Quantcast
Separating frequencies and then reversing phase wow try it - Gearspace.com
The No.1 Website for Pro Audio
Separating frequencies and then reversing phase wow try it
Old 30th December 2016
  #1
Lives for gear
 
🎧 10 years
Separating frequencies and then reversing phase wow try it

I just stumbled upon a super strange trick and wondered if others were trying it. I have NO idea how it works??? I have a pair of Coles 4038s for overheads on a kit. I high pass them at 60hz simply because I want my low end in the centre, not off to the sides. These mics have LOTS of low end and due to mic placement differences in low end in my room, it can sound a bit lop sided.

BUT I stumbled across the idea of duplicating ONE of these overhead tracks and low pass filtering it so only the stuff below 60hz is there. I put that track in the middle in mono. Weirdly, it sounds fantastic when I reverse the phase on that track. So once again, I have the two overhead tracks high passed at 60hz and panned left- right. Then a duplicate of ONE of the tracks I have low passed (leaving only low end under 60) in mono in the centre. I reverse the phase on this track and it sounds awesome with all that controlled low end to add back to taste! Am I the first who has tried this? Surely not! Can anyone understand why it works? I would have assumed it would be a phase nightmare but it works super well.
Old 30th December 2016
  #2
Lives for gear
 
Funny Cat's Avatar
 
🎧 10 years
Yeah this does work great at times. Sometimes not so great. There is actually a plugin which works on this principle too. It's called "Christortion". Made by an independent developer. It's old and separates the frequency into something like 7 bands then you can alter the phase from like 10% to 100% inverted on any given band. Makes for some real cool sounds. I like it on drums. There are a few engineers I know doing great work who play with parallel inverted phase on drums for unique tones too. I "think" I read that Tchad Blake does something similar on kick? Good stuff.
Old 30th December 2016
  #3
Gear Addict
 
James Talus's Avatar
 
🎧 10 years
I'll have to try this... what happens when you low pass the second overhead, move to the center, and invert?
Old 30th December 2016
  #4
Gear Addict
 
🎧 5 years
Depending on the Q, and with the polarity (not phase) reversed, what you may be hearing a is a bump around the crossover point.
Old 30th December 2016 | Show parent
  #5
Tokyo Dawn Labs
 
FabienTDR's Avatar
 
🎧 5 years
Quote:
Originally Posted by Funny Cat ➑️
It's old and separates the frequency into something like 7 bands then you can alter the phase from like 10% to 100% inverted on any given band. Makes for some real cool sounds.
Worth mentioning that the exact same results result could also be achieved more directly via straight forward EQing.
Old 30th December 2016 | Show parent
  #6
Lives for gear
 
Funny Cat's Avatar
 
🎧 10 years
Quote:
Originally Posted by FabienTDR ➑️
Worth mentioning that the exact same results result could also be achieved more directly via straight forward EQing.

Are you certain of this? The Christortion plugin allows phase shifts in increments. Can straight forward EQing emulate this? It sounds quite different to my ears but I haven't examined this in scientific terms by any means. I would also add that using a parallel channel allows the original to stay in tact and unaffected while bringing the phase shifted channel up underneath.
Old 30th December 2016
  #7
Tokyo Dawn Labs
 
FabienTDR's Avatar
 
🎧 5 years
At the end of the day, the complete effect can be characterized by a simple impulse response, at least given the description above (crossover and polarity flip). It is a linear system. So it can easily reproduced with Eqing (or convolution with an impulse).

Christortion is a rather crude polynomial waveshaper with full access to the polynomial coefficients (creates harmonics). You must have a different product in mind? This one doesn't shift phase or split bands.
Old 30th December 2016 | Show parent
  #8
Lives for gear
 
Funny Cat's Avatar
 
🎧 10 years
Quote:
Originally Posted by FabienTDR ➑️
At the end of the day, the complete effect can be characterized by a simple impulse response, at least given the description above (crossover and polarity flip). It is a linear system. So it can easily reproduced with Eqing (or convolution with an impulse).

Christortion is a rather crude polynomial waveshaper with full access to the polynomial coefficients (creates harmonics). You must have a different product in mind? This one doesn't shift phase or split bands.

