Stereo Tool
https://forums.stereotool.com/

Stereo Tool 7.03 BETA
https://forums.stereotool.com/viewtopic.php?t=4448		 Page 86 of 102
Author:  DJ-DOGGY [ Fri Mar 15, 2013 5:41 pm ]
Post subject:  Re: Stereo Tool 7.03 BETA
I like Philips.
I saw modules by them for covering an area for defeat the radio stations in the time of communism here .I was a very stable RF module with parts from Philips. We talk about 30 years ago ( or so ).
Author:  Brian [ Fri Mar 15, 2013 5:57 pm ]
Post subject:  Re: Stereo Tool 7.03 BETA
Quote:
Quote:
Does that make more sense to you now, and do you see why it is of larger scope than just me and my system?
Yes, but this is not happening in my code. Simply because I don't use the dispatcher (except in the command line version). I used to a VERY long time ago (probably version 3.xx or so) but I then discovered that - even on the P4 I had at the time - the dispatching itself and the determination in the compiler whether it should produce multiple versions of a function doesn't work very well; by only allowing one code path (SSE2) I got a performance increase of about 10%.

So there is only one code path, and it either works or crashes if the CPU doesn't support SSE2.

There is also an SSE version (not for the beta's though); the installer checks which of the two you need.
I knew about the SSE version, and I wanted to try that on my system, but it's only for download for the VST version, not DSP.

At any rate, if you're not using the dispatcher, then probably the issue isn't happening. Probably, not definitely.

Found this on the Intel forum:
Quote:
Nick
Sat, 06/10/2006 - 01:05
From premier support: There is not a way to prevent the 32 bit Intel compiler from generating x87 floating point code. The /QxW switch enables generation of SSE2 instructions where that is beneficial, but it does not disable generation of x87.
Uh oh.Silly legacy 32-bit computers.
Are you using /QxW? If so, could you change to /arch:SSE2?

Also, looking at a guide for the Intel 10.1 compiler, /QxO should also work on AMD processors that support SSE3, and it also enables Core microarchitecture optimizations.
Author:  Brian [ Fri Mar 15, 2013 6:52 pm ]
Post subject:  Re: Stereo Tool 7.03 BETA
Quote:
Quote:
On the other hand, Framewave could be absolutely horrific. The thing is though, without trying, there's no way to know.
I know I looked at benchmarks before I went for IPP, unfortunately the site with those benchmarks appears to be down at the moment.

I did find this (I know it's old, but it doesn't look promising at all):
http://virtualdub.org/blog/pivot/entry.php?id=186
5-year old stuff there, on an open source project. Odds are, things have changed.
Author:  Brian [ Fri Mar 15, 2013 6:58 pm ]
Post subject:  Re: Stereo Tool 7.03 BETA
Quote:
Looks like no one is seeing this in beta54 .. Bojcha , Brian, ..others..
I haven't had much time to listen to 54. Also, I'm pretty much not using the new multiband, and I never use anything related to FM, as I'm not a broadcaster.

As for the new multiband stuff, to be totally frank, with the presets I've tried, I truly don't hear anything amazingly different from the classic multiband. This is similar to my opinion about the declipper.

This is why I'm frustrated about the inefficiencies being left in the classic code, because the new code is more demanding of my processor, but doesn't give me any additional audio fidelity that I can HEAR. Maybe it provides something better on a scope, but not to my ears. Maybe that will change with time, as people become more skilled with using the new features and/or any bugs are found and eliminated, but for now, to my ears, there isn't any sort of "WOW" moment, so I don't see the advantage.
Quote:
Strange ....
If i low the latency about 1024 and 512... the issue is even rising . Both bands are affected (first and second) clipper is goin crazy ? May be at 8192 samples this will gone :P
As for what you're seeing with 1024 and 512, that's what I've already reported, except from a performance perspective.
Author:  DJ-DOGGY [ Fri Mar 15, 2013 8:21 pm ]
Post subject:  Re: Stereo Tool 7.03 BETA
I`m not broadcaster too..
But the issue i talk about is not broadcast related , but only from general test on ST, and see some weird spikes when enable the synchro to output... that`s it. only beta54
Author:  hvz [ Sat Mar 16, 2013 2:13 am ]
Post subject:  Re: Stereo Tool 7.03 BETA
Quote:
As for the new multiband stuff, to be totally frank, with the presets I've tried, I truly don't hear anything amazingly different from the classic multiband.
Try rock music. And then switch between the two.
Quote:
This is why I'm frustrated about the inefficiencies being left in the classic code
As I said earlier this week, according to VTune the 'inefficiencies' use less than 1% of the total CPU load.

