This should be a general discussion about phasedeletion (with absolute silence as effect) in the digital world:
My hypothesis:
You can absolutely phasedelete signals (to silence) only with exactly equal ones (which means: same sound, because signals are identical)
Test:
Take two equal signals, phasedelete one with the other. => silence
Change one signal (In a way of your choice) => no absolute phasedeletion, no silence
Change the other signal in the SAME WAY you
changed the first. => silence.
I never got any other results.
This (for example) would prove my hypothesis to be wrong:
I would really really love, if someone could send me 2 waves which sound different and phasedelete each other). So only exactly identical signals can phasedelete each other so that your not able to hear anything after that.
I know we should discuss about audibility also. But this is a gradually thing for sure.
We could ask where from what point on are changes in numbers that one is able to hear?
(db, frq, ...).
We could ask at what point we begin to hear differences when adding signals in every single one we don't hear any difference...
...
Martin
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-27 05:50 ]</font>
General discussion about phase deletion
-
kensuguro
- Posts: 4434
- Joined: Sun Jul 08, 2001 4:00 pm
- Location: BPM 60 to somewhere around 150
- Contact:
I'm going to keep this short.. The conclusion on my end is that you can only phase cancel something that's exactly the same. It doesn't even have to do with the word "phase", but it's just a matter of addition. It's as easy as x + (-x) = 0. I guess taking temporal aspects into consideration, you can also have the exact same signal, but if they're out of phase, you'll get a beating partial cancellation.
But the other thing is this. I guess we can sort of appreciate the whole experimentation, but keep in mind that digitalaudiosoft is a business, and while I'm not very involved in the details of the past DAS issues, you are putting him at an extremely tough position.
I can see from your tests that there is some fishiness going on. Maybe we should leave it at that. Don't make it your goal to draw the definite line to discredit his products and to ruin his business. There is nothing that makes this sort of deliberate hostility a "right" thing to do.
<font size=-1>[ This Message was edited by: kensuguro on 2006-09-27 06:46 ]</font>
But the other thing is this. I guess we can sort of appreciate the whole experimentation, but keep in mind that digitalaudiosoft is a business, and while I'm not very involved in the details of the past DAS issues, you are putting him at an extremely tough position.
I can see from your tests that there is some fishiness going on. Maybe we should leave it at that. Don't make it your goal to draw the definite line to discredit his products and to ruin his business. There is nothing that makes this sort of deliberate hostility a "right" thing to do.
<font size=-1>[ This Message was edited by: kensuguro on 2006-09-27 06:46 ]</font>
Oh. I see. I really don't want that. That would be incredibly bad.
Now i see, that I started this discussion-topic to the wrong time. I will here say NOTHING negative to any specific company or product which is related to SCOPE.
I wanted to start this discussion without that aspect, really. I just want to know, because if someone states other things I think over what I said.
I started this thread to avoid anything that's related to specific companies! Because I'm interested in discussing digital audio!
This post made me think:
"I can see from your tests that there is some fishiness going on. Maybe we should leave it at that. Don't make it your goal to draw the definite line to discredit his products and to ruin his business. There is nothing that makes this sort of deliberate hostility a "right" thing to do."
I really never did want that. If you read the discussion you will see that from my point of view it is stating hypothesis, test, discussion, questioning. Nothing flaming at all. Sometimes reacting on things..., yes. Sorry for that.
The company you're talking about is a great audio company with great products. Which I recommend to buy (in other posts). I never said something bad about them.
If you read the discussion you will see, that ther can't be a final line for anything, because there are arguments. There is a game which I offered and which he accepted. I would have stopped if he had not agreed!
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-27 07:10 ]</font>
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-27 07:11 ]</font>
Now i see, that I started this discussion-topic to the wrong time. I will here say NOTHING negative to any specific company or product which is related to SCOPE.
I wanted to start this discussion without that aspect, really. I just want to know, because if someone states other things I think over what I said.
I started this thread to avoid anything that's related to specific companies! Because I'm interested in discussing digital audio!
