Stereo Tool

The program used is called "Daqarta" and vb-audio

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

My radio station
http://amradio.ddns.net

Hm... my 50Hz sinewave looks definitely better. I generated this one in Goldwave, then saved as .wav.
I got only little bumps (yet audible). This time I used NO (Unveil).



But audio-wise it sounded similar with MPX out.

EDIT: and it's advanced clipper that causes this one. I guess it must introduce some distortion by nature, but IMD?

http://www15.zippyshare.com/v/5169271/file.html (AC-on - 50Hz sinewave is audibly modulated with some other tone, combined make the beat-like phenomenon)
http://www15.zippyshare.com/v/97025526/file.html (AC-off)

when I turn the advanced clipper off completely - IMD is almost completely(?) gone in music too!

Confirmation here:
http://www26.zippyshare.com/v/37465802/file.html

This one is demodulated MPX again Unveil FM with AC off.
I used only simple clipper (100%) and pre-limiter (175%)

Intresting ... The declipper found the midle samples as being clipped and not the peaks..... Btw i will do some tests with the simple clipper.

Kostas

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

My radio station
http://amradio.ddns.net

Glad someone started this. Also i'm glad that sent me this track last night so it's "hearable" for almost everybody.
There is also one IMD, or how i call it "voice vibration" that seems few people hears, and i can hear it only in ST.
To be honest with latest versions i am hear it much less, nut this one is bit different.
I tried BBP and depend on preset used there is also problem there, bit less, but it's there.

Would like to hear same on Orban or Omnia but i bet it would be there too and probably bit less then in ST.

In my latest FM preset i really need to set clipper lower to effect gone 100%.

..

_________________
control point
control point2

Oh really? I have some examples - not the same track, but... you'll still be interested in these:

http://www4.zippyshare.com/v/81778160/file.html (Vaya Con Dios track)
http://www74.zippyshare.com/v/37362489/file.html (speaker's voice and background music)

It's a station nearby the place I live, Złote Przeboje on 91.2 MHz (Katowice, Poland). They have EXTREME intermod of a "vibrating voice" you're talking about. Actually it's a bit similar to this in ST. I have no idea what they're using though (In fact I have been trying to contact their tech department for quite some time with no success). Apart from that, they have the worst sound I've ever heard on-air. Surprisingly - ST is as clean as a crystal on this Vaya Con Dios track:
http://www65.zippyshare.com/v/55670713/file.html

This one on the other hand:
http://www50.zippyshare.com/v/40454995/file.html

comes from the Leif's torture test sample for Optimod 8500FM (Impact preset, 75us).
It's here: http://www.totradio.com/mpx/torturecd/flac/

This distortion in particular is so annoying to me that I recognize it very, very well...

What i have seen over the years ST (and of course all other processors) have songs that have no problem processing and songs that have... For example ST is crystal clear on this song but other processors aren't at all. Another example is that st can generate a VERY clean and open sound but the time you hit it with a sine-wave it "distorts"

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

My radio station
http://amradio.ddns.net

Well, it's definitely the clipper. Because in almost all other filters I have done things to avoid (or at least greatly reduce) this effect, but in the clipper I don't know how to. Well, that is not exactly true - see below.

In case of the test above with Composite Clipper, I just ran some tests here and I've noticed that if you play a 50 Hz tone and play it loud such that it hits the clipper threshold, you can also hear the effect of the stereo pilot and RDS (put them at level '0' and the 50 Hz tone comes out better).

Based on what I've heard in several comparisons, I would say that this problem is less prominent in ST than in almost all other processors (which seems to be confirmed by Slawomir's post). Although after hearing the effect that the Quality slider has on this (I had not expected that) I realized that there is one thing that's specific to ST that might be making it worse (I don't think any other processor uses that same technique). (Note to Bojcha: You were probably right with what you said a few days ago). Note that this does not mean that it can be easily solved...

By the way: There are a few sliders in ST that were made to reduce this effect: Under Advanced Bass Distortion Protection, see the "Constant tone distortion protection" sliders. The first one (Smoothe mid frequencies) affects the CPU load, the 2nd one (Peak detection steepness) controls how easy a tone is recognized as a loud constant sound. If you set it too high it will try to filter out far too much of this effect, which makes the sound of the mid frequencies softer (and you don't want that - because you'll need more clipping to reach the same loudness level). I did a quick test with both sliders at the maximum setting, and it reduced the effect a bit, so the default values are probably not optimal. And since I'm working on ABDP anyway, I'll take another look at that code to see if I can improve it.

That would be yes and no. Apart from that one particular radiostation, it's not that common. Especially this "vibrating voice" one.
Usually, if it occurs, it sounds more-less like at the begining of this sample:
http://www53.zippyshare.com/v/6242239/file.html - recorded off-air (processor unknown) - it simply sounds dirty. Maybe you'll judge the processor from pure MPX https://drive.google.com/file/d/0B1oKe7 ... sp=sharing

and here ST:
http://www50.zippyshare.com/v/86951174/file.html - distinct vibration in the voice (I used old multiband preset on purpose here, just to demonstrate the extremes).
http://www18.zippyshare.com/v/97084409/file.html - stock Unveil preset (FM) - still there, just a bit less. Disappears completely when more sounds of complex spectrum are present (especially beat that occurs regularly).

And I think Bojcha is right on this one too, it indeed appears to be a bit different in ST (maybe it reagrds other part of the spectrum). Hence, the distortion texture is also different (vibrating voice) and specific for ST. Like if it was "modulated" up and down by single, really low frequency tone.

It's also interesting that it occurs in open-sounding presets too but I guess, since there is no limiter, clipper is heavily utilized anyway).

That is indeed the case. This problem occurs mainly when there are asymmetrical sounds with a low bass. Either the bass or the mids can be asymmetrical - for bass this is why I made the "Remove asymmetric bass" checkbox.

What I think happens in most other processors is this:
- Bass, say 35 Hz
- Tone
- Result: Tone gets a drop at every peak and valley of the bass, so 2*35 = 70 times per second. And you hear harmonics of the clipper come and go in the total frequency range.

What happens in Stereo Tool:
- Bass, say 35 Hz
- Tone
- One of the two is asymmetrical
- Holes get punched on one side. This causes drops 35 times per second.

Now of course that would happen in other processors as well. Because in ST there's no harmonics distortion, the distortion with drops per second is far more noticeable than the 70 times per second - in most other processors (excluding BBP and Omnia 9) I would expect the opposite.

One thing that strongly affects asymmetry is the phase rotator, which differs from what other processors use. This might very well explain why ST has problems with other tracks than other processors.


The real problem for ST in this situation is that in case of asymmetry the bass behaves as if it's bass at half the frequency. Which means that a tone at 40 Hz has the same effect as a symmetrical tone at 20 Hz. The bass clipper removes very low bass freqs to avoid this effect, but it doesn't - or at least not enough - take asymmetry in the mids into account.

A possible solution would be to make the mids symmetrical by clipping them harder (for bass there's already a checkbox, so you can just turn that on). The big disadvantage of doing this for mids is that you'll loose loudness. And add more clipping.

I was so curious what the exact frequency of the vibration was (in the Bono's sample), that I physically measured it. It was so strong, that I could see it on a 250 Hz highpassed sample. The vibration turned out to be very low - about 45 Hz. And indeed it can be the half of the bass freq.



audio:
http://www9.zippyshare.com/v/87138567/file.html