By the way, is BETA054 with the new MB still more expensive than the last released version 7.03 on your system? I'll do some measurement later on but on my pc I'm seeing a CPU load of 7% if I load Bojcha's preset - that's the lowest CPU load I've ever seen (for older presets it was always 9-11%). Those are numbers from Task Manager so not reliable - but they usually DO indicate if things are moving in the right direction (just not how much).
Author:  hvz [ Sat Mar 16, 2013 4:03 am ]
Post subject:  Re: Stereo Tool 7.03 BETA
BETA055:
Stand alone: http://www.stereotool.com/download/ster ... 04-055.exe
Winamp DSP: http://www.stereotool.com/download/dsp_ ... 04-055.exe
VST: http://www.stereotool.com/download/vst_ ... 04-055.dll

DONE:
- Multiband: Call Median Reset when loading a preset or hitting Reset
- Multiband: Optimize Median calculation (less frequent updates)
- Compressor side chain causes volume jumps (block overlap issue). -> Unable to reproduce... But this is scary! Earlier I also saw bad things happening to the 2nd channel. That's also gone now.
- Allow saving new MB / SB settings
- Weird spikes in signal when using Synchronize option for FM in stand alone version (DJ DOGGY)
- Still silence using Synchronize option to normal output (?)

PROBABLY DONE:
- Median display issue Probably fixed... But I haven't seen it happen yet so I'm not sure.

TODO:
- Add chain optimization (%) again
- Some default logical Band crossover settings for MB (every number of bands). Best way is probably to ask if defaults should be loaded when changing number of bands, with a 'don't ask again' checkbox. And of course after changing you can always use right click to reset each slider.
- Crash reported by Christian Steifen: If you change to a different category and quickly change something it crashes (VST version). Unable to reproduce, waiting for feedback - was this with BETA054?
- Check new MB limiters and compressors: Optimize to reduce low frequency artifacts. AWAITING FEEDBACK.
- Multiband compressor: Differentiate between Basic/Advanced/Extreme... modes
- Singleband Compressor: Differentiate between Basic/Advanced/Extreme... modes
- Convert low latency monitoring output to use new filters
- Update presets
- Change version number
- Release
Author:  Brian [ Sat Mar 16, 2013 4:30 am ]
Post subject:  Re: Stereo Tool 7.03 BETA
Quote:
Quote:
As for the new multiband stuff, to be totally frank, with the presets I've tried, I truly don't hear anything amazingly different from the classic multiband.
Try rock music. And then switch between the two.
What do you define as "rock"? I have everything from The Rolling Stones, Pink Floyd, Aerosmith, Led Zepplin, Foreigner, Night Ranger, Def Leppard, Van Halen, Scorpions, and Metallica...to stuff like Korn, Disturbed, Marilyn Manson, Dope, Lacuna Coil, Linkin Park, and Rob Zombie.
Quote:
Quote:
This is why I'm frustrated about the inefficiencies being left in the classic code
As I said earlier this week, according to VTune the 'inefficiencies' use less than 1% of the total CPU load.
I don't think you were testing the same thing as I'm talking about, unless you were talking about using all "classic" stuff and NONE of the "new" stuff. In other words, 7.00 to 7.02 or 7.03, but NOT 7.04 beta.

If you truly have gone back in and reworked the classic multiband to remove the additional loop iterations, then I do not recall you mentioning that you had done that. I think you had mentioned that Noise Gate was another spot where the same thing happens, and I do not recall work being done on that, although I could have overlooked it.
Quote:
By the way, is BETA054 with the new MB still more expensive than the last released version 7.03 on your system? I'll do some measurement later on but on my pc I'm seeing a CPU load of 7% if I load Bojcha's preset - that's the lowest CPU load I've ever seen (for older presets it was always 9-11%). Those are numbers from Task Manager so not reliable - but they usually DO indicate if things are moving in the right direction (just not how much).
As I've said, I have not been attempting to use the new MB, because I just do not hear enough of a difference, if any difference at all. What you need to do with that statement is not take offense to it, but realize that I'm saying that your previous work was good enough for me.

If I am to test it though, you'll need to give me what options enabled/disabled in that new stuff is what you believe is a reasonable approximation of the same audio fidelity as the classic multiband at 4096 samples and anywhere from 70-100 quality.