This post made me think:
martinOn 2006-09-27 04:58, tgstgs wrote:
Ive missed this discussion course i was very buissy the last few days;
I dont want to take part of one side or offend or deffend or whatever;
I also dont want to put oil in fire,
but you know this is musicforum isnt it and people discuss aspects of music;
So i feel that i must post this one; Sorry in advance!!
At first to say iv made non of the tests people here did
so dont want to comment this;
to me this phase phenomenon ocurse in the analog world live on stage;
but its digital also;
Test ist made with a song i recorded few months ago; !!!NOT MASTERED!!!
contains full Drum (analog), BAS, Guitar, Organ, Vocal and Percussion
there are also synth involved but only for percusive additional effect;
So i have full spectrum with a analycer;
i took only the left channel duplicated it in Nuendo;
so i have 2 MONO files of same contend on different channels
one panned to the right one panned to the left;
the first one is routed over a Nuendo EQ to asio source to one channel strip of dynamic mixer;
The second one is routed direct to asio source from there to PEQ4 M to another channel strip of Mixer;
I iverted one channel on Dynamic MIxer; whatching the output of this sum;
the dynamic Mixer is 'not phase comp.'!!!
with Bypassed on both EQs i get a Signal to 0 ( first led of VU is blinking)
but on speaker i hear no sound;
So i made a setting on Nuendo EQ and the VU goes up as expected; i can hear now
Then i give the same Numbers in CWEQ to see the result;
after 5 minutes of tweaking the PEQ 4 M
the result goes to 0
check it out
'if i know how i would have posted a picture here but everyone can redo in a few minutes'
So what the F*ck is that!!
Steinacle uses CW Atoms !!
NO
the answer is very simple;
if you look to deep into the structure you finaly see MATH
(of Mr. Einstein the one and only and his Math friends)
does this mean every EQ is the same
NO
its the nuances the makes the difference and developers would be mad if they tell you;
Finally to say you could phaseOUT every wavesource equal what they are passing!!
its a trick engineers use live on stage;
just a good EQ of what sort ever and a good engineer and the signal is gone to 0 (nonhearable)
If some will redo this test
take wave source whatever you like;
take Nueando Sonar VST SX or whatever
route as me
and ply with the EQs finding the point where sum is to 0
the result depends on the source so different source have different settings!!
check it out! it takes just a few minutes
If you want to get better results course one led is still blinking
you had to phase comp. the dynamic mixer and put a delay behind PEQ 4 M
course there are differend runtimes of the signals
its like human ;
we all made of the same substances 99.9%
but this 0.1% is why we all dont look the same and some are good some bad;
its the nuances that makes the difference
i do not reply on this post i dont want to get involved in something i dont have do to with
maybe we discuss the aspects on another tread another time;
good vibes from vienna
"I can see from your tests that there is some fishiness going on. Maybe we should leave it at that. Don't make it your goal to draw the definite line to discredit his products and to ruin his business. There is nothing that makes this sort of deliberate hostility a "right" thing to do."
I really never did want that. If you read the discussion you will see that from my point of view it is stating hypothesis, test, discussion, questioning. Nothing flaming at all. Sometimes reacting on things..., yes. Sorry for that.
The company you're talking about is a great audio company with great products. Which I recommend to buy (in other posts). I never said something bad about them.
If you read the discussion you will see, that ther can't be a final line for anything, because there are arguments. There is a game which I offered and which he accepted. I would have stopped if he had not agreed!
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-27 07:10 ]</font>
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-27 07:11 ]</font>
In fairness, Martin did start these tests, but I'd imagine it wasn't specifically done to make DAS look bad or bad mouth them, as he seems deeply interested in this clinical approach to his testing & developement work & even previously stated that he'd held back from saying anything about any specific plugins.
During his conversations & arguments with DAS he has not once resorted to insulting them personally. He hasn't tried to cause trouble for them needlessly or continue these arguments with alterior motives AFAIK.
IMHO, DAS deserve as much respect as any other Scope users here & I agree that there's an element of distastefullness in talking about clinical tests of their plugins in public, but is that not what plugin developers & sound engineers do occasionally?.
