Who do we trust?

[andreasmaaan]

Sorry I was unclear. A single speaker situated in front of you.

Ok I see. Let's assume for argument's sake that they don't do this comparison. They are nevertheless equalising the components of the mix (and the mix as a whole) so that the sound coming from the L and R speakers sounds the way they want it to sound to a pair of ears positioned equidistantly between the speakers.

I have two points to make.

Firstly, let's say they get it "wrong", i.e. they overcompensate, undercompensate, or incorrectly compensate. Where does that leave someone at home applying a global filter to the master based on an assumption that the engineers mixed the music using only a mono centre channel (which if I'm not mistaken, was your suggestion)?

Secondly, the instruments or voices being recorded are not capturing the general, abstract essence of the voices/instruments that they are recording, such that there is a single "correct" tonal balance that the mixing engineer can apply to accurately recreate that voice/instrument.

Instead, the engineer is working with a recording taken by a microphone that captures only how a voice/instrument sounded at a particular point in space in the room in which it was recorded, as well as the multiple reflections that occurred in that space and which arrived at the mic from a multitude of directions, but which the reproduction system now emits from only X directions (where X = the number of speakers in the reproduction system, i.e. in the case of stereo, two).

Considering all this, it quickly becomes clear that there is no universal "correct" tonal balance for any particular recorded sound. The engineer's job is to judge for themselves what the optimal tonal balance is for a recorded sound using the system with which they are engineering the mix.

In this light, the optimal system for engineering the mix is the same (type of) system upon which the mix will ultimately be reproduced by the end-listener. The tonal balance that the engineer judges to be "correct" (or at least, desired) will best be reproduced by a system conforming to the same general properties as that with which the engineer made the judgement. If both the mixing setup and the listening setup are the same (or of the same type, or as similar as possible), the listener can rest assured that they are hearing the most "correct" tonal balance, as judged by the engineer.

 

[Thomas_A]

If they would mix on a mono centre which has a linear response and then switch to stereo speakers with the same linear response, they may hear the timbral change and compensate for that (but only partly, since full compensation would only apply for the phantom center and not the panned left/right sounds). What I am saying is that I don't think mixing is done in that way normally, but I agree that the engineers mix to their preference and they may be lucky and hit the correct compensation.

If you want to listen to the exact tonal balance as the engineered finally heard then you should go for the same/similar setup. But if you don't believe that the compensation is done in that way during recording you can apply that compensation in your own listening environment.

 

[andreasmaaan]

If they would mix on a mono centre which has a linear response and then switch to stereo speakers with the same linear response, they may hear the timbral change and compensate for that (but only partly, since full compensation would only apply for the phantom center and not the panned left/right sounds). What I am saying is that I don't think mixing is done in that way normally, but I agree that the engineers mix to their preference and they may be lucky and hit the correct compensation.

If you want to listen to the exact tonal balance as the engineered finally heard then you should go for the same/similar setup. But if you don't believe that the compensation is done in that way during recording you can apply that compensation in your own listening environment.

I understand your point But what I'm saying is that I think you're looking at it the wrong way if you see it in terms of "compensation". The engineers mix on a system with (essentially) the same stereo crosstalk as the system you are listening on. There is no correct or incorrect, compensated or uncompensated (unless of course the engineers mixed the recording on a system that exhibits stereo crosstalk, and your system does not, e.g. a stereo mix listened to through a mono bluetooth speaker; in such a case, compensation may be justified).

In other words, it's not a case of the engineers getting "lucky" or not. They are mixing a recording on a particular setup to be listened to on the same type of setup. If you have such a setup, you will hear what the engineers intended. If the timbre of the phantom centre is different to the timbre that the engineers would have created had they referenced their stereo phantom centre to a true mono centre channel, you will have no way of knowing that, and any "compensation" you apply will be completely arbitrary, i.e. the only thing you will know for certain is that you are shifting the tonal balance away from that which the engineers intended.

If you're going to shift the tonal balance away from that which the engineers intended, you may as well do it creatively, i.e. to your own taste, rather than arbitrarily.

