Forum Replies Created
-
Posts
-
ASIO4ALL did work for me when I tested (briefly), but there’s not much point in using in with Breakaway, since Breakaway already supports kernel streaming natively. ASIO4ALL is simply a KS -> Asio wrapper.
Alex, I’m not sure I understand the situation completely. Without using pure ASIO, latency will likely not be low enough to monitor in headphones. So, if the second sound card doesn’t support ASIO, how would you use it?
Also, even if all outputs are in use, why not use a splitter and headphone amplifier to connect the headphones on one of the channels already in use? 🙂
How about also putting a splitter on the input, so that you could feed the same signal to both the ASIO card and the 2nd sound card? That way the second sound card would get an unprocessed signal which your recording program could use.
It *does* get complicated when using ASIO. Normally, you can use Pipelines and virtual cables to route audio every which way, but ASIO is really only meant to turn a computer into a low latency "hardware" processor.. It’s much less flexible when doing things like this — ASIO was never designed to transfer audio between programs. It does one thing only, and it does it extremely well (and quickly) — get audio in and out of *one* program on a computer.
Ain’t the low latency mode nice though? There’s something special about hearing the processed audio immediately as it’s happening 😉.
///Leif
Hi Johnny!
They’re two completely different issues.
Cnbau said: "Many songs end with a fade-out", indicating that it’s the natural fade-out of songs he wants to keep. Breakaway does undo this — since it’s a volume-levelling dynamics processor, it’ll keep turning the volume up (to counteract the fadeout), by the set speed, until it runs out of range. There is also a control that sets how much power it has over the dynamics.
Demodave, on the other hand, noticed a problem with "artificial" fadeouts created by his Karaoke program adjusting the pipeline volume downwards. By default, the pipeline is set to ignore such volume commands and run full blast at all times, so this would completely remove the fadeout.
However, once Demodave has enabled the pipeline volume control, he may still run into the same problem that Cnbau did, and in that case he’d have to turn down the Range control a bit as well.
It’s a tradeoff — too little range means that it won’t be able to normalize songs that are too quiet. Too much range means it will normalize things that were supposed to be quiet. 😉
///Leif
Hi Alex, and welcome!
Algorithm wise, it would be an easy thing to do, especially with an ASIO sound card.
User interface wise, on the other hand, would be very complicated.. Supporting a function like this means it would also have to support all sorts of routing combinations. I’m afraid it’s not feasible, due to the support nightmare it would create.
What about a Y splitter on the instead? Seems like a waste to use a 10in/10out sound card for only stereo 😉.
Hey, if you explain the situation really well, and there really is no easier option, maybe I could add an ini-file option to do it. But, you’ll have to motivate very well 😉.
///Leif
Nice result, Sgeirk! 😀
Regarding Athlon, I don’t know. I don’t have one, but from experience I know that it performs a lot worse on AMDs due to my heavy use of Intel’s Performance Primitives library (for FFT algorithms for example), and they’re more optimized for Intel than for AMD chips, due to.. uh.. marketing reasons. 😉
But, if you happen to have an older 2.4 GHz Athlon with 1.5gb ram (i would guess that you do), please do try it!
If nothing else, I would think it’d run OK in cpu optimized mode!
///Leif
AdamH, x100 server full!
What’s with that? The sound must be too good or something. I guess you’d better upgrade your server 😉.
///Leif
Hey Erwin,
I apologize for being harsh before.
The thing is.. I never liked the sound of that particular processor — I don’t know what aspects of it I would actually emulate.
Here’s something you can try:
Go to http://mpxtool.com and download the torture test recordings. There are recordings from several popular processors — 8500, Omnia 6, 8200, DSP-Xtreme, DSP-X Mini etc. This way you can hear what Omnia 6 sounds like, and decide if this is really the sound you are looking for for your station 🙂.
///Leif
There is a way, Dave.
Download the attached Breakaway Pipeline control panel. Check the Volume control box for Cable 1, press Set. That should do it!
///Leif
Oops! My bad, nice catch, Stuart 🙂.
Then it’s definitely either the song or the headphones. Breakaway Personal does not add distortion, unless you turn the bass up way way too high.
///Leif
Are you looking at input or output?
If it’s the input, the distortion is in the song already.
If it’s output, turn Final Drive down 😉.
///Leif
Bass boost, when turned up, will:
Apply Parametric EQ before the multiband
AND raise the b1/b2 agc thresholds
AND raise the b1/b2 out mixBass boost when turned down will only:
Lower the b1/b2 agc thresholdsShape adjusts the shape of the PEQ as well as how much of b1 vs b2 to adjust.. shape minimum will adjust more of B1, shape maximum will adjust more of b2.
The crossover frequency (-3dB) point between B1 and B2 is 36hz, however the filters are nice and gentle, so in reality B1 is sub bass and B2 is mid bass.
🙂
///Leif
LPY, your stream sounds great! Processing suits the format, very clean.
It might be unnecessarily quiet though with the 2dB attenuation.
It’s probably not worth trying to out-think the users.. The overshoots you’d get from running 0dB into a 128kbps mp3 codec are no worse than what you get from the average CD encoded to MP3 (in fact probably less of an issue!), and the users who are worried about decoder overshoots are probably already using in_mad with clip protection, or [x] Fast layer 3 EQ, and have pre-amp at -2 in the Winamp Equalizer 🙂. Both methods avoid clipping, but when the stream is also attenuated, all of a sudden we’re attenuating 4dB instead of just 2.
///Leif
It’s indeed the sharc eval board 🙂. Video was shot over 2 years ago, the port was a project for Linear Acoustic. SHARC is a nice DSP indeed.. Way easy to develop for. Interestingly, my reference x86 code is all fixed point — I actually had to *port to floating point* to make it run on the sharc 😉.
///Leif
It stopped playing? That shouldn’t be — looks like you have found a bug for me, thank you 🙂.
It should just play the ad. I’ll look into it.
I haven’t had a chance to think about the multi instance licensing yet, but I will soon!
///Leif
MUCH better, Adam!
It’s still an absolutely crazy preset (NYC was made to be as loud as stations there, not to actually sound good), but X100’s sound is night and day compared to before. There are transients, no directly audible IM, and it’s listenable! I’ve actually listened for 2 whole songs now, with no urge to turn it off, whereas I barely lasted 30 seconds before.. and it’s still BLISTERINGLY loud, in fact even more so than before, since it’s more effortless sounding now!
It’s nearly impossible to be use the space in the waveform more efficiently than this 😉.
Nice one.
By the way, I wouldn’t worry about decoder overshoots for MP3. They’re miniscule (a sample or three) and basically inaudible. This is the whole purpose of the 15us pre + de-emphasis process — to make the overshoots more benign.
aacPlus, however, is a different story! Those overshoots are much bigger, and they get dealt with by a limiter inside the decoder, not a clipper.. thus, overshoots cause IM distortion. I would absolutely attenuate 2dB before the aacPlus encoder. I’ve attached a few attenuator DSP plug-ins i wrote recently — they may come in handy.
BTW, I love your faq! Especially the "must be in the form of a question" part 😉.
///Leif