I'm not a plugin designer so I may be explaining this wrong, however the plugin divides the signal into ten actual bands and the description states "It is also possible to invert the phase of each harmonic. This concept guarantees a very surgical control over the distortion produced."

You can also control the amount of phase "inversion" introduced with the sliders. It does not sound like straight forward EQ when I compared it. If this isn't the same/similar process as the OP (only in a plugin) then pardon the OT.





P.S. The Tchad Blake? Kick drum technique takes a copy of the kick, EQ's and flips the phase then blends it back in. I know another fantastic engineer who does this. Not sure how or why this works (seems like it would be a mess) but it just works. Does not sound like straight up EQ to me either.



https://www.kvraudio.com/product/chr...hristian_budde
Old 31st December 2016 | Show parent
  #9
Lives for gear
 
🎧 10 years
Quote:
Originally Posted by jamestalus ➑️
I'll have to try this... what happens when you low pass the second overhead, move to the center, and invert?
It actually works best (in my room) to only use one mic as the low mono one. I suspect the phase is better just using one. Thanks for all responses very interesting. I'm surprised more people don't discuss it. Very good way of centering any low end in overheads for a start.
Old 31st December 2016 | Show parent
  #10
Tokyo Dawn Labs
 
FabienTDR's Avatar
 
🎧 5 years
Quote:
Originally Posted by Funny Cat ➑️
I'm not a plugin designer so I may be explaining this wrong, however the plugin divides the signal into ten actual bands and the description states "It is also possible to invert the phase of each harmonic. This concept guarantees a very surgical control over the distortion produced."

https://www.kvraudio.com/product/chr...hristian_budde

It does not say this. At no point does it mention "the plugin divides the signal into ten actual bands". Where did you pick that up? It doesn't appear at all in the product description.

It clearly says that it produces up to 10 harmonics, not bands! And you can control their amplitude and phase. Phase of the harmonics, not of the signal itself.

There is no crossover inside this plugin. Christortion is a pure waveshaper, a distortion device. It is not related to the OP at all.
https://www.kvraudio.com/product/chr...hristian_budde

Following the OP's description, we have a crossover splitting the signal into bands and means of flipping the phase of each band. That's nothing else than EQing. Overcomplicated, crazy unpredictable back flip EQing. But still EQing.
Old 31st December 2016
  #11
Lives for gear
 
Funny Cat's Avatar
 
🎧 10 years
@ FabienTDR


The mention of bands was mine not from the product description. It was my personal laymen interpretation of what the sliders were.

The part from my post in quotes is the part from the plugin product page. By the way, "quotes" means it's a direct reference. Understand how that works?

Maybe drop the pedantic and condescending tone. You're a smart guy clearly but perhaps you are not using your ears.

The method the OP is describing is not simple straight forward EQing...because it yields a different result. It does NOT sound the same. This can be proven easily.

Folks like Tchad Blake AND my other A-list mixer friend would not use it to mix top selling hits if it was a matter of "overcomplicated, crazy unpredictable back flip EQ'ing".
Old 31st December 2016 | Show parent
  #12
Gear Addict
 
🎧 5 years
Quote:
Originally Posted by FabienTDR ➑️
Following the OP's description, we have a crossover splitting the signal into bands and means of flipping the phase of each band. That's nothing else than EQing. Overcomplicated, crazy unpredictable back flip EQing. But still EQing.
I know that EQ is not linear phase (except for some special digital ones,) but I don't see how you can call complete polarity reversal of one band of frequencies EQ. Give us an example of what settings you would use on an EQ to provide the same effect as the OP describes.
Old 2nd January 2017
  #13
Lives for gear
 
haysonics's Avatar
 
🎧 5 years
Quote:
Originally Posted by Ebeowulf17 ➑️
I know that EQ is not linear phase (except for some special digital ones,) but I don't see how you can call complete polarity reversal of one band of frequencies EQ. Give us an example of what settings you would use on an EQ to provide the same effect as the OP describes.
You are (inadvertently) asking Fabien to create complicated EQ modelling for you. I imagine you would need to advise what percentage (to a tenth, eg 56.8%) of the polarity reversed mono track you would be mixing in, and what your cut-off frequency is. If your point had been about the simplicity of the trick the OP describes then I would understand but to be fair to Fabien he used the word "overcomplicated" in terms of modelling the result. How much are you willing to pony up for this new plugin you are asking him to create?
Old 2nd January 2017 | Show parent
  #14
Gear Maniac
 