 

[LeftCoastTim]

I'm very late to this thread. I think the audiophile community has two poles: the art-appreciation vs engineering of audio equipment. I clearly fall in the engineering side.

Therefore, I trust:
https://www.amazon.com/Sound-Reproduction-Psychoacoustics-Loudspeakers-Engineering/dp/0240520092

and
https://www.listeninc.com/wp/media/Perception_and_-Measurement_of_Headphones_Sean_Olive.pdf

In the above PDF, page 115 entitled "Retail Price Versus Sound Quality" is an accurate representation of the state of hifi.

 

[andreasmaaan]

Another quick question @Thomas_A: Why would a mixing engineer who mixed their stereo phantom centre by:

• firstly, mixing on a true mono centre channel
• secondly, referencing their stereo mix to this true mono centre's tonal balance

arrive at a different tonal balance than if they'd skipped the true mono reference and mixed only in stereo?

Wouldn't they arrive at the same tonal balance (i.e. the one they believed to sound most correct/preferred), regardless of the path they used to get there?

 

[DDF]

Another quick question @Thomas_A: Why would a mixing engineer who mixed their stereo phantom centre by:

• firstly, mixing on a true mono centre channel
• secondly, referencing their stereo mix to this true mono centre's tonal balance

arrive at a different tonal balance than if they'd skipped the true mono reference and mixed only in stereo?

Wouldn't they arrive at the same tonal balance (i.e. the one they believed to sound most correct/preferred), regardless of the path they used to get there?

I'm not Thomas but if I understand your question correctly, I answered it here
https://www.diyaudio.com/forums/mul...t-vs-accurate-hi-fidelity-15.html#post4482884
and here
https://www.diyaudio.com/forums/multi-way/103978-flat-inaccurate.html

 

[Thomas_A]

I understand your point But what I'm saying is that I think you're looking at it the wrong way if you see it in terms of "compensation". The engineers mix on a system with (essentially) the same stereo crosstalk as the system you are listening on. There is no correct or incorrect, compensated or uncompensated (unless of course the engineers mixed the recording on a system that exhibits stereo crosstalk, and your system does not, e.g. a stereo mix listened to through a mono bluetooth speaker; in such a case, compensation may be justified).

In other words, it's not a case of the engineers getting "lucky" or not. They are mixing a recording on a particular setup to be listened to on the same type of setup. If you have such a setup, you will hear what the engineers intended. If the timbre of the phantom centre is different to the timbre that the engineers would have created had they referenced their stereo phantom centre to a true mono centre channel, you will have no way of knowing that, and any "compensation" you apply will be completely arbitrary, i.e. the only thing you will know for certain is that you are shifting the tonal balance away from that which the engineers intended.

If you're going to shift the tonal balance away from that which the engineers intended, you may as well do it creatively, i.e. to your own taste, rather than arbitrarily.

I perfectly understand what you are saying as well. But I don't believe that the engineers use a standardised way to compensate and that, overall, there is an error in most recorded work compared to the live sound source. You are right, I am suggesting a small compensation (about +/-1 dB) that changes the timbre away from the engineers intended sound, just because of the above.

 

[Thomas_A]

I'm very late to this thread. I think the audiophile community has two poles: the art-appreciation vs engineering of audio equipment. I clearly fall in the engineering side.

Therefore, I trust:
https://www.amazon.com/Sound-Reproduction-Psychoacoustics-Loudspeakers-Engineering/dp/0240520092

and
https://www.listeninc.com/wp/media/Perception_and_-Measurement_of_Headphones_Sean_Olive.pdf

In the above PDF, page 115 entitled "Retail Price Versus Sound Quality" is an accurate representation of the state of hifi.

Toole is basically correct but as far as I know he never investigates the small corrections (about +/-1 dB) that are required at certain frequencies in terms of what listeners prefer.

