Stereo Tool

Nice progress!
I have to tweak some settings to get similar sound as at 4096 mode, but anyway much better.

I just realized one thing about ST DSP. This low latency modes are not so needed when FM is used.
Buffer can't be under 100ms anyway.

what a nice track xD

Why is that? It should be possible to set it much lower, especially when using ASIO... (Unless maybe when you're using the FM Transmitter Calibration settings)

Bojcha => i have listen your track but excuse me ,but this song is "beurk"

I prefer 1000x james last

I am not using ST DSP, but i want to know how to use asio with Winamp and ST DSP.
That's example for testing lowlatency modes

Bojcha => here a good asio plugin for winamp and with control panel access for asio configuration

http://www.megaupload.com/?d=1UFHQ0E1

I know for that Garfield...
But what is i already use other plugin like SQR Crossfade? I can't use both in same time.
Actually there is no need to use Asio in Winamp+ST DSP, because you actually have realtime "play/stop" in all modes.

I found a bug with BETA601-004. I am using the standalone version with custom MB center frequencies. When I use the higher latencies of 2048 and 4096 the upper 2 bands of the MB are completely cut off. These are the settings for MB center frequencies:
Center frequency - Band -1=45
Center frequency - Band 0=90
Center frequency - Band 1=140
Center frequency - Band 2=350
Center frequency - Band 3=1100
Center frequency - Band 4=4200
Center frequency - Band 5=6200
Center frequency - Band 6=8100
Center frequency - Band 7=10500
Center frequency - Band 8=12800

The lower latencies work fine. I have customized the upper bands because I am broadcasting 75uS stereo FM and the narrower higher bands allow better control of the compression and clipping (IMO) because of the pre-emphasis.

_________________
J0HN II

Band -1?!

Backwards compatibility. I started with 8 bands (1-8), then I added 2 at the bottom...

Ah, fortunately that one is easy to explain (and relatively easy to fix).

Currently, each band has maximum effect from 0.9 upto 1.1 times its center frequency.

In your case, that meas that band 9 (7) runs upto 10500 * 1.1 = 11550 Hz
Band 10 (8) starts at 12800 * .9 = 11520 Hz

So when I calculate how steep the transition is I get a negative value... Oops... This value (-30 Hz) is divided by the filter precision - 86 and 43 at latencies 512 and 1024 (which results in a value that's rounded up to 0), 21 and 10.something for latencies 2048 and 4096 (leading to -1 and -2).

But the real issue here is that the 0.9 and 1.1 calculation is not valid anymore if you change the center frequencies. I actually need to do something with the distance between the current and next peak level.


Edit: Solved, will be available in the next BETA.