Stereo Tool

-8 and disabling the oversampling for clipping and limiting now have me at about the same as beta 79. Not much of an improvement... I don't know if it is a cleaner signal either... vs. beta 79.

Correction: I'm now up to 0 with the oversampling turned off and have only 5-10% additional load vs. 79. Is the oversampling still needed?

Ok, I'm now making some further improvements. Basically, the filters have been optimized a lot and should run at at least twice the speed that ran at in BETA079. But I also changed the filtering to be as conservative as possible. That means that I need way more processing steps (making things slow), but also that the end result is louder and tighter clipped. Anyway, for lower CPU usage modes I'm now going back to the processing that I used earlier; it should be possible to get a reasonably tightly clipped signal using no more processing power than BETA079 used.

Oversampling: Don't know. Depends on the playback hardware etc. For example, if during playback Windows or the sound card oversamples the audio, it might clip - oversampling during clipping avoids that. On the other hand, if you're not pushing the volume to 100% (for example for web streaming), it might not matter much.

LOL... I just discovered that one change that I made after talking to Leif is actually making the sound worse! (cold and metalic) So I'm going back to my old design now. Which is also slightly more efficient. And it *may* allow for some extra loudness, but I'm not sure yet (first I need to fully restore my own clipping method).

Ok, I'm building a new version that:
1. Sounds cleaner and should perform (very) slightly better
2. Is clipped stricter at low quality settings. And I added some more low quality performance optimizations.
At quality -6 the output is already reasonably flat.

I've also re-enabled the 'increase dynamics' slider. After careful testing I think you shouldn't set it higher than about 20% - unless you like distortion... At 20% the loudness is on average increased by about 0.15 dB (at 100% that's 0.3 dB), but that's on average - it does occasionally sound a bit more constant. But also more distorted.

BETA089:
- Improved clipper audio quality
- Slightly improved clipper performance
- Clipping is stricter at low clipper quality settings

Winamp DSP plugin: http://www.stereotool.com/download/dsp_ ... 20-089.exe
Stand alone version: http://www.stereotool.com/download/ster ... 20-089.exe
VST version: http://www.stereotool.com/download/vst_ ... 20-089.dll
VST version (No SSE2): http://www.stereotool.com/download/vst_ ... 20-089.dll
Command line version: http://www.stereotool.com/download/ster ... 20-089.exe
Linux command line version: http://www.stereotool.com/download/ster ... ETA620-089 NOT AVAILABLE
Linux GUI version: http://www.stereotool.com/download/ster ... ETA620-089 NOT AVAILABLE

