Forum Replies Created
-
Posts
-
Code and gui complexity, Stuart. The key to a user friendly product (and maintainable code) is to keep it simple.
It’s simply not useful enough at the output to warrant the complication.
I am however adding an ITU loudness meter to MpxTool! There, it makes sense, because you can then numerically compare your loudness to other stations.
///Leif
Interesting!
I wouldn’t have thought it’d matter. I’ll be sure to test it on a fresh install without video drivers, and see if I can reproduce it.
Best,
///LeifPhase linear high pass filters cause LOTS of delay. A 30hz brick-wall high pass filter such as in BBP uses 16384 taps (!!). This means a delay of 16384 samples. (Yes, theoretically only half, but you have to buffer at both the input and output.)
A 20hz brick-wall high pass would probably need to step up the number of taps.. That is, 32768 samples, or a full 743 milliseconds (yes that’s seven hundred and forty-three), JUST for the filter. It doesn’t seem worth it.
If I made the filter non-phase-linear, it wouldn’t add any delay at all (at midrange frequencies), but then we’d lose phase linearity — not an option for me. However, you can always add your own filter as an effect plug-in!
Best,
///LeifHi Ken!
It’s not out yet, but it will be shortly. It will be $299, and it will in fact replace a Live + BBP combo!
That is, you’ll be able to run just BBP, and get both FM output and low latency output.
It will be a FREE upgrade for people who have previously bought both BBP and Live, or simply a $100 upgrade from standard BBP.
Best,
///LeifHowdy!
The one reason for keeping it at the input instead of the output of a radio processor, is that the input loudness meter will greatly help set the input level, whereas for the output, there’s not much to show. Levels are very uniform after Breakaway (especially with radio-level processing), so there’s never a reason to use a loudness meter to set the output level.
For a TV processor, on the other hand, you’d need loudness meters on both input and output, since you’re not processing nearly as hard, and there are actual loudness standards to follow.
Best,
///LeifTried it. Even compared to the 8500 stream, the new one is wimpy and weak sounding, and also neither loud nor clean.
I can’t find anything worth emulating on that stream.
LuK, you never actually responded, what did you think about the clips I posted? What are your comments?
///Leif
Hi Scotty!
Indeed, Breakaway is doing the conversion from 44.1 to 48.0. You can test the performance of the SRC by hitting Bypass and then running a sweep through. Should look very nice and clean.
Coming out of Beta will be a while! The reason is that I’m still adding features. I thought they were stable, and then I realized all these other things I had to add.
For example, ITU BS.1770 loudness meters on the input:
Also, while still in Beta, I can still replace presets without people complaining too much. After official release, that gets much trickier, and rightfully so.
Best,
///LeifHi Scotty!
Nice instrument!
YES – this absolutely means the laptop’s onboard audio is really running at 48,000. Most consumer-grade audio chipsets do.
This is why the default Input Rate in Breakaway products is 44100, and the default Output Rate in Breakaway products is 48000. My SRC does NOT do what you see in that analysis screenshot 🙂.
Best,
///LeifHey guys, I’m loving this thread! I no longer have to look very far to find excellent sounding stations to listen to 😉.
Keep up the good work.
Best,
///LeifHi Scotty!
Set up that way, Windows will be doing the sample rate conversion from 44.1 to 48 kHz. It will *not* do a good job. Create a 44100hz file with a linear sweep from 20 to 20000hz (5 seconds long is enough), play it through your player and listen. You should hear all sorts of "birdies" as the sweep goes up in frequency. Then, create the same sweep in a 48000hz file and play that — it should sound completely clean.
(BBP does not add audible aliasing while clipping.)
A pipeline is a virtual audio device with two ends. It looks like a sound card in both ends, but they’re connected, so as you play into "Pipeline 1 Out", it will appear on "Pipeline 1 In".
DS = DirectSound, Wave = MME. They’re both older, standardized ways to access audio cards, and they are in fact translated to KS (Kernel Streaming) by Windows. KS is the native language of almost all sound card drivers (including VAC / Pipeline), thus applications that support KS correctly can bypass several layers of windows.
Running the pipeline at 48000hz should not yield fewer glitches than running it at 44100, but who knows? Interesting find!
Best,
///LeifWait a second, what happened to the old "perfect sound" I spent a significant amount of time to emulate? How much could your ears have changed in a month?
///Leif
Hi Thagurt!
Positive Tilt (outwards tilt) basically never occurs, but if it does, you would correct it with a capacitor *in series*.
Negative Tilt can be corrected by adding large capacitors in parallel with the existing ones. To find which capacitor is causing the tilt, you will need to probe the circuit with an oscilloscope while running, to see where the tilt happens.This is tricky when using on-board sound (since you must then probe a powered-on motherboard without crashing the computer) but it’s possible — I’ve done it.
Disclaimer: PROCEED AT YOUR OWN RISK. It is absolutely possible to fry your motherboard if you make a mistake! You have been warned.
You’ll need to take the computer completely apart and assemble it on the bench, using the minimum parts possible. You’ll need to be able to access the underside of the board (or sound card) while probing with an oscilloscope, and if using any pci cards they must be kept steady throughout the whole ordeal or the computer will just freeze. This can be accomplished with a rubber band, to lock the card leaning to one side.
Then:
Launch BBP and set it to output squarewave, and start by probing the output, making sure you see the tilt.
Look for electrolytic capacitors near the output. Probe both legs (while keeping the oscilloscope probe ground connected to the computer ground). If you find a capacitor where one pin shows more tilt than the other, then you have found a culprit. Add a much larger value capacitor in parallel with it, taking care to match the polarity. DO NOT remove the old one! It works even better with two in parallel, for noise filtering purposes.
After adding a capacitor, check the output again. If it’s flat enough, you’re done. Otherwise, look for more capacitors.This is how I modified by M Audio Delta AP192 for near perfect squarewave performance.
Good luck!
Best regards,
///LeifIt depends on the motherboard.
On the very cheapest motherboards, the manufacturer may neglect to even put a coupling capacitor in. Then, you’ll see perfect DC straight performance (although with a significant DC offset) and no roll-off whatsoever.
On more expensive motherboards though, the manufacturer helpfully puts a couple of capacitors in, but perhaps cheaps out and puts a small one in. Hello tilt, hello premature roll-off!
Funny how that works out, isn’t it? You just never know until you measure it for yourself. This is why I recommend to everyone: Buy a freaking oscilloscope!
A used oscilloscope is NOT expensive — you can find one on eBay for half the price of BBP. Compared to even the cheapest hardware processor, even counting a dedicated computer, a used oscilloscope, etc — you’re still saving a LOT of money, and getting the best possible performance.
Best,
///Leif
