’s only a matter of having the correct algorithms in the software, but I have never seen a sound card that support 384 kHz sampling rate -- at least not an affordable one.

///Leif

 

[fl-lpfm]

Hello,

Thanks for the quick reply. For how long will the trial work? I will like to test this on the air.

 

[BobOnTheJob]

What is the difference between a Composite Input and a Mono Input?

Mono inputs are generally 600 ohm balanced & designed to pass up to 15khz. Composite inputs are higher impedance (20k+?) and pass the entire stereo spectrum to at least 53khz...generally they are capable of passing higher frequencies than that. But for best stereo separation, the composite input must have equal response (plus or minus very few tenths of a db) from 53khz on down to the lower limit of audible hearing.

 

[konbaasiang]

fl-lpfm, the trial runs forever, but plays a commercial for breakaway broadcast, roughly once every 40 minutes.

The easiest way might be to set up your backup exciter fed by Breakaway Broadcast on a different frequency. This way you could switch back and forth between listening to your normal transmitter and the backup exciter with BBP, and you'll hear exactly the difference it will make, without the commercial interfering with your regular broadcast.

BobOnTheJob, thank you!

///Leif

 

[BobOnTheJob]

Playing with Personal edition again & I see something I don't quite understand. When I apply bass boost, the amount of multi-band gain reduction on the lowest band does not seem to increase, but the VU meters on my audio console go higher along with the bass boost. In my world, any kind of EQ needs to be applied before the multi-band stage. Leif or Jesse, can you enlighten me as to why that doesn't seem to be the case here?

 

[jesseg]

Playing with Personal edition again & I see something I don't quite understand. When I apply bass boost, the amount of multi-band gain reduction on the lowest band does not seem to increase, but the VU meters on my audio console go higher along with the bass boost. In my world, any kind of EQ needs to be applied before the multi-band stage. Leif or Jesse, can you enlighten me as to why that doesn't seem to be the case here?

It's after the multiband but before the bass clipper. There's a bass clipper on bands 1 & 2 (which never change crossover frequency btw). Drive adjusts the output drive of the bottom two bands, and shape lowers & raises one or the other of bands 1 and 2, plus tunes an extra parametric EQ which IS before the multiband. The EQ isn't much, it's just enough to shape the balance of the tone of the bass by a few db before the multiband, so it's not something that'll be readily apparent if you're not running test tones.

Also, the final drive can change the type of bass clipper, which audibly mainly is changing it's aggressiveness. Above +3.0db the clipper will go up another level. Most of the presets use Type1 by default which is the most musical and generates the least harmonics. So above +3.0db they will switch to Type2 (or3) which is more controlled and generates more harmonics. And nearing +6db, they will all use Type3 which is near brick-walled and generates lots of harmonics on lots of things.

[edit]
the exception to this rule is the protection clip presets.
[/edit]

The final 2-band limiter is also part of the process.

Final Drive also changes the threshold of the bass clipper (lower), and hence the drive into the 2-band (less), and the drive into the final clipper (different). More final drive = more bass clipper clipping. Generally this is a good thing because the reduced peaks are compensated by harmonics and the final clipper is just going to reduce the real bass to make room for the treble if it needs to anyways. The nice thing is it does the reducing in the clipper, not in safety limiters. In fact there are no safety limiters of any kind in the bass clipper or final clipper, other than what the preset designer does with multi-band 1-2 limiters and band1 of the final limiters.
;D
So this ability generates next to zero undesired IMD, and is MUCH more accurate in what it's reducing than processors that use protection limiters. A lot of the loudness on-air comes from this revolutionary level of precision in the final clipper.

New York preset is a great example to listen to, if you wanna hear the final clipper reduce certain aspects of bass when it needs to. It's definitely on the extreme edge of what I would ever consider putting on air, already at 0db, but hey... that's NYC (and they go a LOT farther), yet the preset ends up perceptively sounding a good 1-2db louder than most "cranked" 8500s - and they are clipping composite.

 

[BobOnTheJob]

Playing with Personal edition again & I see something I don't quite understand. When I apply bass boost, the amount of multi-band gain reduction on the lowest band does not seem to increase, but the VU meters on my audio console go higher along with the bass boost. In my world, any kind of EQ needs to be applied before the multi-band stage. Leif or Jesse, can you enlighten me as to why that doesn't seem to be the case here?

It's after the multiband but before the bass clipper. There's a bass clipper on bands 1 & 2 (which never change crossover frequency btw). Drive adjusts the output drive of the bottom two bands, and shape lowers & raises one or the other of bands 1 and 2, plus tunes an extra parametric EQ which IS before the multiband. The EQ isn't much, it's just enough to shape the balance of the tone of the bass by a few db before the multiband, so it's not something that'll be readily apparent if you're not running test tones.

