Polarity Inversion Overdrive

[Marco Raaphorst]

All distortion devices in Reason are asymmetrical. This means that both sides of the signal (+ and -), are not treated equally. I discovered this by experimenting and listening. It also means that you can get a different sound out of these saturation devices in Reason simply by inverting the polarity of the signal!

Please read my blogpost and listen to the included audio demo.

[selig]

Interesting topic, as always!

There are a couple of explanations for this, one being that we actually only detect the change (in air pressure) when we switch the examples back and forth as you have done in your blog, and another being that most speakers are not 100% symmetrical to begin with which leads to the difference some claim to hear.

In your example I "THINK" I may be hearing a difference, but that was only after listening more closely the second time around. A double blind test would probably be totally impossible for either of us to pass since the sounds aren't played back to back. 

There has been quite a lot posted on "absolute phase" over the years, with the only real conclusions I've found being "maybe" or at best "that depends". 

In this case the differences appear to be quite infinitesimal to my ear, and that's listening to the guitar soloed - in a track these differences would be imperceptible IMO).

Just to put things in perspective, if you changed pick position by even a fraction of an inch it would probably have a MUCH larger affect on the tone than polarity. 

[Marco Raaphorst]

:s0403:

There's a difference in sound which can also be detected with the spectrum analyser in Reason. For example the low end is different. 

I must say that with a live guitar the sound to me sounds different. I found out about it that way. Then I did my own blind tests and I could prove it. I also did some test with just a bassdrum and the Pulveriser and there was a very obvious difference in low end. There's some interesting things going on here, but sometimes indeed, subtle. 

You are probably right that in the mix the differences are gone. As most subtleties are.

The polarity inversion on distortion is also used by Tchad Blake who I admire. He uses that in a parallel (phase reversed) so only distortion will be added, not the direct signal. I have build a few Combinator patches with these kind of technique. These are not subtle but very useful...

And yes, pick position is most important. Yesterday I was playing Amp and Cab in Ableton which I believe is really bad guitar simulation, but in a way it was working for me. It sounded ok and I could play some chords. I admit that in general I way too much of a perfectionist Maybe I'll do a lofi album one day. Being a fan of Latin Playboys...

[selig]

OK, let's do a little "logic" test here.

1- If two files completely cancel out when polarity inverted and summed, we can assume they are the exactly the same, agreed?

2- If you add your two audio examples together, the original and the polarity inverted, they will cancel 100%, right?

3- If so, how then can they be different in any way?

(Did I miss anything? I'm just asking the questions here, as I don't yet have a solid opinion on this subject!)

[selig]

:s0403:

There's a difference in sound which can also be detected with the spectrum analyser in Reason. For example the low end is different. 

I must say that with a live guitar the sound to me sounds different. I found out about it that way. Then I did my own blind tests and I could prove it. I also did some test with just a bassdrum and the Pulveriser and there was a very obvious difference in low end. There's some interesting things going on here, but sometimes indeed, subtle. 

Again, the test may lead to the conclusions. If you leave silence between the examples when doing an A/B, can you still pick out the inverted version 100% of the time? 

You are probably right that in the mix the differences are gone. As most subtleties are.

The polarity inversion on distortion is also used by Tchad Blake who I admire. He uses that in a parallel (phase reversed) so only distortion will be added, not the direct signal. I have build a few Combinator patches with these kind of technique. These are not subtle but very useful...  

That effect is so difficult to setup because the distorted signal is dynamically compressed as well as having additional harmonics. I've tried to "remove" the original signal from a distorted signal many times (didn't know Tchad did this), and that's the big rub - there's no single level where you'll get constant canceling since the dynamic range is different on one signal. It would be like trying to extract the distortion added by a compressor by the same approach. In my experience you get too much level change as the levels line up and cancel, then move apart and get louder (many times a second as the track plays). Hope I'm being clear here, really wish we were sitting side by side and comparing audio examples here!!!

And yes, pick position is most important. Yesterday I was playing Amp and Cab in Ableton which I believe is really bad guitar simulation, but in a way it was working for me. It sounded ok and I could play some chords. I admit that in general I way too much of a perfectionist Maybe I'll do a lofi album one day. Being a fan of Latin Playboys...

[Marco Raaphorst]

1- If two files completely cancel out when polarity inverted and summed, we can assume they are the exactly the same, agreed?

Yes

2- If you add your two audio examples together, the original and the polarity inverted, they will cancel 100%, right?

yes

3- If so, how then can they be different in any way?

they can't, they are the same. but that's not the case here. if you split a signal and send one into a Pulveriser and another one with reversed polarity into another Pulverser they won't cancel. so both versions are not affected in the same way by the Pulveriser. 

