Stereo Tool

Do you have a C:\temp directory? If so, can you check if there's an exception file in it and send it to me or post it? Thanks...

BETA751-022

It took some time to find out why the sound was disappearing in some combinations of settings after running fine for upto multiple hours - but I have found it now and fixed it. Also, I have added a lot of extra logging that's disabled but I can turn it on again if this problem ever resurfaces.

Windows 32 bit:
Windows stand alone: http://www.stereotool.com/download/ster ... 51-022.exe
Winamp/DSP plugin: http://www.stereotool.com/download/dsp_ ... 51-022.exe

Windows 64 bit:
Windows stand alone: http://www.stereotool.com/download/ster ... 51-023.exe

CHANGES:
Audio drop in FM output with Tilt correction enabled (which was caused by the clipper!) solved.
Latency 4096 + 1024 bands in pre-limiter sounds good again.


TO DO:
- Fix RDS clock on 64 bit
- Add visual to show Pre-limiter effect.
- Check Pre-limiter to see if I can improve how it works, or make it more logical / easier to setup.
- Make DIFF work for all pre-limiter settings.
- Exception in: GUI Event Handler (Exception 0xc0000005 at address 0xf5f5b8, read/write=1, r/w location=0x0). Offset-info: 0xeb0c84. Reported by Fario and 2Sense. Unfortunately the crash occurs in a function called __intel_memset and I cannot see where it's called from. (Well, the GUI obviously...)
- Add presets, and remove a lot of older ones
- Improve web interface further. -> Postpone, can be done together with hardware GUI development. (very soon, among others preset selection)


OLDER:
- Stokkemask in strict mode (checkboxes for 'skip smoothing' and 'sloppy smoothing' disabled) is *slightly* less strict than before, but the sound quality is improved a lot (effect on the audio of the Stokkemask filter is reduced by more than half).
- You can set Strictness (CPU) to 3 instead of 2 as maximum. 2 should normally be sufficient though, and it's the default setting.
Stokkemask CPU load optimized for all settings.
Bug in Stokkemask and Multipath Clipper fixed that makes it work better.
- Added faster processing paths for SSE3, SSE4.1/2, AXV and AVX2.
- Optimized Stokkemask filter (so far only in Faster mode, will optimize the other modes later). Filtering is slightly less good, although I don't see it in the measurement results.
- Changed band steepness minimum values for multiband from 6 to 3 dB/octave. (Ooh. I hope i changed MB and not ND...).
- Added a bunch of new presets. Some need to be updated again. And the Jeroen Platenkamp HvZ mod version is currently identical to the original.
- Web interface drawing of equalizer fixed
- Crash when loading some presets fixed
- Bug fix: Prelimiter acted up at certain latency settings.
- Web interface dragging fixed.
- Bug that caused sliders and top menu to stop working fixed.
- Calculation of FM MPX left/right peaks was wrong when using Composite LIMITER, fixed.
- Made left/right display green (with above 100% yellow) as suggested by mpex2006km.
- Improved web interface a lot more ("1000" is gone, more displays are working, right part of display auto-scales to device width)
- Fix MPX Power calculator
- Fix resetting of tilt calibration when loading a preset
- Add volume meters which show > 100% modulation for IBC demo
- Bug fix: Single-channel file in plugin version caused crashes when using the composite clipper.
- Compressor/limiters: Moved limiter before compressor. This slows down the attack speed when limiting occurs, which gives a more stable sound.
- Improved web interface. (Not yet finished).
- There's some weird sound audible when you enable 'DIFFERENCE' mode. Even if nothing happens. This means that the audio is somehow damaged. It's worse at lower latency settings. FIXED
- Make highs clipping level configurable. DONE
- Fix composite clipper mode. DONE
- See a few posts up. I've now added 2 sliders that can be used to reduce the effect when there's not clearly a tone present. Tones only reduces the effect when there isn't a single tone that clearly sticks out - basically it checks how loud the strongest band is vs. the total of all bands, and adjusts the effect based on the result of that calculation. Since there are almost always also other tones present, setting this too high will even reduce tones that do clearly stick out. Tones threshold can be used to turn the filter off completely if there isn't a single tone that clearly sticks out (it makes the slope of the Tones only effect steeper).
- Changed some things back to how they were in version 005 - Bojcha reported (and I confirmed) some weird artifacts in version 006. If you enable METHOD2 you get the 006 behavior, that's for my own tests and I will remove METHOD2 later.
- Added Smother filter. The higher you set it, the more it will ONLY affect the loudest sounds. If you have multiple loud tones (for example a bass and a loud voice), setting this very high might cause only one of the two to be effected - setting it very low will affect everything though.
- Added multiband pre-limiter/phase optimizer to the Advanced Clipper.
- Trying a different multiband limiter in the Advanced Clipper. This time without phase optimizer. I'm very curious which version you guys prefer!

Correct me if i`m wrong but is the prelimiter "limits" all the frequency band of the songs?
I mean when give it to the lowest bands and press "difference" , i see limited signals even above pilot tone . Is it a displaying bug in the difference option only or it is real result. Must be checked also.

As i know prelimiter is in the clipper structure , so it must limit everything in the bandpass sound only - from 30-15000 ? am i right ?
Or may be the limiter is before the bandpass filter ?

Just added the 64 bit version to the post above!

@DJ-DOGGY: I will add some visuals to show what the pre-limiter is doing. It could impact the L-R because it works on the two channels separately.

Ok, will see , but it was strange . I`m not talking about the L-R . It was clear regognizable that it is on the L+R signal and continue to the pilot or above it. Which causes extend of the L-R too , yes.

Why would anyone use that prelimiter anyway ? Its completely useless imo, its not a clipper but its not a proper look ahead peak limiter either, so whatever it is its not really doing a great job.

It doesn't do what you think it does.

The intention is to remove very loud sounds before entering the clipper, because in the clipper those sounds would cause pumping - and they are too loud to get through anyway. This also protects (unfortunately not enough, not yet anyway) a bit against intermodulation distortion from loud bass tones.

It is however very difficult to configure it correctly, which is why I want to look into it again the next few days.

DIFF is unfiltered output, which may be weird. I'll see if I can improve on that.

Ok, i understand now. Will wait to test the fixes or improvements.

I found some "bugs" (more like design flaws actually) in the pre-limiter. However, after fixing them it doesn't work as it should anymore. Tomorrow I have no time to work on it so I'll probably have a new version next Friday.