I will try to use it tomorrow or on Sunday. My Mom has not been feeling well, and I've been driving her places because she does not feel like she is safe to drive (jittery), thus I have not had a lot of time for extended listening.
Author:  Brian [ Sat Mar 16, 2013 5:42 am ]
Post subject:  Re: Stereo Tool 7.03 BETA
First attempt at listening to the new stuff:

I picked "80s style". I disabled declipper, oversampling, reduced Strictness (CPU) from 13 to 6, and enabled Hard Limit.

Track I unfortunately picked was Pink Floyd - Echoes.

I began playing, and it sounded like there was a lot less "noise". Going back and looking at that now that I've stopped playing that unfortunate track selection, yep, the noise threshold is significantly higher, meaning more things are considered to be "noise".

Why was the track selection "unfortunate"? I wanted to save the settings, so I clicked Save -> All Settings (Export). Well, to perhaps convey how bog slow this new GUI and new everything has been, when I hit the export, it was at 2m 45s into the track. The track length is 23m 28s. The file saved and I finally got control back after the track had completely finished playing. Yep, 20 minutes and 43 seconds to write 44500 bytes of data. Multitasking? I think not.

Created: Yesterday, March 15, 2013, 11:44:48 PM
Modified: Today, March 16, 2013, 12:05:28 AM

Well, per the OS, it was actually 20 minutes and 40 seconds.

So, well, the track's highs do sound "crisper" with the new stuff, but the preset also has some serious effort in making sure highs are loud. The bass, on the other hand, is significantly weaker.

My preset that I'm in the process of finalizing that uses the Classic stuff has a load examined strictly in Task Manager of approximately 70%. The new stuff? 80-85%. Both values were taken at 4096 samples, 100 quality, and the same strictness settings in Advanced Clipper.

In my opinion, based on previous statements, Hans, you likely really like that preset because of the highs. I personally think that I can likely increase the amount of highs in mine without sacrificing too much bass to where the two are roughly the same in the amount of highs content.

One thing I did NOT yet test for is to see if it is only an illusion of brighter highs due to a higher RMS level. I suspect 80s style has a higher RMS, because it seemed to stay pegged at full more than mine did. If that's the case, I'll try adjusting the volume with Simple Clipper and get the two presets at the same approximate RMS and see what it sounds like.

The takeaway here though is that something that several people have said they really like can most likely be replicated on my system with the classic stuff for less CPU load, at least as far as my ears are concerned. Maybe the new stuff looks fantastically better on a scope, but I don't listen with a scope.

Edit:

Yep, 80s style is louder, which means my level of highs could be similar. I'll have to recalibrate to find out...after sleep.

-----------------------------------------------
Statistics for: 80s_Pink Floyd - 06 - Echoes.wav
Number of Samples: 62100335
-----------------------------------------------

left right

Peak value: -0.45 dB --- -0.45 dB
Avg RMS: -6.17 dB --- -5.99 dB
DR channel: 4.40 dB --- 4.30 dB
-----------------------------------------------

Official DR value: DR4
===============================================

-----------------------------------------------
Statistics for: ClassicCobalt_Pink Floyd - 06 - Echoes.wav
Number of Samples: 62100335
-----------------------------------------------

left right

Peak value: -0.45 dB --- -0.45 dB
Avg RMS: -7.80 dB --- -7.66 dB
DR channel: 5.99 dB --- 5.71 dB
-----------------------------------------------

Official DR value: DR6
===============================================
Author:  DJ-DOGGY [ Sat Mar 16, 2013 9:33 am ]
Post subject:  Re: Stereo Tool 7.03 BETA
Reported Spikes in the signal seems to be gone .
Phase diff ( yellow circle ) seems to be OK, but sometimes when i set the phase to 90 deg and then turn it back to 0 this difference is stay like in the yellow circle and after some time slowly disappears(very slowly). Again when i disable the MB it completely disappears and when turn it back on, phase diff is there and continue to slowly disappear.

For the Synchro....Sometimes i see a very little spikes down to the last db`s of the MPX GUI scale ( it appears only in the L+R side ) .
When disable it, this is gone... think that something is still happened in this option .

Something in new sliders push them to a step, that i cannot set it exactly to 1.00 . It`s 0.997 or 1.003.... why is that ?
Page 86 of 102 All times are UTC+02:00
Powered by phpBB® Forum Software © phpBB Limited
https://www.phpbb.com/