Shroomz
(staying neutral like a good swiss man)
<font size=-1>[ This Message was edited by: Shroomz on 2006-09-27 07:45 ]</font>
During his conversations & arguments with DAS he has not once resorted to insulting them personally. He hasn't tried to cause trouble for them needlessly or continue these arguments with alterior motives AFAIK.
IMHO, DAS deserve as much respect as any other Scope users here & I agree that there's an element of distastefullness in talking about clinical tests of their plugins in public, but is that not what plugin developers & sound engineers do occasionally?.
Shroomz
(staying neutral like a good swiss man)
<font size=-1>[ This Message was edited by: Shroomz on 2006-09-27 07:45 ]</font>
Math 0 is a ideal you never reach in real world;
in real world its to 0 which is non hearable ; neglactable
same i dont say this
on my example iv 700HZ with q 0.7 and gain 4.5 in nuendo and 720HZ with q 1.2 and gain 4.5 type peak with PEQ result is -65.1DB first led of VU is blinking;
Thats less than pulsarIN with nothing connected (No cable)! = neglectable
but the same settings is a good starting point
wait a few months and we can discuss
final greetings and good vibes from vienna
in real world its to 0 which is non hearable ; neglactable
same i dont say this
on my example iv 700HZ with q 0.7 and gain 4.5 in nuendo and 720HZ with q 1.2 and gain 4.5 type peak with PEQ result is -65.1DB first led of VU is blinking;
Thats less than pulsarIN with nothing connected (No cable)! = neglectable
but the same settings is a good starting point
wait a few months and we can discuss
final greetings and good vibes from vienna
tgstgs i like your style of argumentation!
I think there is the point where it really becomes interesting.
I bought a behringer ADA8000 and I like it much but I see the argument, that when using it over and over with 100 Tracks I may hear something that was previously under that audible db value!
B.t.w. I could try to cancel A16 with ADA8000. that would be a great experiment cause I think they sound different!
For that reason I use A16 and differnent preamps too - to avoid this "doubling".
If talking about that single EQ you tried one could say: If you can phasedelete it to audible 0 you won't (very probably?)not able to tell which one is sounding better.
My thesis to this point is:
You won't be able to delete an EQ to that level (let's say 64, 80, 90, 100?), which sounds audible lets say 'better' than the other.
That does not mean, that you will hear a difference when adding 30 of these EQs.
You may be than able to decide: I like EQ a better than b because mix x (with a used) sounds much better than y (with b used).
Maybe a cancellation level of lets say -80db leads to audible changes when adding 3 of the
different plugins each! I don't know. That's what I am about to discuss.
Martin
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-27 07:39 ]</font>
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-27 07:41 ]</font>
I think there is the point where it really becomes interesting.
I bought a behringer ADA8000 and I like it much but I see the argument, that when using it over and over with 100 Tracks I may hear something that was previously under that audible db value!
B.t.w. I could try to cancel A16 with ADA8000. that would be a great experiment cause I think they sound different!
For that reason I use A16 and differnent preamps too - to avoid this "doubling".
If talking about that single EQ you tried one could say: If you can phasedelete it to audible 0 you won't (very probably?)not able to tell which one is sounding better.
My thesis to this point is:
You won't be able to delete an EQ to that level (let's say 64, 80, 90, 100?), which sounds audible lets say 'better' than the other.
That does not mean, that you will hear a difference when adding 30 of these EQs.
You may be than able to decide: I like EQ a better than b because mix x (with a used) sounds much better than y (with b used).
Maybe a cancellation level of lets say -80db leads to audible changes when adding 3 of the
different plugins each! I don't know. That's what I am about to discuss.
Martin
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-27 07:39 ]</font>
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-27 07:41 ]</font>
No Ken, he is putting CWA in that positionOn 2006-09-27 06:42, kensuguro wrote:
... but keep in mind that digitalaudiosoft is a business, and while I'm not very involved in the details of the past DAS issues, you are putting him at an extremely tough position...
if you load enough into and outof a project, route things here and tweak things there, suddenly some mixerchannels (at least there it's to be observed...) are delayed in the range of 1-4 samples.
