[diablo9]

first let's forget about the coloration from source, IC, AMP, HD, let's say the pure "duplicate" of the original sound is considered as "perfect" SQ. there is one way to perfectly measure SQ:
1. take a dummy head, install Hi-End microphone in it, like the ones recording studio used to record concert.
2. insert your IEM, put on your full-size HD or whatever HD on this dummy head.
3. play the source, use mic in the dummy head to record, convert the recorded sound into wave form
4. rip the origin CD into wave form
5. most important step, figuring out alogirithm to compare the two wave forms and find percentage of difference.
The lower the percentage is, the better the signal chain is.
Isn't that right?

 

[ken36]

Way over my head.

 

[braillediver]

How about the visceral feel of the music? The feel of the music penetrating your body, muscles, bones, soul.

At a Pink Floyd Concert once I was thinking that my brain cells were actually feeling the music they were remembering.

I sat with a musician friend a few weeks ago and listened while he did sound checks on the gear and set up microphones. I doubt a musical reproduction system could ever reproduce the intensity and feeling that the live instruments I was hearing produced. It was a whole body experience.


Mitch

 

[rsaavedra]

Quote:

Originally Posted by diablo9 first let's forget about the coloration from source, IC, AMP, HD, let's say the pure "duplicate" of the original sound is considered as "perfect" SQ. there is one way to perfectly measure SQ: 1. take a dummy head, install Hi-End microphone in it, like the ones recording studio used to record concert. 2. insert your IEM, put on your full-size HD or whatever HD on this dummy head. 3. play the source, use mic in the dummy head to record, convert the recorded sound into wave form 4. rip the origin CD into wave form 5. most important step, figuring out alogirithm to compare the two wave forms and find percentage of difference. The lower the percentage is, the better the signal chain is. Isn't that right?

Yes in theory, in practice there are several issues to deal with:
- Frequency response of the microphone. (Btw I believe the microphone you would need inside a dummy head is different from the one used in recording studios. Would need to be the kinds of microphones used for binaural recordings.)
- "Convert recorded sound into wave form" --> choice of analog-to-digital converter.
- Time shifts between encoding coming directly from CD, vs coming from that converter.
- Resonances in the chamber of air between IEM and dummy ear canal and microphone will be different from resonance between IEM and real ear canal of someone, and a real tympanic membrane.

Even if the IEM had perfectly flat freq. response all of the above will be troublesome in practice.

But even in theory I think you can't ultimately measure sound reproduction accurately because the fact that you are measuring is an intrusion in the dynamics of the system that alters what you are trying to measure. (This evokes the uncertainty principle in physics-chemistry.) That doesn't mean you can't measure and approximate better and better though.

PS. What are you using HD for? Headphones?

 

[diablo9]

Quote:

Originally Posted by braillediver How about the visceral feel of the music? The feel of the music penetrating your body, muscles, bones, soul. At a Pink Floyd Concert once I was thinking that my brain cells were actually feeling the music they were remembering. I sat with a musician friend a few weeks ago and listened while he did sound checks on the gear and set up microphones. I doubt a musical reproduction system could ever reproduce the intensity and feeling that the live instruments I was hearing produced. It was a whole body experience. Mitch

From physics, sound only involves two factors, frequency and SPL in db. So I believe that you "feel" the music in your muscle, bone, soul is only because SPL is big enough that it vibrate your viscera. I know objective test of SQ involve alot of things in psychology, but to measure SQ in "scientific" way, I think that's the way to go

 

[diablo9]

Yes in theory, in practice there are several issues to deal with:
- Frequency response of the microphone. I believe the microphone you would need inside a dummy head is different from the one used in recording studios. Would need to be the kinds of microphones used for binaural recordings.
I agree with you on that, either the dummy or mic needs some mod, though, I just need the "most accurate" ones.

- "Convert recorded sound into wave form" --> choice of analog-to-digital converter.
actually recording industry already covered that. We are NOT afraid of ADC used in recording studio and just trust whatever ADC they used and assume that one perfectly replicate the analog recording to digital, we should trust ADC used in this experiment, too. just use exactly the same ADC they used when producing CD.

- Time shifts between encoding coming directly from CD, and wave coming from that converter.
this should be covered by the algorithm. I am not expert on this but I am sure mathematician can figure it out.

