Forum Replies Created
-
Posts
-
Wow, I will definitely take you up on that! I’ve never been to Greece — this could be a lot of fun!
What airport is closest to you?
///Leif
Peatgy, I don’t know what your budget is, but if you’re looking for excellent and inexpensive, I’ll second Jesse’s recommendation for the Broadcast Warehouse (BW) PLL+ 1W is excellent! Rock-solid modulator, no overshoots, no tilt. It’s about $200, and it’s also available at http://www.progressive-concepts.com/info/item.php?id=45 . I’ve heard it’s difficult to buy it directly from BW in europe, but this american company ships anywhere.
It is, however, only 1 watt, so you will definitely need an amplifier to reach the power level you need.
I have an older version of this 100w amplifier from Aareff: http://www.aareff.com/en/100w-fm-transm … lifier.htm . I have successfully used it with my BW exciter.
Taking my limited RF experience into account, I can recommend this amplifier.
However, DON’T buy an exciter from aareff — it’s useless (overshoot city). Low frequency response is anything but flat.
Here’s some images to show you what I mean.
Test material is a falling slow sweep, 200hz to 2 (!!) hz, length 30 seconds.
Marian Trace Pro (completely DC straight output) -> Exciter -> Belar Wizard modulation monitor (essentially DC straight MPX output) -> Hardware Oscilloscope and Marian Trace Pro Input
The BW PLL+ was completely flat at 2 Hz (according to the hardware oscilloscope), however the trace pro input is not DC coupled, so there is a slight rolloff in the extreme infrasonic range of this image:
This is what a perfect exciter looks like, recorded through that sound card.
And for comparison, here’s what a bad exciter (i.e. Aareff’s) will do:
I contacted them, and they were not able to resolve the performance to acceptable levels.
Here’s another way of looking at it.
30 Hz squarewave through BW PLL+ 1W:
And here’s the aareff exciter:
Honestly I didn’t know whether to laugh or cry what I saw it.
Despite this, I can recommend Aareff’s amplifier, just not Aareff’s exciter. Take it with a grain of salt though — I have extensive MPX experience but limited RF experience, and I do not yet own a spectrum analyzer so I have not looked at the actual spectrum RF spectrum.
///Leif
Sigmacom, that looks amazing!
Hey, I’m in europe (Twente, Netherlands) all summer. Maybe I should come visit you, show you the Omnia.9 and talk more about future cooperation. 🙂
I’d also like to buy one of those exciters from you. Heck, it sounds like Greece’s economy can use it right now 😉.
///Leif
If you’re got the cash laying around, I can *warmly* recommend the MOTU 828 mk3.
It’s pricey, but it’s an incredible sound card. External (1U rackmount), supports both USB 2 and Firewire, it has a crapload of inputs and outputs, the outputs are DC coupled, and it of course supports 192 KHz!
There’s a smaller, slightly cheaper version too — MOTU UltraLite mk3, which has 8 inputs and 8 dc coupled outputs.
The 828 mk3 is $750, and the UltraLite mk3 is $550. In this case, you get what you pay for, though.
…unlike for example the RME HDSPe AIO, which is $800, PCI express, but NO dc coupled outputs, which means tilt correction is necessary, and tilt correction isn’t perfect. I bought one of these, and actually sent it back! I rarely send something back (usually not worth the time), but this time I did it on principle. But, you can have my MOTU when you pry it from my cold, dead hands 😉
///Leif
Jesse is 100% correct. Peak-control is a very finicky thing. It’s hard to achieve, and extremely easy to lose.
The high pass filter in breakaway live isn’t phase linear (as a phase linear high pass filter adds hundreds of milliseconds of delay, like the one in bbp in phase linear mode), and as such causes tilt, so if you start with a clipped waveform (squarewave), there will be overshoots.
If you look at the oscilloscope in breakaway live, you will see what’s going on. I didn’t include it just for shits and giggles, it’s actually an indispensable tool.
The sample rate converter can also reveal inter-sample peaks, but it will not introduce peaks on a TRUE peak and bandwidth-limited signal such as the output of BBP!
///Leif
Apart from the usual 4, 5, 6 and 7 band modes, the upcoming Breakaway versions will also have 3- and 2-band modes. 2-band seems to be excellent for voice processing, so maybe the next Breakaway Live version will have something that fits the bill 🙂.
///Leif
Syam, you asked me, and I wrote that pseudo code for you.
I thought you wanted to make a stereo coder. If I write the pseudo code, and someone else writes the code, then what part are you doing??
Honestly, if you can’t write the code yourself even with the pseudo code, then perhaps you should start with a simpler programming project before you attempt realtime audio processing. This is not an easy project, and neither is being a programmer! Being a programmer means being able to do it yourself even though it’s difficult. You mentioned this is a postgraduate project — didn’t they teach you how to write programs at school? If not, a programming project may not be the best choice for you.
///Leif
Jesse, I think he’s asking for Mini-ITX, not Micro-ATX.
Indeed, I’ve never seen a socket 1155 mini itx board with PCI either.