Once a project 'crosses' that boarder, you can trash it - the state is irreversible and not even constant - your results will change with every launch.
This is of course only relevant if your sound depends on that type of sample accuracy - otherwise it would be an academic question.
imagine a phase extinction based De-esser (for example) that you've tweaked to an optimal setting on a vocal track, starts with a new value the next session...
Olive from DAS has more or less confirmed this, and as he's doing studio work for a living it's not uncommon to him with other gear - in fact his words were that CWA is among the best performers in this context.
cheers, Tom
<font size=-1>[ This Message was edited by: astroman on 2006-09-27 08:07 ]</font>
-
hubird
I think the word "phasedelete" is quite strange on its own...
if you mix 2 exactly same waveforms, one iverted and one not, they will mix up to complete silence. And only if they are exactly the same they will completly result in silence.
"in phase" or "not in phase" does more or less describe, if you play the 2 waveforms with a delay in between or not.
Then again, if "not played in phase" two exact same waveforms (one of them inverted) will not result in complete silence. they more or less add and subtract in amplitude.
this illustration will make it easier to understand.
http://www.sayedsaad.com/fundmental/5_S ... 0Phase.htm
so I think "phasechancel" is slightly the wrong word for this phaenomen.
<font size=-1>[ This Message was edited by: hifiboom on 2006-09-27 14:55 ]</font>
if you mix 2 exactly same waveforms, one iverted and one not, they will mix up to complete silence. And only if they are exactly the same they will completly result in silence.
"in phase" or "not in phase" does more or less describe, if you play the 2 waveforms with a delay in between or not.
Then again, if "not played in phase" two exact same waveforms (one of them inverted) will not result in complete silence. they more or less add and subtract in amplitude.
this illustration will make it easier to understand.
http://www.sayedsaad.com/fundmental/5_S ... 0Phase.htm
so I think "phasechancel" is slightly the wrong word for this phaenomen.
<font size=-1>[ This Message was edited by: hifiboom on 2006-09-27 14:55 ]</font>
-
hubird
yes, I'm glad you see the point now, Ken 
Not much in the world is as inexorable as phase cancelation, even at -91 db.
MCCyrano deserves a 'planetz developers price' for his scientific research and attitude.
The whole story also showed the selfcontrolled and selfmoderating capability of all planetz members, isn't it?
Imagine this whole thread on KVR or something, it would have been a real dirty internet war
Not much in the world is as inexorable as phase cancelation, even at -91 db.
MCCyrano deserves a 'planetz developers price' for his scientific research and attitude.
The whole story also showed the selfcontrolled and selfmoderating capability of all planetz members, isn't it?
Imagine this whole thread on KVR or something, it would have been a real dirty internet war
-
kensuguro
- Posts: 4434
- Joined: Sun Jul 08, 2001 4:00 pm
- Location: BPM 60 to somewhere around 150
- Contact:
took some time to make some of my own observations. No pictures, but to make my story short, I confirmed mccyrano, although I didn't spend too much time getting all the way to -90db.
left channel of -18db white noise played through ASIO in 32 bit.
inverted and added in modIII, using adern's Val Monitor to check for minute fluxuations.
phase invert checked all combination of:
original / polteq / p4eq
in bypass mode with successful complete cancellation
polteq / peq4 (peq4 inversed and added to polteq)
turning on polteq, with all bands off leaves a total signal of -162db, but amounts to digital silence (0bit) recorded.
Turning on all bands (at default position) leaves -40db (audible) signal.
Bottom 3 groups:
the 10k channel:
fiddled as much as possible, most cancellation was at -61dB
Polteq's "bandwidth" knob response is very different from peq4. In peq4's "Q" numbers, the "bandwidth" know stays below 9, 10 for most of the rotation, and then shoots into high resonance.
Polteq's frequency values seemed to not affect the bandwidth/atten values. So once I got the peq4 settings to match a certain set of values in polteq, then if I changed polteq's frequency, I just needed to match the frequency on peq4.
Combination:
With both 10K channel and 500 channel turned on and set to random settings, I was able to tweak down to -61dB.
