Stereo Tool

Actually i was just trying multiband + hard limit (!?) .. nothing else
And in "low latency mode" there is strange "harmonics" (i think that is the word). i can hear more clear if i up first two sliders in EQ. With "post filter" there is more.

Ok... In that case, they should be gone in the new version (there isn't even a 'post filter' anymore).

However, if I implement low latency clipping, it will return... (only if you enable it, of course)

I've put something online for testing. Set it to Low Latency mode to test the new Low Latency processing.

Note #1: Pre-emphasis is currently NOT working, so this version is not really suitable for FM broadcasts. I'll try if I can create a 0-latency pre-emphasis tomorrow - if not I can definitely create one with a very small latency.

Note #2: There's a second input channel active now, please set it to an input that reads silence or you might get really strange effects. I first started with the low latency and silence detection stuff before I discovered that there may be better ways to reduce the latency...

Note #3: For some reason the multiband filter in low latency mode swaps the channels...

Note #4: If the difference with the default pre limiter/multiband settings gets bigger, the sound may differ more from that with low latency disabled... The multiband frequency response control is also a lot less precise because the separation between bands is very weak.

Note #5: Clipping and "limit output" currently don't work in low latency mode.

Winamp DSP plugin: http://www.stereotool.com/download/dsp_ ... r_BETA.exe
Stand alone version: http://www.stereotool.com/download/ster ... r_BETA.exe
Command line: http://www.stereotool.com/download/ster ... d_BETA.exe CRASHES
VST version: http://www.stereotool.com/download/vst_ ... l_BETA.dll



Note-to-self: Create very-small-latency pre/de emphasis + lowpass filter by bypassing any low frequency filtering!

Hi Hans,
This working very good.
I do quick test and make MIC preset with multiband/eq, noisegate, and simple compressor, sounds good and latency is still low.
one question, will meters work in multiband for lowlatency mode?

I found what causes strangde harmonics/chorus in lowlatency mode.. and it's not multiband .. it's bandpass filter.

Hi,

The meters will start working again. I'll try if I can fix the bandpass filter...

Here is my measurements:

multiband + noisegate + simplecompressor = 11ms , and that is fantastic.
But, That is without asio latency!

on 48kHz input, it's 38ms (asio only - ST bypass) + 11ms processing = 49ms

on 192kHz it's less but i can't measure because "monitor" on my e-mu 0202 not working 192kHz

Then i try breakaway live.
Interest is that only you can choose 44,1 or 48kHz for i/o. Processing + asio (overall) is 15-17ms.

so, what is it with ST asio ?

Hi Bojcha,

Previously, I had some filters that I had to process in blocks. Now, if a block is 15 ms, and the latency is less than 15 ms, I would run out of data.

That's why I filled in a minimum latency of 20 ms - and until now that wasn't too bad compared to the total latency. I'll lower it!

(By the way, 11 ms?! It should be less than 1... Will look into that)

I found a bug in the multiband processing --> will also fix that later tonight, fixing it should give better control over high frequencies.
And, Pre-emphasis is working now, also with a latency of 2 samples. The difference with the original pre-emphasis filter for frequencies below 14 kHz is less than 0.2 (50 us)/0.4 (75 us) dB! And de-emphasis should be possible with the same latency.

I measure simple...
reset .. i turned on multiband, noisegate, singleband compressor and hard limit... (hq on, lowlatency on and input 48kHz input)
I measured with bypass and without bypass... and one is 49ms and other 38ms.
So 11ms is for processing right?
Even better if this 11ms can be 1ms.

I am now more worry about why is more latency is with 48kHz and less with 192kHz.

I think I figured it out now.

In my stand-alone implementation I'm using the same "block size" (bunch of samples for which I call the processing) that Winamp uses: 1152 samples.

1152 / 44100 = 0.026 s = 26 ms
1152 / 192000 = 0.006 = 6 ms

I'll make a few quick changes to reduce the ASIO latency a lot more.

Since I installed Vista64 my Juli@ ASIO support is completely broken. So I cannot test anything

In about 1-2 hours I'll upload a version where I changed 2 things:
- Minimum ASIO latency setting can be set as low as 2 ms instead of 20.
- Processing blocks when using Low Latency mode are 64 samples instead of 1152.

But for now I really don't know if it will work or not.