Unanswered topics | Active topics


» »

All times are UTC+02:00



Post new topic  Reply to topic  Page 9 of 11
  [ 105 posts ]  Go to page Previous 1 … 7 8 9 10 11 Next
  Print view Previous topic | Next topic 
Author Message
hvz
PostPosted: Tue Jul 22, 2014 11:19 pm 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11241
BETA745-009 (RELEASE CANDIDATE 1):

Windows 32 bit:
Windows stand alone: http://www.stereotool.com/download/ster ... 45-009.exe
WINAMP DSP: http://www.stereotool.com/download/dsp_ ... 45-009.exe


CHANGES:
- Check CPU load issue since merging HQ mode code in. -> REMOVED EXTRA THREAD,
- Fix new method Attack (first kick is too soft; amount_of_attack / attenuation_level is too low because attenuation_level is too high at the start of compression).
- Clean up stuff from stereo widener that's no longer used (2 detection algorithms etc.).
- Update Ibiza preset
- Think of a better name (ACR Stereo: Attack Controlled Repeat stereo; E.D.L. Stereo - Eighties Delayed Limited; PEDL Stereo - Punchy Eighties Delayed Limited; DWEL - Delayed Wide Eighties Limited; PEDRL - Punchy Eighties Delayed Repeat Limited stereo; DREL; DRL; LDR; ...). Keep ACR?
- High input level appears to activate limiter -> NO.


OLDER:
Most important change: New attack detection mechanism in stereo widener that - I think - more accurately controls L-R volume boosts.
Before, I used 2 compressors with different release speeds and then calculated the difference in level between the two; when both hit the attack they would go the the same level (maximum stereo increment), then it would drop again. But, this tended to cause 'jumps' in level, and in audio without big volume differences the amount of stereo increment would be high because both would stabilize at approximately the same level. Which is bad: The idea is to boost transients more than constant tones.

The new version detects when attack happens and normalizes that to match the amount of attack with the amount of compression. A completely separate release is used to smoothe the effect. The L-R part of constant tones is hardly boosted at all anymore. Also there are far less jumps in the volume, overall it sounds far more stable.

Note: Some default values are changed for a better sound with this new detection mechanism!

- Implemented new transient detection method. (Move the 'Method' slider from 0 to 1; 1 gives the new behavior, 0 the old). To test, slide between method 0 (old) and 1 (new) with the new Method slider. I'm planning to remove method 0.
- Add all stereo types on/off switches in main Stereo panel.
- Add ACR Stereo indicator in main panel
- "Hide" Extreme Tweaker mode
- Make Punch behavior less complex!! Remove settings that aren't needed.
- Optimize stereo widener code (remove pow's)
- Add Bojcha's Ibiza preset
- Change version number to 7.50
- Look into amateur/HAM VST plugin settings (email Dave) WAITING FOR FEEDBACK

- Fixed 2 bugs in jump detection. Until now some spikes could be 'missed', and the same audio could come out differently each time you played it, that is no longer the case. Also adjusted some default settings and ranges.

This is how it works:
- We have an attack and release, which are used to follow the peak levels.
- Volume: There's a slower release ("relative volume release"). The difference in level (the version with slower release gets equal during attack, but drops down more slowly) is used to adjust the volume. If you set Volume effect higher this has more effect and you can clearly hear what it does. 100% is roughly equal to the Spartacus plugin.
- Adjustment speed: Speedup drop speed controls how quickly after an attack has occurred (according to main attack/release) the speedup stops again. Relative speedup controls how much effect this has. Set to 0 to turn off.

Both can be best adjusted when using Difference mode.

The Volume effect can be heard best when you set Volume effect to 100%. If needed set Limit below and Stop above to their maximum values.
The adjustment speed can best be heard when you set Volume effect to 0%, Limit below and Stop above at their default values (or lower), and L-R level boost to 300%. Then find tracks with punchy sounds and you can hear what happens when you change the Speedup drop and Relative speedup sliders.
- New jump detection. Now working in Spartacus-style, but without certain bad things in Spartacus (if you send in audio at a diffe
rent level the amount of stereo widening in Spartacus varies greatly, which is bad).
- Added L+R loudness jump detection - on L+R loudness jumps we can adjust the L-R level at a MUCH higher speed without causing noticeable effects. Which helps to create a much more constant stereo image.

This is a bit difficult to set up (default settings MIGHT be ok, but better settings might be possible); I might try to simplify it later or even remove parts of it completely.

Here's what I did:
- There's a new panel for Punch Detection
- It contains an attack and release setting which behave as a normal (extremely simple) attack and release on the signal.
- The next thing that happens is that the amount of attack that occurs is added up, and reduced again. The reduction speed is set by Peak Detection Drop.
- Relative Speedup determines how much faster all the settings under Response Speeds (except Peak Follow) are working relative to the value found in the previous step (Peak Detection with Drop.).
- This would be completely impossible to configure. But the next two sliders will help here. If you set Volume effect strength to 100%, stereo enhancement will only occur at a Punch peak. Volume effect can be used to adjust the relative level (compared to the Peak Detection level). Note: With a slow Peak Detection Drop speed this can be used to replicate the effect of the Spartacus VST plugin, which in turn was made to replicate a StereoMaxx. I would *not* advice very high levels here (values around 30-50% work pretty well though).