TODO:
- Fix loading changed multiband frequencies
- Add buffer and filter for SCA output (SCA1 ok, SCA2 ok)
- Finish AGC improvement - make mono value configurable (replace checkbox by slider) 1 hour --> NO, not needed - anything else needed? -> NO
- Fix Punch
- Check what to do with new filters (such as bass AGC) - keep them, remove them, change them? --> KEEP
- Save new BASS_AGC setting in VST mode
- Loudness: Annoying cracking sound in bass. Slightly present in 5.00, worse in 6.00, maybe even worse in 6.10. Only when bass is too loud. Much worse than in Final Limiter (at same input level!) - so this clearly indicates a bug. Most likely cause: The filter that was added to remove bass artifacts....... - No, it's the louder bass. But it can be fixed by changing some settings. Default settings updated, and behavior for 'not Very strict' improved. Also Deep bass boost and Very deep bass protection are enabled for latency 512 now.
- Fix crash at program close
- Fix VST plugin version (does not run)
- Dynamically drop 'Allow louder highs, even if it causes vibrations' to 0 when bass filter suspects noticeable voice vibrations.
- Reduce Loudness CPU load
- Check and remove static variables
- Finish new de-essing filter (check what to do with the settings, remove at least some!)
- Convert Multiband input to MONO, then use arrays [2][4096] --> should give speedup. - FAILED
- Natural Dynamics: Fix or remove transient boost
- Natural Dynamics: Add expected + strength slider per band
- Finish declipping filter (clipping level detection + level reduction in dB). 1. Figure out why removal of unwanted frequencies causes flat lines at high quality setting with small overlap. This causes distortion, with this fixed repairs are MUCH better. 2. Fix MP3 correction, automatically scale down when this deteriorates the sound. --> TOP part fails!
- Declipping filter: Fix low latency behavior
- Always oversample clipping (configurable)
- Declipping filter: Change detection at lower input levels.
- "Test Right Channel" in "FM Transmitter Calibration" does not work since v6.10 (standalone) (bojcha)
- Declipper window close function - check! - Seems ok
- Scopes black background?
- Add AGC start level
- Declipping filter: Add comparting of sample history to make sure loud bursts are still detected properly (now, with 16 blocks of history, 65 samples are removed even when ignoring just 0.1%). I should also check the current block (probably with a margin of a factor 2).
- Declipping filter: Use reliability of tilt detection to determine movement speed; slowly move to default tilt (configurable!)
- Declipping filter: Far too many samples are marked as 'maybe/probably' clipped. Histogram not used or not cleared?
- Declipping filter: Optimize tilt detection for performance. SSE2 for maximum. And keep separate smaller histograms to determine the maximum (should perform much better).
- Declipping filter: Cleaned histogram must be mirrorred. After cleanup, mirror center is place where SIGMA above == SIGMA below! Only use on the lower end. Then run cleanup again! The result should contain approximately the same amount of data above and below. If not, that indicates incorrect detection.
- Declipping filter: Add a long term histogram, and use maximum of short term and long term start clipping level detection.
- Declipping filter: Check if 'No declipping when tilt cannot be determined' can safely be turned on.
- Declipping filter: Add setting and saving of new sliders (history size, percentage of highest samples to drop, tilt detection range start, end and precision)
- Declipping filter: Check why higher Quality settings causes cracking sounds. Answer: Rounding error, and too low filtering values to remove it.
- Move pre-emphasis to end of processing
- AGC: Separate 'Remove remaining peaks' settings for band 1 and 2
- Channels L/R swap in stand alone version when changing filtering/quality (eldoradofm)
- Declipping filter: Optimize peak matching for performance: Move determination to extra preprocessing step DONE; try to change if statements to min/max or something IMPOSSIBLE.
- Fix changing value of 'final limit' in Multiband when changing Loudness settings.
- Save bandpass filter order
- Save all new settings, also through VST interface
- Declipping filter: Change default values for Maximum distortion and Step size (see Bojcha's post)
- Try to improve reconstruction match (should remove artifacts for low bass + soft highs) ■ FAILED
- Declipper: Fix loud bass with soft higher frequencies issue. Do something with loudest frequency vs. peak level - if it cannot fit (Madonna, Metallica) there must be clipping, otherwise there might not be. On top of this I should somehow try to preserve deformed bass sounds (if they are not hard clipped). Maybe this will work: Find loudest frequency, if it's very low. Than assume that it's clipped at the maximum restoration level (?); generate a new tone (using lowpass filter) of the same frequency and clip that at some margin below the maximum peak level. Then filter the input signal based on the FFT of the clipped signal. Finally analyze this new signal; if it is indeed clipped the filtered output will resemble a clipped waveform (visible in sample histogram; no tilt issues!) instead of a sine wave. Use the output of this as input to determine how strict the clipping detection for the input waveform should be (eg. based on normal or sinusoid signals). This is all way too difficult and it just won't work. Here's the real solution: Detect if a loud bass sound is present, if so make sure that symmetrical clipping detection is used, possibly with automatic 'maybe clipped' level adjustment. This should only slightly affect Metallica (which still sounds very reasonable with symmetrical enabled), but has a big effect on DBN - Jack is Back (which is far less affected with symmetrical enabled; auto likely/maybe adjustment based on symmetrical should help even more). Using veil solution now
