Stereo Tool

Update is online. "Remove remaining peaks" and "Stereo Image" are working now.

LowLatency-beta2:
I forgot this, Hard limit is working now.. and i think it is good.
Loudness ... actually .. works very well (mostly i look bass) ..but with both ILDP and TWO STEPS enabled.

For normal mode .. i'll look again and better compare it with default v4.22

Stereo Tool (latest beta2) LowLatecy my test stream 64k AAC+ http://85.25.73.243:10100/listen.pls
sometimes will not be aveliable.
Preset is one i am working on for web.

For testing purposes I'm building a second version that's closer to the original 4.22 - but still has the same latency.

I had added a few extra steps in the BETA2 Loudness filter, in the version I'm building now I've turned them off again...


Version with processing closer to the original 4.22:
Winamp DSP plugin: http://www.stereotool.com/download/dsp_ ... _BETA3.exe
Stand alone version: http://www.stereotool.com/download/ster ... _BETA3.exe
VST version: http://www.stereotool.com/download/vst_ ... _BETA3.dll

The Loudness peak level control is not very good in this version - but if it sounds better than the BETA2 (and maybe than the original 4.22?) I'll use this version as a starting point for further developments.

Normal mode: HardLimit .. in beta3 .. i can see peeks in output meter (postAmp at -2dB)

beta3 with one preset without loudness one PC (amd dualcore) have 60-70% cpu usage. with default v4.22 cpu usage is 35-45%.

tomorow i'll look others.

Just checked: BETA2 and BETA3 with only Hard Limit on, Pre Amp at 20x - and I don't see any peaks in the output meter...

The CPU load is currently indeed higher (although just the Loudness CPU load should be roughly the same), I'll work on this later. The differences I see are a lot lower than what you're reporting though (20-22% for the original 4.22, 24-27% for the BETA2, 21-23% for BETA3 - that's all based on the FM Loud Bass Boost preset.

I checked the CPU load on another PC, and there I saw: BETA2 40% higher, BETA3 25% higher.

I still have to work on the CPU load, first I want to get the quality right. When that's done, I'll look into the CPU load - I expect to be able to lower it a lot, probably below 4.22 levels (for normal latency and high quality).

I've just done a quick test, and I managed to roughly divide the CPU load by 2 with little (but some) impact on the sound.

For now, I'll keep using the "full quality" version until everything is finished. And there are also a few other things that I can do to reduce the CPU load that won't impact the quality. Eventually I'll make it configurable.

In short,

Step #1: Get the sound in Normal Latency mode as close as possible to that of 4.22 (or better) - OR restore some of the latency in Normal Latency mode.

- Final Limiter "Reduced CPU usage" and "Avoid Distortion" switches
- Bass protection on very low latency settings. (although it's not really bad either)
- Check the bass problem without ILDP reported by Bojcha

Step #2: Reduce CPU load

Currently I'm processing some pieces of audio that are later thrown away, avoiding that might reduce the CPU load considerably.

Step #3: Make CPU load of the remaining processing configurable.

Yeah, good work Hans! If you have a version ready with reduced CPU usage, I will test it in a vmware environment.

To Bojcha: I've seen the problem that you mentioned (Hard Limit with peaks in the output) here too now. I will look into it.

Edit: I have now also found some tracks where the new version sounds clearly worse than the original one. I'll look into it and either fix it or restore the longer latency for the pieces of processing that are causing it.


I'll upload a new BETA3 where the Hard Limit problem is solved and the Loudness peak control is better in an hour.

Neither of the betas are performing well with SAM.

Beta 2 crackles incessantly, no matter how many times I reload it. Beta 3 forces me to close SAM altogether, as it completely takes my system with it!