I think this description is probably unclear, but just take this from it:
- If you set Volume Effect to 100% and enable Difference mode, you'll hear that L-R is only widened after an L+R kick. What you need to know is that the speed at which settings are adjusted is relative to the output level with these settings. When there's silence you get the behavior of the previous releases (you also get that when you set Relative Speedup to 0).


- Added Balance (for input sound card)
- Added gate to stereo delay
- Added extra L-R vs L+R detection with different attack/release timings, actions are now taken based on the minimum of these two.
- Fixed lowpass filter slope and added option to set the level above the maximum lowpass frequency.
- Highpass and lowpass freqs are now locked (you cannot set the 2nd below the 1st.
- Changed ranges and default settings.
- Added settings to control limiter response speed.
- Changed default speed settings to get a much more constant sound level without jumping up and down
- Highpass and lowpass (!) filter added
- Added option to also limit the input stereo level.
- Stereo Delay now also visible in other than 'Extreme Tweaker' mode
- Stereo Delay limiter added. Also in case of existing extreme stereo it is switched off completely.
- Stereo Delay now also works with 0 delay (behaves more like Stereo Boost).
- Stereo Delay does not boost bass below 150 Hz, with a slope going up to 300 Hz.
- Natural Dynamics worked slightly different in version 7.43 and 7.44, and it turns out that it sounded worse (the bass is boosted much longer, causing excessive bass). Sound of version 7.42 is restored.
- New stereo widener added, based on techniques used in some famous stereo wideners since the 1980's. For now (I will change this) only available in Extreme Tweakers mode. See below.

TODO:
- Move Volune effect slider to Advanced settings.
- Check nelson c's 7.44 problem (Juli@ driver issue?) (TEST IS RUNNING IN MY OFFICE - unable to reproduce so far)
Top
   
Reply with quote  
mpex2006km
PostPosted: Wed Jul 23, 2014 12:26 am 
User avatar

Joined: Fri Jun 27, 2014 4:20 pm
Posts: 211
Quote:
Take some double core K10 Amd Athlon II X2 or Phenom II X2 and you can unlock the Phenom to 4 cores if you have some compatible model of Mainboard. But remember , only K10 family or later .
I found an AMD Phenom II X2 550 (HDZ550WFK2DGI). I have a gigabyte AM3 motherboard wich supports acc-ec and i'll go the way you sugested. Thank you very much ...
Quote:
Quote:
The bad thing is that my cpu's heatsink base broke last night and for now no more single core cpu to test newer versions.

One thing that i would like to tell me what am2 cpu's to look for stereo tool.
http://www.cpubenchmark.net/cpu_list.php
Look them up in this list.

Then calculate the speed per core as follows:

1. Divide total speed by #cores
2. If the CPU supports HyperThreading (probably not, I think only Intel has that), divide the result by 1.3.

The higher this number is, the more of ST you can run. Hint: Quality from 0% to 100% makes a 50% difference in CPU load. So, if you now run at Quality 20%, and you want to run at 100%, the new CPU must be at least (200 / 120) = 1.67 times as fast as your current (if you have multiply cores that helps as well of course, but not everything in ST runs on multiple cores. Current ST uses 2 cores at most, plus one for the GUI).
The benchmark gives 2040. Now if divide 2040/2=1020.My old AMD Athlon 64 3500 had 531 so i think a big improvement...


Thanks to Both of you.

_________________
Very proud user of Stereo Tool since 2010. Thanks Hans.

My radio station
http://amradio.ddns.net
Top
   
Reply with quote  
Brian
PostPosted: Wed Jul 23, 2014 12:33 am 

Joined: Sun Dec 12, 2010 2:26 pm
Posts: 885
Quote:
BETA745-009

CHANGES:
- Check CPU load issue since merging HQ mode code in. -> REMOVED EXTRA THREAD, waiting for feedback...
That seems to have taken care of it. However, mpex has a multi-core system (Phenom 9550).
Quote:
Quote:
Quote:
@Bojcha: Are the EQ graphs still heavy? I thought I had fixed (or at least greatly improved) that a while ago...
I remember when you improve it. But it still makes problems. As i said, when CPU is on edge even hovering over any place in gui makes stuttering. I asked for option to turn off scope and i am now asking same for FM graph.

There must be way to fully avoid that.
+1. I'd like that as well.

EDIT: I can't run composite cliper (allone not any other feature) in full strickness with any mouse mouse movement .Even with quality set to 20% and gui display set to 0%. My pc uses an AMD Phenom 9550 Quad-Core 2.2Ghz and for ram i have 2.5GB OS: Win7

The commonality actually seems to be AMD, if Bojcha can confirm that his issues are on the AMD and not the Core2 systems.