- Move veil processing to detilted signal
- Switch to symmetrical detection if very loud very low frequencies are present NO, not needed anymore because artifacts are fixed using another mechanism
- Fix veil and detilting display (in declipper window; separate windows for separate channels; no change in DrawWave code needed!)
- Fix declipper display cleanup
- If the reconstructed samples are close to the original samples, and there are a lot of 'clipped' samples OR there are very loud low frequencies present, don't declip. Also add texture based filtering, and use a slider to set the sensitivity. Also increase the effect of bass and reduce overall level.
- Save veil settings and 'not reconstruct if like original' (STS + VST)
- Fix clipper (idea: Clip, filter, filter diff against clip diff, apply).
- Can I combine 2 FFT windows into 1? That would give half of the improvement of the smooth window, without CPU load increase.
- Old clipper seems to sound better. (New with less steps also sounds good). Use new clipper AFTER old clipper, and use already clipped and cleaned signal as 'clean_input'. Or try a bunch of short clips, than a long one to clean up, then some short (but not very short) ones to clean up pre/post ringing.
- Declipper EXTREMELY slow when feeding DC input.
- Big window output is smaller than normal one - no it's not.
- Bass AGC RMS display does not work (always displays 0.00)
- Longer history for maximum level?
- In FM calibration 38 kHz is NOT working. EDIT: And more bugs: 23 kHz is not at 23 kHz 28 kHz is not at 28 kHz 33 kHz is not at 33 kHz 43 kHz is not at 43 kHz 48 kHz is not at 48 kHz 53 kHz is not at 53 kHz. - not true, tested and all seems to work fine. There is another issue though: This filter still uses the very old windowing method and sounds crappy (artifacts). To be looked into for the next Stereo Tool version.
- Remove Natural Dynamics - Partially done, does the crash come from here? No, for some reason ANY call on my system to InternetOpenUrl in a Winamp plugin causes crashes. But this happens even with version 3.03, so it's not something new - it's only something new on my system...
- When "Enable FM transmitter calibration" is OFF and you click "Calibrate FM transmitter ..." and then close calibration screen again, then "Enable FM transmitter calibration" is turned ON automatically.
- Use DLL name in INI file name (requested by Jazler)
- Add registration to command line versions
- AGC: Added 'Force level below' slider for improved peak control. If the 'down speed' causes the volume to be much higher than the calculated 'target' volume, with this slider you can force the volume to go down anyway if the difference gets too big.
- Registration key must be written to DEFAULT ini file. And for stand alone version, ini file StereoTool.ini should automatically be converted to stereo_tool.ini to match existing ini files.
- FM Calibration beep tones are still not corrected.
- Stand alone: Check box to FM Calibration on/off missing in the FM transmitter calibration window.
- Declipper artifacts ( viewtopic.php?f=14&t=3065&p=12527#p12527 ): Adjust defaults or increase bass sensitivity. - No, replaced MORE switch by slider.
- Finish blind interface WAITING FOR FEEDBACK TODO: TEST: Only display this window when needed.
- RDS issue reported here: viewtopic.php?f=15&t=3703&p=11524#p11524 partially solved
- Optimize clipper code: Reflection, smaller overlap (more needed?)
- SSE version crashes (reported in email by Francis)
- Save new clipper options (STS+VST)
- Smarter lows volume drop against highs protection: First clip using reduced highs (highs protection filter), then do it again but do NOT lower the lows below the level found in the 1st step.
- Add composite clipper
- Highpass before Multiband etc?
- Update presets? (BASS_AGC etc.) 1 day
- Change version number 1 hour
- Release 1 hour
- Add lowpass filter for stereo signal (will cause a lot of extra latency!) - it might be possible to avoid this latency using a Hilbert transform
- Add smarter clipping detection. Maybe something much simpler suffices: Current clipping detection with threshold + flat line detection
- Declipping filter: Automatically override the 3 clipping level sliders if the clipping level is detected very clearly (clear thin spike in sample value histogram). - NONSENSE, this is already done by the histogram function. But it can be made a BIT better - I think - by automatically LOWERING the 'always clipped' slider if a lot of data is present at the highest few bins (but care is needed for DBN - Jack is Back like tracks)
- Declipping filter: Create a new histogram after declipping, clean up and subtract from original.
- Change stand alone version to 32 bits.
- Optimization: Reduce CPU usage when input is silence.
- Non phase linear highpass filter: Make steeper
- Natural Dynamics: Attempt to reduce number of bands by 2 (memory + cpu DIV 2), using bands with sloppy slopes (no interaction between bands -> memory usage DIV 2, out-of-cache reads gone, hopefull final filter gone too -> CPU DIV 2 again).
- Natural Dynamics: Voices, especially in chorus, still sound weird.
- Natural Dynamics: Smooth out different bands more to reduce low-bitrate-MP3-like sounds.
- Natural Dynamics: Optimize situation with no interaction between bands for performance.
- Natural Dynamics: Save settings in VST version
- AGC: Sudden loud kicks (32:40) - much faster drop! IDEA: IGNORE FAST 2ND STAGE DROP IF THE EFFECT IS LOW.
- AGC: Try frequency separator implementation (2 or 3 different volume measurements, take lowest). Should protect loud voices much better.
- FM Calibration causes artifacts and latency.