- Resonances in the chamber of air between IEM and dummy ear canal and microphone will be different from resonance between IEM and real ear canal of someone, and a real tympanic membrane.
This shouldn't be worried. we dont' worry about resonance when we wear HD or IEM, right? mic here just funcation as "machine ear" to fully memorize what it hear and never forget.
Even if the IEM had perfectly flat freq. response all of the above will be troublesome in practice.

But even in theory I think you can't ultimately measure sound reproduction accurately because the fact that you are measuring is an intrusion in the dynamics of the system that alters what you are trying to measure. (This evokes the uncertainty principle in physics-chemistry.) That doesn't mean you can't measure and approximate better and better though。
I know you are talking about Heisenberg theory, right? I don't think that can be applied here.

 

[diablo9]

oops, HD should be HP, headphone, my bad.

 

[rsaavedra]

Quote:

Originally Posted by diablo9 I know you are talking about Heisenberg theory, right?

Yep, Heisenbergs Uncertainty Principle. I think it does apply to some extent.

 

[rsaavedra]

Quote:

Originally Posted by diablo9 This shouldn't be worried. we dont' worry about resonance when we wear HD or IEM, right?

We largely do worry about those resonances. Have you heard of HRTF (Head-related Transfer Functions), more recently called ATF (Anatomical Transfer Functions)? Also, IEM manufacturers take into account typical resonances in the ear canal to shape the desired response of the IEM's for more reallistic sound reproduction.

 

[diablo9]

Quote:

Originally Posted by rsaavedra Yep, Heisenbergs Uncertainty Principle. I think it does apply to some extent.

emm, dummy is of course different from real head.

mic is of course different from ear drum,
but my point is: if you go to a concert, you listen to a concert right at the "optimal position", OR, if there is accurate mic in a dummy head at exactly same position. what you hear should be exactly the same with the mic, given the design of "dummy-mic" combo is optimal.
so my definition of perfect SQ is that, your source-IC-AMP-HP signal can perfectly replicate what the dummy heard that give it to you.
equation is:
IF in the concert, what dummy heard (m)= what you heard (n)
Since we mostly trust recording studio, so we suppose everything that's recorded are put on CD, in digital form. so CD content (x)=m
IF signal chain is perfect, your rig should perfectly replicate the digital content on CD into analog, so x= your rig's playback sound (y)
so y=n, IF your rig is PERFECT.
the further away y is from n, the worse your signal chain is.

 

[diablo9]

Quote:

Originally Posted by rsaavedra We largely do worry about those resonances. Have you heard of HRTF (Head-related Transfer Functions), more recently called ATF (Anatomical Transfer Functions)? Also, IEM manufacturers take into account typical resonances in the ear canal to shape the desired response of the IEM's for more reallistic sound reproduction.

I really have not heard of that HRTF before, yes, in IEM and closed structure HP, resonances are causing problems. I know designers are having hard time figuring out ways to get rid of resonance.
So what if I take out IEM and closed structure HP in my original thread and only talk about open HP here?

 

[Ferbose]

Point #5 is wrong, or at least I disagree.
This algorithm does not exist today and likely never will.
The word "sound" implies there is an ear-brain to interpret the pressure variations and movement in air.
Sound quality is determined by ear-brain but we can't measure inside the ear-brain with any accuary at all.
You can devise all the algorithm you want to quantify the measurable differences in your point #5, be it THD or IMD or whatever distortion index you design, but that quantified index will have little correlation with perceived sound quality in ear-brian, unless you know how the ear-brain actually works.

 

[diablo9]

in this case, we don't need to know how the ear/brain works, we just need to know the frequency/SPL/timing of each wave form and see if your rig is accurately duplicating it. in other words,
concert--------------->your ear/brain
concert->CD->your rig->your ear/brain
In these two cases,
I ONLY care if your ear/brain is receiving the identical/similar information, but I don't care how your ear/brain is interpreting it. that's why I don't need to know how they work.
How your ear/brain work surely matters, but they just don't matter in this particular comparison case.

 

[Emon]

Quote:

Originally Posted by Ferbose Point #5 is wrong, or at least I disagree. This algorithm does not exist today and likely never will.

What?? Are you saying we can't tell the difference between two waveforms? You crazy!

 

[Ferbose]

Of course there is a difference, but how do you relate the difference to sound quality. There is no perfect transducer or microphone. In point #5 there will be a measurable difference for sure. Different->better? or differet->worse? or different inaudible? How do you quantify, without understanding the ear-brain?