Also, the final drive can change the type of bass clipper, which audibly mainly is changing it's aggressiveness. Above +3.0db the clipper will go up another level. Most of the presets use Type1 by default which is the most musical and generates the least harmonics. So above +3.0db they will switch to Type2 (or3) which is more controlled and generates more harmonics. And nearing +6db, they will all use Type3 which is near brick-walled and generates lots of harmonics on lots of things.

[edit]
the exception to this rule is the protection clip presets.
[/edit]

The final 2-band limiter is also part of the process.

Final Drive also changes the threshold of the bass clipper (lower), and hence the drive into the 2-band (less), and the drive into the final clipper (different). More final drive = more bass clipper clipping. Generally this is a good thing because the reduced peaks are compensated by harmonics and the final clipper is just going to reduce the real bass to make room for the treble if it needs to anyways. The nice thing is it does the reducing in the clipper, not in safety limiters. In fact there are no safety limiters of any kind in the bass clipper or final clipper, other than what the preset designer does with multi-band 1-2 limiters and band1 of the final limiters.
;D
So this ability generates next to zero undesired IMD, and is MUCH more accurate in what it's reducing than processors that use protection limiters. A lot of the loudness on-air comes from this revolutionary level of precision in the final clipper.

New York preset is a great example to listen to, if you wanna hear the final clipper reduce certain aspects of bass when it needs to. It's definitely on the extreme edge of what I would ever consider putting on air, already at 0db, but hey... that's NYC (and they go a LOT farther), yet the preset ends up perceptively sounding a good 1-2db louder than most "cranked" 8500s - and they are clipping composite.

Thanks for the in-depth analysis...much appreciated. I'm trying to follow the audio trail & wanted to make certain that this explanation applies to the Personal version before I dive any deeper into the thought process.

 

[jesseg]

Thanks for the in-depth analysis...much appreciated. I'm trying to follow the audio trail & wanted to make certain that this explanation applies to the Personal version before I dive any deeper into the thought process.

It does with the exception of the final clipper, which Breakaway Personal does not have, and the final 2-band limiter is tuned differently too.

Of course Breakaway Personal (unlike Broadcast and Live) doesn't not have phase-linear filters, so you can have things happen like pre-clipped program material getting the square-peaks rotated down the sides of the waveforms, and then re-clipped by the bass clipper, and re-brickwall-limited. But that's the nature of the low-latency game (almost all other processors do it too), and it's after all only $30.

 

[gtomas]

Of course Breakaway Personal (unlike Broadcast and Live) doesn't not have phase-linear filters, so you can have things happen like pre-clipped program material getting the square-peaks rotated down the sides of the waveforms, and then re-clipped by the bass clipper, and re-brickwall-limited. But that's the nature of the low-latency game (almost all other processors do it too), and it's after all only $30.

If you have an active phase rotator in the processor, it doesn't matter if the filters have phase linear response (or how much latency they produce). As soon as your square wave resembling program material passes through a phase rotator, it's all gone. Bob and Frank elaborated on that quite well.


Regards,
Goran Tomas

 

[konbaasiang]

Depends on what frequencies you rotate, Goran.

Traditional clippers seem to have problems with voices, thus traditional low frequency phase rotation (generally all-pass filters centered around 200hz) is necessary. That kind of phase rotator makes regular speaking voices nice and symmetrical.

My clipper isn't a traditional clipper, and doesn't have that problem. Instead, the phase scrambler (not the same as a phase rotator) in BBP affects high frequencies. You can see the effect very clearly by running a squarewave through and toggling the phase scrambler off and on. The edges get completely scrambled, the timbre changes slightly, but the flat top stays put. What *really* gains an advantage from this type of empirically developed phase scrambling is sounds like saxophones, sharp singing voices, sharp electronic instruments and other HF-harmonic rich instruments. These usually have a tremendous amount of IM through traditional processors, with or without phase rotation, and come through much cleaner in BBP.

///Leif

 

[gtomas]

If you don't use low frequency phase rotation to symmetrize (male) voices, my comment doesn't apply. Other processors have phase rotation option not because "they have problems with voices", but because there are noticeable gains in reducing required speech limiting with fairly small audible penalties.

I was trying to point out that even if you have phase linear crossover(s) (and most of today's digital processor on the market have) as soon as you turn on the phase rotation, that fact becomes pretty much irrelevant. As far as square waves are concerned...


Regards,
Goran Tomas