Edit:

Very confusing here... :? Mpex now references an Athlon64 3500+, which is indeed a single-core, but also mentioned the Phenom 9550... :?:
Top
   
Reply with quote  
mpex2006km
PostPosted: Wed Jul 23, 2014 1:13 am 
User avatar

Joined: Fri Jun 27, 2014 4:20 pm
Posts: 211
Quote:
Quote:
BETA745-009

CHANGES:
- Check CPU load issue since merging HQ mode code in. -> REMOVED EXTRA THREAD, waiting for feedback...
That seems to have taken care of it. However, mpex has a multi-core system (Phenom 9550).


The commonality actually seems to be AMD, if Bojcha can confirm that his issues are on the AMD and not the Core2 systems.

Edit:

Very confusing here... :? Mpex now references an Athlon64 3500+, which is indeed a single-core, but also mentioned the Phenom 9550... :?:
Ok let me explain. My personal pc has an Phenom 9550. My http - mysql - icecast - radiodj + stereo tool (winamp dsp as a pluging on the radiodj) pc was running on Athlon64 3500 (single-core).

The bad thing is that my single-core cpu is not usable anymore because its heatsink's base broke.

Sorry for my bad English. THANKS

_________________
Very proud user of Stereo Tool since 2010. Thanks Hans.

My radio station
http://amradio.ddns.net
Top
   
Reply with quote  
hvz
PostPosted: Wed Jul 23, 2014 1:19 am 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11241
Quote:
Quote:
BETA745-009

CHANGES:
- Check CPU load issue since merging HQ mode code in. -> REMOVED EXTRA THREAD, waiting for feedback...
That seems to have taken care of it.
Great! You might wonder why I made that extra thread in the first place - it was to use the same code for Stereo Tool and for another product; I hadn't expected any performance impact from it. Anyway, now using separate code again (only a few lines so it's not really a problem).
Top
   
Reply with quote  
Brian
PostPosted: Wed Jul 23, 2014 1:20 am 

Joined: Sun Dec 12, 2010 2:26 pm
Posts: 885
Quote:
Ok let me explain. My personal pc has an Phenom 9550. My http - mysql - icecast - radiodj + stereo tool (winamp dsp as a pluging on the radiodj) pc was running on Athlon64 3500 (single-core).
OK. Got it.

This version took care of the issue I had with the preset list causing stuttering. Has it taken care of any issues for you on either computer?

@Hans: Instead of removing the code, perhaps just put a check for a multi-core system around that code? Not sure what the purpose of doing what you did was though. I guess the question would be was there any ultimate benefit in audio quality to be obtained, or was it just for separating processing of GUI and the host program?
Top
   
Reply with quote  
hvz
PostPosted: Wed Jul 23, 2014 1:56 am 
Site Admin
User avatar

Joined: Mon Mar 17, 2008 1:40 am
Posts: 11241
Stand alone version 009 (Win32) is now also available. See a few posts back.
Top
   
Reply with quote  
Brian
PostPosted: Wed Jul 23, 2014 5:28 am 

Joined: Sun Dec 12, 2010 2:26 pm
Posts: 885
Not so sure about this ACR Stereo...

Check "Time" (Pink Floyd) and see if that is the desired effect at the end of the clocks...
Top
   
Reply with quote  
mpex2006km
PostPosted: Wed Jul 23, 2014 9:39 am 
User avatar

Joined: Fri Jun 27, 2014 4:20 pm
Posts: 211
Quote:
Quote:
Ok let me explain. My personal pc has an Phenom 9550. My http - mysql - icecast - radiodj + stereo tool (winamp dsp as a pluging on the radiodj) pc was running on Athlon64 3500 (single-core).
OK. Got it.

This version took care of the issue I had with the preset list causing stuttering. Has it taken care of any issues for you on either computer?

@Hans: Instead of removing the code, perhaps just put a check for a multi-core system around that code? Not sure what the purpose of doing what you did was though. I guess the question would be was there any ultimate benefit in audio quality to be obtained, or was it just for separating processing of GUI and the host program?

The good news are that This version seems more stable when i run it on a single-core/multi-core. and the problem with the presets is gone + i can see less cpu load when i move the mouse over the frontpanel.

_________________
Very proud user of Stereo Tool since 2010. Thanks Hans.

My radio station
http://amradio.ddns.net
Top
   
Reply with quote  
mpex2006km
PostPosted: Wed Jul 23, 2014 12:50 pm 
User avatar

Joined: Fri Jun 27, 2014 4:20 pm
Posts: 211
Hans can you please explain to me how ST uses a multicore cpu. I think it uses 2 cores for audio amd one for GUI. And my second problem is why I have shuttering even with 60% of core usage ( this is the most used core)???

Thanks

_________________
Very proud user of Stereo Tool since 2010. Thanks Hans.

My radio station
http://amradio.ddns.net
Top
   
Reply with quote  
Display posts from previous:  Sort by  
Post new topic  Reply to topic  Page 9 of 11
  [ 105 posts ]  Go to page Previous 1 … 7 8 9 10 11 Next

» »

All times are UTC+02:00

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Limited