🎧 5 years
Quote:
Originally Posted by Funny Cat ➑️
Maybe drop the pedantic and condescending tone.
He is not being pedantic, nor condescending. He is trying to explain something.
Old 2nd January 2017 | Show parent
  #15
Lives for gear
 
Funny Cat's Avatar
 
🎧 10 years
Quote:
Originally Posted by CaptainMarc ➑️
He is not being pedantic, nor condescending. He is trying to explain something.

Yes, he is trying to explain it in a "pedantic and condescending" manner. He even started his comments stating how Christortion is "rather crude", which could be taken as a bit offensive especially coming from another plugin maker.

He later went on to say it's "nothing else than EQing. Overcomplicated, crazy unpredictable back flip EQing. But still EQing." which is most definitely condescending...especially to the OP and anyone else who has tried it and thinks this is a valid mixing technique to achieve a particular result.

I let that slide and even came back respectfully "asking" are you sure simple EQing can do this just for the sake of deeper discussion and elaboration from him (I personally don't think it's possible based on my rudimentary testing).

I first discovered the method from my buddy who mixes a bunch of top 40 pop etc. Then I started reading about others who use this as it was so new to me. Maybe you should try it for yourself and then chime in to the thread with your own real world assessment.

Last edited by Funny Cat; 2nd January 2017 at 05:00 PM.. Reason: Edited for clarification
Old 2nd January 2017 | Show parent
  #16
Gear Addict
 
🎧 5 years
Quote:
Originally Posted by haysonics ➑️
You are (inadvertently) asking Fabien to create complicated EQ modelling for you. I imagine you would need to advise what percentage (to a tenth, eg 56.8%) of the polarity reversed mono track you would be mixing in, and what your cut-off frequency is. If your point had been about the simplicity of the trick the OP describes then I would understand but to be fair to Fabien he used the word "overcomplicated" in terms of modelling the result. How much are you willing to pony up for this new plugin you are asking him to create?
I didn't mean to ruffle any feathers, and I certainly don't want anyone designing a new plugin for me.

Fabien said that "the exact same results result could also be achieved more directly via straight forward EQing." To me, that implies that Fabien believes the EQ approach to be simpler or easier.

I would've thought that hitting very specific phase relationships like that with only EQ would be quite difficult, and be the realm of electrical engineers or software developers. In my opinion, if it's simpler or easier, there should be simple, easy to share parameters, like cut x dB at y cutoff frequency with z octave bandwidth. If it is in fact so complicated that I should have to pay someone to develop a new plugin for it, then it seems to me that the OP's method is much simpler and easier. Anyone working itb, and in fact anyone working on hardware with a respectable studio console, has enough simple HP and LP filters and polarity switches to try the OP's method.

So, if the OP suggests a simple method that anyone can try at home, what benefit is there in pointing out that there's a "straight forward" way to do it with just EQ, if "straight forward" secretly means hire a plugin developer?
Old 2nd January 2017 | Show parent
  #17
Gear Maniac
 
🎧 5 years
Quote:
Originally Posted by Funny Cat ➑️
Yes, he is trying to explain it in a "pedantic and condescending" manner. He even started his comments stating how Christortion is "rather crude", which could be taken as a bit offensive especially coming from another plugin maker.

He later went on to say it's "nothing else than EQing. Overcomplicated, crazy unpredictable back flip EQing. But still EQing." which is most definitely condescending...especially to the OP and anyone else who has tried it and thinks this is a valid mixing technique to achieve a particular result.

I let that slide and even came back respectfully "asking" are you sure simple EQing can do this just for the sake of deeper discussion and elaboration from him (I personally don't think it's possible based on my rudimentary testing).