[Marco Raaphorst]

That effect is so difficult to setup because the distorted signal is dynamically compressed as well as having additional harmonics. I've tried to "remove" the original signal from a distorted signal many times (didn't know Tchad did this), and that's the big rub - there's no single level where you'll get constant canceling since the dynamic range is different on one signal. It would be like trying to extract the distortion added by a compressor by the same approach. In my experience you get too much level change as the levels line up and cancel, then move apart and get louder (many times a second as the track plays). Hope I'm being clear here, really wish we were sitting side by side and comparing audio examples here!!!

Yes, I agree that it's very difficult. I should add that I am not trying to totally remove the direct channel just that I love the sound of phase inverting an saturation channel and add that as send or parallel channel. You can mix in more dirtyness that way which sounds different from a non inverted channel. With the Pulveriser you can do a lot of the things I love. 

[selig]

1- If two files completely cancel out when polarity inverted and summed, we can assume they are the exactly the same, agreed?

Yes

2- If you add your two audio examples together, the original and the polarity inverted, they will cancel 100%, right?

yes

3- If so, how then can they be different in any way?

they can't, they are the same. but that's not the case here. if you split a signal and send one into a Pulveriser and another one with reversed polarity into another Pulverser they won't cancel. so both versions are not affected in the same way by the Pulveriser. 

IF they ARE the same, as you say above, then why do you hear a difference? Could be your speakers aren't symmetrical?

Is what you describe above not different from what you first described? In your original audio example, you have taken the output of one Pulveriser and inverted the polarity, then switched it back (and repeated), right? 

BTW, two Pulverisers with 50% "dirt" cancel almost 100% leaving about -90 dBFS residual level. Pretty damn close in my book - is this different than what you described?
[edit for clarity: split the signal, inverted polarity going INTO the second Pulveriser, summed both Pulverisers to mono, near perfect cancellation.]

[Marco Raaphorst]

don't know how to quote on this forum. I do it like this;

IF they ARE the same, as you say above, then why do you hear a difference? Could be your speakers aren't symmetrical?

Is what you describe above not different from what you first described? In your original audio example, you have taken the output of one Pulveriser and inverted the polarity, then switched it back (and repeated), right? 

BTW, two Pulverisers with 50% "dirt" cancel almost 100% leaving about -90 dBFS residual level. Pretty damn close in my book - is this different than what you described?
[edit for clarity: split the signal, inverted polarity going INTO the second Pulveriser, summed both Pulverisers to mono, near perfect cancellation.

They are not the same. The signal don't cancel. And yes I can hear the difference with a silence in between them. Try it on bassdrums, some will sound lower because of inversion. You can check the spectrum for it. It's not huge but it is something. I am afraid I am way too much of a Steely Dan freak

The Dirt of the Pulveriser needs to be indeed be opened because the overdrive is asymetical. It's not huge, but it is there. A sine wave might be way too simple to use hear. I have only used my guitars for these kind of effects. 

I am not sure if my speakers are asymetrical. I hope not. I do tests with frequencies sweeps a lot. I don't hear unregularities. And I do check these things on 2 headphones as well. So if everything is broken, it would be wrong.

The Polarity INV trick is rather nice on guitar. It really work by simply playing on of my Combi's and switching between + or - polarity. The sound is different and in my opinion.

[Marco Raaphorst]

let's do this test.
download this testfile 
2 pulverisers with the same settings, one has an inverted input. with a silence in between switching between both channels. do you hear the difference? I hear it. one sounds lower! I  am now working on headphones. both channels will cancel so they should be the same, so no difference should be heard, but I am hearing it. this is weird and I can't explain.

and I must say I have tested these things over and over again. is this caused by the speakers? all my speakers inc 2 headphones systems are causing this then. and for guitar it's really like your biasing the tubes or using different speaker damping... mmm, interesting

[pedrocaetanos]

Please read my blogpost and listen to the included audio demo.

Marco, your blog post seems to have vanished...

[Marco Raaphorst]

Yes, both the post and soundcloud audio I have put offline (on private). I will do a proper post soon but I have been testing and working on this for a couple of days. New conclusions will soon be available

[Jagwah]

"Polarity Inversion Overdrive"

^^^ Damn those words are fancy! 

[rvman]

When I listened to the examples on the blog, I couldn't tell a difference. Granted my hearing is not fantastic and I was using ear buds.

I would get more of a difference with picking techniques and different picks.

[motuscott]

I lurves me some Latin Playboys.