Looking for non-linearity:
Polteq's Atten/Boost seems to change its response. If I set peq4 to match Polsteq at default Atten/Boost settings, the more Atten/Boost I apply, the more I need to adjust peq4's settings.
Top 2 groups:
Polteq's "Cut" seems to be a fixed frequency shelving filter of some sort, but it cuts a little more than the peq4's counterpart.
Polteq's upper 2 "Boost" is a fixed Q peaking. These I could match quite well with peq4's peaking, cancellation up to -69dB.
With real audio material:
When phase cancelling at -60dB or so, you really can't any hear any audio slipping through. So basically, in practical terms, it's a near perfect cancellation. Although I didn't do a spectral comparison, I would bet that the two will be darn close. The fact that I have to use "arrow tapping" to input values I can't see can leave a margin for error.
I would also guess that step amounts can also make a difference. Not sure if it applies here, but what I'm thinking of is: 0-20000 in 300 discrete steps versus 249 discrete steps. And since "arrow tapping" is discrete, then there's a chance that we're just landing on different values, causing the minute phase difference. (leaving us -60dB worth of "something") Given the polteq and peq4 use the same modules. But in that case, there's a low chance that I would have landed most everytime at a maximum cancellation of -61dB, so it's questionable.
The verdict:
As far as these tests show, there is very low chance that poleq truely has a unique sound, or atleast a sound that is too far from what peq4 can create. Even with a listening test, there was no marginal difference. True, the signal cancellation didn't result in a 100% cancellation, leaving -61dB of difference, but again, an inaudible difference seems to make no practical difference. The verdict is, basically these 2 sound similar enough to be considered the same in practical use.
The Pros:
In usage, polteq's design may prompt the user to focus more on the individual critical bands, speeding up sound tweaking. This may also have an effect on when you have a problem sound that you're stumped on, since the critical bands are laid out for you. (so you can do some trial and error) And finally, the polteq has a final out volume knob that's missing in peq4. Also, polteq has a total of 5 bands.
Defining Emulation:
I am not about to break this product as I have respect of the engineering and time that must have gone into this device. So perhaps we need to think about the definition of emulation? (local to the pultec, of course)
And idea that came to mind is that, perhaps it's not a circuit emulation with audio consequences, but maybe one that has do with usability and control? Like, perhaps the emulation focused on the response of the knobs, and tuning the bands to the same with the original, that sort of thing. If so, then I really don't know the original so I'll leave that judgement to someone else. I can still see though, that the knob response and band tuning does give this device some character in terms of usability. Especially the "bandwidth" knob, as it is very sensitive in the low to mid Q zone, which is very easy to fly past with peq4's tiny and highly responsive "Q" knobs. Things like that.
About Math behind EQ:
Well, I would agree that there are mathematical filter models that everyone uses. I know there are many famous ones, and I wouldn't be surprised if peq4 uses them (since it's stock eq), and if the atoms were there for the developers. I mean, there are C libraries with models for all sorts of filters. I wouldn't be surprised if the state variable filter in Max/MSP could be completely cancelled by its counterpart in csound. They probably both use the same "most recently made by some professor at MIT" model. (with lots of mysterious comments in the code and all)
But of course, when it comes to vintage modeling, it's sort of strange to be using these stock models.. but again, I think it's all a matter of which aspect was to be emulated, so..
About Math behind phase cancellation:
There is only one way a signal can be entirely cancelled, and it is by the inverted copy of itself. That is the only way, period. A waveform is an array of values, and the only way to bring that value to 0, is to subtract (or add the negative) the same value. This of course is only if you're mixing the signal in the digital domain. Once the audio is out in the air, well, it's a different story.
<font size=-1>[ This Message was edited by: kensuguro on 2006-09-27 21:48 ]</font>
left channel of -18db white noise played through ASIO in 32 bit.
inverted and added in modIII, using adern's Val Monitor to check for minute fluxuations.
phase invert checked all combination of:
original / polteq / p4eq
in bypass mode with successful complete cancellation
polteq / peq4 (peq4 inversed and added to polteq)
turning on polteq, with all bands off leaves a total signal of -162db, but amounts to digital silence (0bit) recorded.