I first discovered the method from my buddy who mixes a bunch of top 40 pop etc. Then I started reading about others who use this as it was so new to me. Maybe you should try it for yourself and then chime in to the thread with your own real world assessment.
Christortion is crude. That doesn't mean it's bad, crude can be good.

"Overcomplicated crazy unpredictable backflip EQing" - this is not condescending, it is entirely accurate. Anybody who knows how to build an eq can see this. This doesn't mean the method is without merit, if it works, it works. Personally, when I like what I hear, it's a keeper, regardless of how I achieved it.

Rudimentary testing: two overhead tracks, hard panned L & R, HPF at some cutoff point. Take one of the tracks, pan in the centre, LPF at same cutoff point, adjust eq to taste, applying particular care to the area around the cutoff point. Voila! It's hardly rocket science, it's hardly a new technique.

Similarly with a mono kick track. As Fabien said "At the end of the day, the complete effect can be characterized by a simple impulse response, at least given the description above (crossover and polarity flip). It is a linear system. So it can easily reproduced with Eqing (or convolution with an impulse).". I mean, this is totally obvious, isn't it?

Anyway, if you can get results quickly by eq on a flipped phase copy, that's great. I would also recommend this plug in for even more exciting phase adventures.

You need to chill mate. Nobody is having a go at you. People are just trying to explain stuff.
Old 2nd January 2017
  #18
Tokyo Dawn Labs
 
FabienTDR's Avatar
 
🎧 5 years
Funny Cat, I really didn't mean to be offensive or arrogant.

Allow me to clarify my posts above. Carefully following the OP's scenario, we essentially have:

A. Two hard panned, high passed mono tracks.
B. One channel, say, R being low passed and mixed equally into the above tracks.

Now, before we flip the phase of B, let's consider a second what we have now.
Given the LP filter in B is the exact inverse of the HP in A (say, both are first pole and set to same cutoff), the right channel will now be equivalent to the original input (no changes at all). The left channel will now contain the LF of the right channel, and given a high correlation between L and R, will sound very much like the original. If the correlation is low (high difference between L and R, in this case under 60Hz), we might get some bump or strong cancellations in the left channel.

More clearly. The result will be the original signal + an unpredictable bass in the left channel. You can do the same with a dual mono EQ.

Now let's flip the phase of B, as described.

Given the same assumptions as above (HP/LP are inverses of each other), With B now polarity flipped, the result will be:

L = the same unpredictable Bass region as above.
R = a huge bump over the cutoff area (with "holes" around it).

Again, you could have achieved the same by simply launching a dual mono EQ and blindly messing around with the lowest bands. In doubt, try the same with white noise and measure the output freq response.



************************************************



As for the Christortion process, it does something very different. Give me a min to explain how it works, it will probably make clear what I meant with crude (I'll try in simple words and intentionally leave out a few details).

The process works like this, the plugin takes the input signal and generates various harmonics. It can generate 2nd harmonics, 3rd, 4th, and so on. Each harmonic can be specified in form of a coefficient describing how strong this harmonic will be. Each harmonic can also have certain polarity.

At runtime, the plugin simply does this ("*" stands for multiplication):


// second harmonic
y = x * x * secondCoefficient

// third harmonic
y = x * x * x * thirdCoefficient

// fourth harmonic
y = x * x * x * x * fourthCoefficient

(and so on...)

All partials are then simply being added to the original signal. Finito!

This is a super powerful construct called the Taylor series. It can approximate most type of functions. Most function such as log(), sin(), sqrt() have their own Taylor series.

Now the truly crude aspect is that the coefficients mentioned above are super sensitive and their effect hard to control. A bit like lottery. Try and error does not work. This would be somewhat equivalent to offering an EQ in biquad form. It's super flexible, but setting cutoff and Q properly is rocket science. That's why you typically have an additional layer in between giving these low level coefficients some "tactile sense" (say, by offering params such as "drive", "colour", "hardness", "symmetry", "shape" and so on).