Marian has PCIe cards now, but they’re pricey. Seraph AD2 is €389.
Most on-board audio supports 192k nowadays, but even with tilt correction it’s not 100% accurate.
///Leif
I do.
Standard Breakaway Broadcast should run fine on an Atom D525 (and on an Atom N280 in low cpu mode), but ASIO could be a problem at the lowest latency (17ms).
The clipper internally works on blocks of 512 samples, but for the lowest latency we have to run the sound card at 256 samples.
256 samples means one block of audio every 1.33 milliseconds.
So, here’s how the work load looks for the CPU:
Block 0: No work (buffering)
Block 1: 2.66ms of audio, must be finished in 1.33ms
Block 2: No work (buffering)
Block 3: 2.66ms of audio, must be finished in 1.33ms
Block 4: No work (buffering)
…No block must ever be late, or there would be a glitch in the audio. So for low latency, we must have lots of CPU headroom.
Anyway, even if the current version runs OK on an Atom, it doesn’t leave any headroom for future improvements. With a faster CPU, we can do better processing. So, if you’re building a machine specifically for BBP, I definitely do not recommend going with an Atom!
If you have an Atom laying around, then there’s no need to guess — download the demo and try it.
If I was building a Mini-ITX system today, I would base it on the Pentium Dual Core G620T. It’s 35 watts (perfect for Mini-ITX) but it’s Core i3 class (Sandy Bridge) — big improvement from Core 2.
http://www.newegg.com/Product/Product.a … 6819116401In fact, I just bought a CPU like that for a new Mini-ITX system for my car audio PC 🙂.
Best regards,
///LeifMerry Xmas and happy new year to everyone, too!
Thanks, Jesse!
About that BBP bug, I spentmonths trying to reproduce it, but was never able to. It never did it for me on any of my 5+ test computers. That time, the customer reported it to be a dusty CPU fan / heatsink causing heat buildup and cpu protection throttling. When he cleaned it, the problem went away.
Anyway, if anything like this happens in the future (on the version built on the new framework), the framework will allow me to more easily add cpu benchmark figures for individual parts of the audio engine, which will help in narrowing down the problem!
Best,
///LeifCamclone, I’m sorry if I wasn’t clear before:
30hz HPF on peak-controlled audio causes overshoots.
Before:
After:
I’m not sure exactly what’s going on in your system, but this is what happens to the signal. The overshoots aren’t big, which means there wasn’t much infrasonic in the signal to begin with, but they’re there, clear as day.
I just did an experiment, high-pass filtering the signal at 10hz with a standard filter (this means TILT, which causes severe overshoots) and then applying a 30hz phase linear high pass filter to see if it would help at least a little bit.. It didn’t.
I know the problem is not with Breakaway’s peak limiting — it’s more accurate than the competition — so that leaves the sound card or the exciters. Since you say the other products do not cause the problem, that eliminates the exciters, leaving only the sound card!
You may have to pick up a Marian Trace Alpha. If this has really been causing problems, that would make them go away completely.
///Leif
[quote author=”camclone”]Jessie,
WHY Leif didn’t put high pass filter on the final stage of the sound ? ( after the left right clipper ? )
[/quote]
…because that causes overshoots. When clipping, both harmonics and subharmonics (through difference-frequency IM) are created. These harmonics are necessary to keep the peak level of the signal down (that’s the whole point of clipping), so they must not be filtered out. This is why Breakaway — and every other FM processor — has the high pass filter at the very beginning of the chain! Check the block diagrams.
Squarewaves (which I happen to know is your favourite waveform) contain infrasonic information, so without infrasonics, there can be no squarewave.
However, unlike the other processors, BBP (in phase linear mode) has an extremely sharp, phase linear filter. When you set the filter at 30hz, it’s almost dead silent at 25hz, whereas the other processors are only about 6dB down at that frequency. For those familiar with filter design, it’s an 8192-point FIR filter(!) In short, the filter in BBP is *more* effective, so I don’t believe it is the cause of the problems you’re seeing.
If you are seeing lots of infrasonic, you may just be clipping too much, but since you say you’re NOT seeing the same problems with other processors, I don’t believe the problem is related to bass.
I’ve seen receivers skip stations due to RDS settings. It could also be overmod, perhaps? Please check a hardware oscilloscope in parallel with the output to make sure there are no overshoots, whether low or high frequencies.
///Leif
VuvuX DSP plug-in by yours truly!
viewtopic.php?f=5&t=1105&start=0
By the way, Jesse, it almost completely filters out the bees in that video you posted 🙂.
///Leif
[quote author=”Martin S”]Can anyone recommend certain types of cable with low capacitance pr. length?[/quote]
Yes, Belden 1694A. I used 600ft of it for a 200ft component video run out to my outdoor theater projector, and got a crisp shadow-free 720p image. Best coax I’ve ever seen, I hear it was the de-facto standard for internal runs at TV stations. Of course, this is total overkill for MPX. Component video at 720p is 20 MHz (if i remember correctly), MPX is 60 kHz.
Any decent audio cable should be fine. Reasonable overkill would be a composite video cable.
///Leif