Turning on all bands (at default position) leaves -40db (audible) signal.
Bottom 3 groups:
the 10k channel:
fiddled as much as possible, most cancellation was at -61dB
Polteq's "bandwidth" knob response is very different from peq4. In peq4's "Q" numbers, the "bandwidth" know stays below 9, 10 for most of the rotation, and then shoots into high resonance.
Polteq's frequency values seemed to not affect the bandwidth/atten values. So once I got the peq4 settings to match a certain set of values in polteq, then if I changed polteq's frequency, I just needed to match the frequency on peq4.
Combination:
With both 10K channel and 500 channel turned on and set to random settings, I was able to tweak down to -61dB.
Looking for non-linearity:
Polteq's Atten/Boost seems to change its response. If I set peq4 to match Polsteq at default Atten/Boost settings, the more Atten/Boost I apply, the more I need to adjust peq4's settings.
Top 2 groups:
Polteq's "Cut" seems to be a fixed frequency shelving filter of some sort, but it cuts a little more than the peq4's counterpart.
Polteq's upper 2 "Boost" is a fixed Q peaking. These I could match quite well with peq4's peaking, cancellation up to -69dB.
With real audio material:
When phase cancelling at -60dB or so, you really can't any hear any audio slipping through. So basically, in practical terms, it's a near perfect cancellation. Although I didn't do a spectral comparison, I would bet that the two will be darn close. The fact that I have to use "arrow tapping" to input values I can't see can leave a margin for error.
I would also guess that step amounts can also make a difference. Not sure if it applies here, but what I'm thinking of is: 0-20000 in 300 discrete steps versus 249 discrete steps. And since "arrow tapping" is discrete, then there's a chance that we're just landing on different values, causing the minute phase difference. (leaving us -60dB worth of "something") Given the polteq and peq4 use the same modules. But in that case, there's a low chance that I would have landed most everytime at a maximum cancellation of -61dB, so it's questionable.
The verdict:
As far as these tests show, there is very low chance that poleq truely has a unique sound, or atleast a sound that is too far from what peq4 can create. Even with a listening test, there was no marginal difference. True, the signal cancellation didn't result in a 100% cancellation, leaving -61dB of difference, but again, an inaudible difference seems to make no practical difference. The verdict is, basically these 2 sound similar enough to be considered the same in practical use.
The Pros:
In usage, polteq's design may prompt the user to focus more on the individual critical bands, speeding up sound tweaking. This may also have an effect on when you have a problem sound that you're stumped on, since the critical bands are laid out for you. (so you can do some trial and error) And finally, the polteq has a final out volume knob that's missing in peq4. Also, polteq has a total of 5 bands.
Defining Emulation:
I am not about to break this product as I have respect of the engineering and time that must have gone into this device. So perhaps we need to think about the definition of emulation? (local to the pultec, of course)
And idea that came to mind is that, perhaps it's not a circuit emulation with audio consequences, but maybe one that has do with usability and control? Like, perhaps the emulation focused on the response of the knobs, and tuning the bands to the same with the original, that sort of thing. If so, then I really don't know the original so I'll leave that judgement to someone else. I can still see though, that the knob response and band tuning does give this device some character in terms of usability. Especially the "bandwidth" knob, as it is very sensitive in the low to mid Q zone, which is very easy to fly past with peq4's tiny and highly responsive "Q" knobs. Things like that.
About Math behind EQ:
Well, I would agree that there are mathematical filter models that everyone uses. I know there are many famous ones, and I wouldn't be surprised if peq4 uses them (since it's stock eq), and if the atoms were there for the developers. I mean, there are C libraries with models for all sorts of filters. I wouldn't be surprised if the state variable filter in Max/MSP could be completely cancelled by its counterpart in csound. They probably both use the same "most recently made by some professor at MIT" model. (with lots of mysterious comments in the code and all)
But of course, when it comes to vintage modeling, it's sort of strange to be using these stock models.. but again, I think it's all a matter of which aspect was to be emulated, so..