I didn't mean to call your decision "crude" or "bad". Just saying that it's not what the OP did.
Old 2nd January 2017 | Show parent
  #19
Lives for gear
 
Funny Cat's Avatar
 
🎧 10 years
Quote:
Originally Posted by FabienTDR ➑️
Funny Cat, I really didn't mean to be offensive or arrogant.

Allow me to clarify my posts above. Carefully following the OP's scenario, we essentially have:

A. Two hard panned, high passed mono tracks.
B. One channel, say, R being low passed and mixed equally into the above tracks.

Now, before we flip the phase of B, let's consider a second what we have now.
Given the LP filter in B is the exact inverse of the HP in A (say, both are first pole and set to same cutoff), the right channel will now be equivalent to the original input (no changes at all). The left channel will now contain the LF of the right channel, and given a high correlation between L and R, will sound very much like the original. If the correlation is low (high difference between L and R, in this case under 60Hz), we might get some bump or strong cancellations in the left channel.

More clearly. The result will be the original signal + an unpredictable bass in the left channel. You can do the same with a dual mono EQ.

Now let's flip the phase of B, as described.

Given the same assumptions as above (HP/LP are inverses of each other), With B now polarity flipped, the result will be:

L = the same unpredictable Bass region as above.
R = a huge bump over the cutoff area (with "holes" around it).

Again, you could have achieved the same by simply launching a dual mono EQ and blindly messing around with the lowest bands. In doubt, try the same with white noise and measure the output freq response.



************************************************



As for the Christortion process, it does something very different. Give me a min to explain how it works, it will probably make clear what I meant with crude (I'll try in simple words and intentionally leave out a few details).

The process works like this, the plugin takes the input signal and generates various harmonics. It can generate 2nd harmonics, 3rd, 4th, and so on. Each harmonic can be specified in form of a coefficient describing how strong this harmonic will be. Each harmonic can also have certain polarity.

At runtime, the plugin simply does this ("*" stands for multiplication):


// second harmonic
y = x * x * secondCoefficient

// third harmonic
y = x * x * x * thirdCoefficient

// fourth harmonic
y = x * x * x * x * fourthCoefficient

(and so on...)

All partials are then simply being added to the original signal. Finito!

This is a super powerful construct called the Taylor series. It can approximate most type of functions. Most function such as log(), sin(), sqrt() have their own Taylor series.

Now the truly crude aspect is that the coefficients mentioned above are super sensitive and their effect hard to control. A bit like lottery. Try and error does not work. This would be somewhat equivalent to offering an EQ in biquad form. It's super flexible, but setting cutoff and Q properly is rocket science. That's why you typically have an additional layer in between giving these low level coefficients some "tactile sense" (say, by offering params such as "drive", "colour", "hardness", "symmetry", "shape" and so on).

I didn't mean to call your decision "crude" or "bad". Just saying that it's not what the OP did.



Thank you Fabien for taking the time to put together a well thought out, detailed and tempered reply. I appreciate it. I can't honestly say I follow the entire logic, I am no plugin designer obviously but I will play around with some more dual mono EQ's to see if I can get a similar effect.

One thing I will say is I have been trying to figure out how to get the low end extension like my buddy gets and he initially tipped me off by saying "play with EQ's and phase on the drums and just use your ears until it sounds right" which is how I stumbled on the OP's premise.

I had been toying with this for over a year and finally figured it out. I really spent a lot of hours trying this out on mixes and to my ears it just does not sound like any simple EQ I had tried previously, since there is actually an audible "distortion" (for lack of a better word) I can hear in the processed signal.

It's like the transients become softer and rounder due to the phase anomalies but in a really pleasing way while bumping and extending the low freq's in a very unique way. Maybe if I can find the file I'll post an example of the effect I'm speaking of from I mix I did last year. Again I will take more time to test this.

Last edited by Funny Cat; 3rd January 2017 at 02:13 AM.. Reason: grammar
Old 2nd January 2017
  #20
Lives for gear
 
Funny Cat's Avatar
 
🎧 10 years
@ FabienTDR



OK just so happens that I have a copy of a mix I did for an artist where I played with EQ's and phase on the low end on this laptop so have a listen and you'll see what I mean by "audible distortion" in the low end. Let me know if you think this is an effect a simple EQ can achieve in your opinion. Thanks and Happy New Year to you.