beta 89... Slider at 0. Same performance levels... Beta 79 level if I leave oversampling off and use shortcuts. Increase of 20-25% in load relative to the 79 baseline if I don't check the shortcuts checkbox. Increase of ~100% if I check oversampling and shortcuts. Increase of 150% relative to the 79 baseline if I check oversampling and don't check shortcuts.

@Hans
Beta89 uses too much CPU...I guess 79 is the limit for my PC (Athlon X2 4800+). I can't use 89 at full quality so forget it. I'm sticking with 79 for now. I know I need a CPU upgrade but I can't afford it right now.
I know there is only so much you can do to reduce CPU usage but try your hardest to keep it as low as possible. When you say you are working on the clipper, I assume you mean the one in the Multiband window above the Equalizer. I LOVE CLIPPING so make it as strict as possible and keep the CPU as low as you can for us slow CPU people.
I'll be upgrading to an i7 as soon as I can and then it won't matter anymore
Keep it up Hans!!!!!!

Hm, just to check:

Brian: So with the slider in the CENTER you're now getting similar performance values (at least without oversampling) as in BETA079, right? So basically this version is - at least without oversampling - no longer worse than BETA079? Do I recall correctly that you're using an Intel P4?

baires0314: Could you compare oversampled vs. non-oversampled performance? Please turn everything else off, especially FM processing (because that forces oversampling on). If my guess WHY there's much more difference on an AMD is correct, you should see a much bigger jump in performance than Brian when switching oversampling on.

Your current comparison was WITH oversampling I suppose? (It was if you didn't turn it off AND turned FM off).

Here are my measurements:
If I don't use oversampling (not: you have to turn FM off for that as well!), and only enable Loudness, and disable Advanced highs protection (that makes a big difference!), I'm getting IDENTICAL performance in version 6.10 and BETA089 with quality set to -3.

With oversampling, to get a similar behavior I need to go to between -8 and -9.


Edit: I found at least 1 more optimization that I can perform and - if I'm right - should reduce the CPU load especially on an AMD further. But I'm first waiting for baires response (there might be much better solutions available, depending on what he reports).

You do not recall correctly. My CPU is the Athlon 64 3700+ (San Diego core, 1MB cache).

Your other statement is right though. It's roughly equivalent, without oversampling AND enabling "Take some shortcuts".

Ok, I found a few other very minor optimizations. And one bigger one (but that needs more work - I could do part of the processing using 2x oversampling instead of 4x oversampling. That's a bit less precise, but the difference in the output should not be that big if I use 4x oversampling at least in the final step(s).

Now building the 'minor optimizations' version.