About Math behind phase cancellation:
There is only one way a signal can be entirely cancelled, and it is by the inverted copy of itself. That is the only way, period. A waveform is an array of values, and the only way to bring that value to 0, is to subtract (or add the negative) the same value. This of course is only if you're mixing the signal in the digital domain. Once the audio is out in the air, well, it's a different story.
<font size=-1>[ This Message was edited by: kensuguro on 2006-09-27 21:48 ]</font>
'About Math behind phase cancellation:
There is only one way a signal can be entirely cancelled, and it is by the inverted copy of itself.'
true
there is math 0
NO ONE reaches it with the tests whatever tests be made;
you get -60 to -90dB
-60 in 5 minutes of tweaking -90 or so in hours
thats not audible
true
but not math 0
i showed with my nuendoEQ test that
YOU CAN DO THIS 'TEST' WITH EVERY EQ ON THIS PLANET
even with non digital analog HW!!!
with same result!!!!!!!!!!!!!
sorry to say but THIS TEST IS WORTHLESS
if some dont understand this (I see but i dont meen you kensuguro) they dont want to (for other reason)
take EQ whatever you like and you can bring the signal 'to 0' down under -60dB with another good EQ (the second one should be a good one) in a few minutes;
so this test doesnt meen anything sorry to say
WER MISST MISST MIST
i learned in the 1st lesson of electrical measurment technics 25 years ago
who measures measures rubbish
sorry in the translation the sense may be lost
those who want to see will see
greetings from vienna
There is only one way a signal can be entirely cancelled, and it is by the inverted copy of itself.'
true
there is math 0
NO ONE reaches it with the tests whatever tests be made;
you get -60 to -90dB
-60 in 5 minutes of tweaking -90 or so in hours
thats not audible
true
but not math 0
i showed with my nuendoEQ test that
YOU CAN DO THIS 'TEST' WITH EVERY EQ ON THIS PLANET
even with non digital analog HW!!!
with same result!!!!!!!!!!!!!
sorry to say but THIS TEST IS WORTHLESS
if some dont understand this (I see but i dont meen you kensuguro) they dont want to (for other reason)
take EQ whatever you like and you can bring the signal 'to 0' down under -60dB with another good EQ (the second one should be a good one) in a few minutes;
so this test doesnt meen anything sorry to say
WER MISST MISST MIST
i learned in the 1st lesson of electrical measurment technics 25 years ago
who measures measures rubbish
sorry in the translation the sense may be lost
those who want to see will see
greetings from vienna
-
kensuguro
- Posts: 4434
- Joined: Sun Jul 08, 2001 4:00 pm
- Location: BPM 60 to somewhere around 150
- Contact:
right, the test still does leave a margin for an obviously eccentric sounding EQ. (one that obviously doesn't use an established algo)
I guess bottom line, unless it's a 100% cancellation it's not numerically correct to say they're identical... But I still think that if the eq characteristics were different enough, it would produce different results.
But, comparing 2 eq that has a high probability of using mathematicall correct filter models is a bad comparison. Like nuendo EQ vs. samplitude EQ vs. peq4.. they're all sort of 'standard' eqs, so I wouldn't be surprised if they all responded very similarily.
Anyway, I need to find some obviously different ones to compare. Because now the focus should shift to "if you can hear the difference between 2 eqs, can they cancel each other out?"
I guess bottom line, unless it's a 100% cancellation it's not numerically correct to say they're identical... But I still think that if the eq characteristics were different enough, it would produce different results.
But, comparing 2 eq that has a high probability of using mathematicall correct filter models is a bad comparison. Like nuendo EQ vs. samplitude EQ vs. peq4.. they're all sort of 'standard' eqs, so I wouldn't be surprised if they all responded very similarily.
Anyway, I need to find some obviously different ones to compare. Because now the focus should shift to "if you can hear the difference between 2 eqs, can they cancel each other out?"
take the EQ of your HWmixer i guess you have one and try it out
i can do this with a soundcraft and a PEQ4 too!!
and i guess soundcraft dont uses CWAtoms
the interesting point for me
thats why i would continue this discuss is
when theres only a difference down under -60 to -90 dB
Why do they sound so different to my ears?
i can do this with a soundcraft and a PEQ4 too!!
and i guess soundcraft dont uses CWAtoms
the interesting point for me
thats why i would continue this discuss is
when theres only a difference down under -60 to -90 dB
Why do they sound so different to my ears?