[EDIT] Sorry for some reason I can't upload the file. Says it uploads successfully but then the file disappears. :0/ I'll try again later...


[EDIT2] Still can't upload the file. Tried converting it to a lower bit rate mp3 and still no love. Any of the mods know what's going on with the file upload feature?

Last edited by Funny Cat; 3rd January 2017 at 02:18 AM..
Old 3rd January 2017 | Show parent
  #21
Gear Addict
 
🎧 5 years
Quote:
Originally Posted by FabienTDR ➑️
Funny Cat, I really didn't mean to be offensive or arrogant.

Allow me to clarify my posts above. Carefully following the OP's scenario, we essentially have:

A. Two hard panned, high passed mono tracks.
B. One channel, say, R being low passed and mixed equally into the above tracks.

Now, before we flip the phase of B, let's consider a second what we have now.
Given the LP filter in B is the exact inverse of the HP in A (say, both are first pole and set to same cutoff), the right channel will now be equivalent to the original input (no changes at all). The left channel will now contain the LF of the right channel, and given a high correlation between L and R, will sound very much like the original. If the correlation is low (high difference between L and R, in this case under 60Hz), we might get some bump or strong cancellations in the left channel.

More clearly. The result will be the original signal + an unpredictable bass in the left channel. You can do the same with a dual mono EQ.

Now let's flip the phase of B, as described.

Given the same assumptions as above (HP/LP are inverses of each other), With B now polarity flipped, the result will be:

L = the same unpredictable Bass region as above.
R = a huge bump over the cutoff area (with "holes" around it).

Again, you could have achieved the same by simply launching a dual mono EQ and blindly messing around with the lowest bands. In doubt, try the same with white noise and measure the output freq response.



************************************************



As for the Christortion process, it does something very different. Give me a min to explain how it works, it will probably make clear what I meant with crude (I'll try in simple words and intentionally leave out a few details).

The process works like this, the plugin takes the input signal and generates various harmonics. It can generate 2nd harmonics, 3rd, 4th, and so on. Each harmonic can be specified in form of a coefficient describing how strong this harmonic will be. Each harmonic can also have certain polarity.

At runtime, the plugin simply does this ("*" stands for multiplication):


// second harmonic
y = x * x * secondCoefficient

// third harmonic
y = x * x * x * thirdCoefficient

// fourth harmonic
y = x * x * x * x * fourthCoefficient

(and so on...)

All partials are then simply being added to the original signal. Finito!

This is a super powerful construct called the Taylor series. It can approximate most type of functions. Most function such as log(), sin(), sqrt() have their own Taylor series.

Now the truly crude aspect is that the coefficients mentioned above are super sensitive and their effect hard to control. A bit like lottery. Try and error does not work. This would be somewhat equivalent to offering an EQ in biquad form. It's super flexible, but setting cutoff and Q properly is rocket science. That's why you typically have an additional layer in between giving these low level coefficients some "tactile sense" (say, by offering params such as "drive", "colour", "hardness", "symmetry", "shape" and so on).

I didn't mean to call your decision "crude" or "bad". Just saying that it's not what the OP did.
Thanks for explaining your eq reasoning there. Much appreciated. It makes more sense to me the way you've explained it now. I apologize if I seemed flippant or combative before.

I do wonder if the polarity reversal in the OP's method might have effects in the mix beyond what you'd hear from just the overhead mics. Perhaps the mics were placed to achieve good phase relationships on snare, but kick benefits from the phase reversal. Of course, if that were the case, simply flipping the kick channel might make more sense. Dunno, just thinking out loud now.
πŸ“ Reply

Similar Threads

Thread / Thread Starter Replies / Views Last Post
replies: 5883 views: 1882054
Avatar for nobtwiddler
nobtwiddler 1 day ago
replies: 57 views: 6052
Avatar for johnjacobniles
johnjacobniles 29th March 2021
Topic:
Post Reply

Welcome to the Gearspace Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…

Forum Jump
Forum Jump