I can't get even close to -90 db with analogue hardware! Could you post a setup and result-pictures? And some settings... How much EQs did you use at what gain?
But I also would draw -60db from our list. -60 is not enough.
-120 db is possible without problems for single bands in the tests with the well discussed digital ones. Logically when being able to really get the right values from the connected knobs there is (& must be) more (than -90db) possible with identical EQs. When being able to reproduce right values what depends on the value matrix of the knobs.
Martin
Demo wavs would be unbelievable GREAT. It's not a big thing, just recording 10 seconds of your test.
This would become a fantastic recource on audio-knowhow!!!
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-28 06:25 ]</font>
But I also would draw -60db from our list. -60 is not enough.
-120 db is possible without problems for single bands in the tests with the well discussed digital ones. Logically when being able to really get the right values from the connected knobs there is (& must be) more (than -90db) possible with identical EQs. When being able to reproduce right values what depends on the value matrix of the knobs.
Martin
Demo wavs would be unbelievable GREAT. It's not a big thing, just recording 10 seconds of your test.
This would become a fantastic recource on audio-knowhow!!!
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-28 06:25 ]</font>
well, that's an easy question to answerOn 2006-09-27 23:15, tgstgs wrote:
...when theres only a difference down under -60 to -90 dB
Why do they sound so different to my ears?
none of the EQs discussed in this context is a phase-linear type, so the filter process does alter the phase (by nature AND by intention - it's what makes the EQ 'sound')
the human ear is far more sensitive to phase deviations between 2 sources than to a their difference in loudness.
The latter is the large scale information, giving you a rough impression of the distance, while the signals phase difference between left and right ear will tell you about the location and movement.
it's amazing how much detail can be in the 'background noise' around you, once there's a reason to focus on it.
For example I'd never do bicycle rides with earphones on, as they hide some vital information (in the true sense of the word) about what's going on behind you - or think about the tones to guide blind people at traffic lights.
cheers, Tom
Additional thesis concerning analogue hardware:
Much depends on the setup. If you record your analogue EQ from the mixer and the singnal to be edited with digital Para-EQ for complete deletion through substraction through the same mixer comparability to SCOPE EQ is influeced:
Can you bypass that EQ on the mixer? If not, some sound-influencing things (which make you feel the EQ sounds different) are recorded for that test in both signals.
ParaEQ is a very neutral EQ which has not much sound on it's own. So it is not the same editing a signal (lets take for example white noise) in Computer or white noise which has allready gone through that sound-making mixer.
I say: If both parts for deletion go through that same mixer, soundwise both signals are allready influenced by the EQ sound of the mixer (when it is not able to be bypassed).
I ask: Can we (if you can 'phasedelte' these signals to lets say -80, -90, -100db) hear differences in these signals in a blind test?
Simple thing to find out:
Record them and we'll design a blind test.
Martin
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-28 06:26 ]</font>
Much depends on the setup. If you record your analogue EQ from the mixer and the singnal to be edited with digital Para-EQ for complete deletion through substraction through the same mixer comparability to SCOPE EQ is influeced:
Can you bypass that EQ on the mixer? If not, some sound-influencing things (which make you feel the EQ sounds different) are recorded for that test in both signals.
ParaEQ is a very neutral EQ which has not much sound on it's own. So it is not the same editing a signal (lets take for example white noise) in Computer or white noise which has allready gone through that sound-making mixer.
I say: If both parts for deletion go through that same mixer, soundwise both signals are allready influenced by the EQ sound of the mixer (when it is not able to be bypassed).
I ask: Can we (if you can 'phasedelte' these signals to lets say -80, -90, -100db) hear differences in these signals in a blind test?
Simple thing to find out:
Record them and we'll design a blind test.
Martin
<font size=-1>[ This Message was edited by: MCCYRANO on 2006-09-28 06:26 ]</font>