 

[Thomas_A]

I'm not Thomas but if I understand your question correctly, I answered it here
https://www.diyaudio.com/forums/mul...t-vs-accurate-hi-fidelity-15.html#post4482884
and here
https://www.diyaudio.com/forums/multi-way/103978-flat-inaccurate.html

You answered it and curves are also presented in Toole's work (citation of Shirley).

 

[RayDunzl]

Toole is basically correct but as far as I know he never investigates the small corrections (about +/-1 dB) that are required at certain frequencies in terms of what listeners prefer.

Maybe he - @Floyd Toole - will comment...

 

[Thomas_A]

Maybe he - @Floyd Toole - will comment...

That would be nice!

 

[andreasmaaan]

But I don't believe that the engineers use a standardised way to compensate and that, overall, there is an error in most recorded work compared to the live sound source.

The thing is, the live sound source doesn’t have a singular or “correct” tonal balance.

Any live source has a theoretically infinite number of possible tonal balances: one for every point in space at which the recording microphone could have been placed.

None of these is the singularly “true” or “correct” or inherent tonal balance of the instrument. You might think about this in similar terms to photography: each photo represents just one of an infinite number of possible perspectives. No single perspective captures the whole truth.

Similarly, in audio, at any given mic location, the tonal balance of the recorded sound will be slightly different than at any other possible location (as will the relationship between direct and reflected sounds).

It’s therefore nonsensical to compare a recorded sound to a live sound source and say the former contains an “error”; rather, any recorded sound is nothing more than a single, incomplete perspective.

The job of an engineer is to work with recorded sounds on the setup they have to create a representation of the live sound that artistically or aesthetically conforms to the “truth” of the live sound as it appears to them.

If you listen back on the same setup, you will hear the same version of this truth that the engineer created.

 

[Thomas_A]

The thing is, the live sound source doesn’t have a singular or “correct” tonal balance.

Any live source has a theoretically infinite number of possible tonal balances: one for every point in space at which the recording microphone could have been placed.

None of these is the singularly “true” or “correct” or inherent tonal balance of the instrument. You might think about this in similar terms to photography: each photo represents just one of an infinite number of possible perspectives. No single perspective captures the whole truth.

Similarly, in audio, at any given mic location, the tonal balance of the recorded sound will be slightly different than at any other possible location (as will the relationship between direct and reflected sounds).

It’s therefore nonsensical to compare a recorded sound to a live sound source and say the former contains an “error”; rather, any recorded sound is nothing more than a single, incomplete perspective.

The job of an engineer is to work with recorded sounds on the setup they have to create a representation of the live sound that artistically or aesthetically conforms to the “truth” of the live sound as it appears to them.

If you listen back on the same setup, you will hear the same version of this truth that the engineer created.

You could say that everything is in the mix already, or not. I think the experiment for preference has not been done as it would implicate testing changes within +/-1 dB and only at certain frequencies. It would be nice if Toole could comment if such experiments have been conducted. I have not seen them though.

 

[Blumlein 88]

You could say that everything is in the mix already, or not. I think the experiment for preference has not been done as it would implicate testing changes within +/-1 dB and only at certain frequencies. It would be nice if Toole could comment if such experiments have been conducted. I have not seen them though.

So is this about altering response from using left and right speakers that you have in mind?

Toole has done most testing with a single mono speaker. He did do testing with stereo speakers, and found the same type of response was preferred by people, but the results were more consistent and discriminating with one speaker vs two.

 

[Thomas_A]

So is this about altering response from using left and right speakers that you have in mind?

Toole has done most testing with a single mono speaker. He did do testing with stereo speakers, and found the same type of response was preferred by people, but the results were more consistent and discriminating with one speaker vs two.

It would imply a compensation for the timbral change that two stereo speakers give compared to a centre mono speaker and whether the compensated stereo speaker would be preferred over the non-compensated. Has that been done?

 

[andreasmaaan]

You could say that everything is in the mix already, or not.

Could you explain this please?

 

[Thomas_A]

Could you explain this please?

That the engineers go for a desired sound that includes the above said compensation.

 

[Blumlein 88]

It would imply a compensation for the timbral change that two stereo speakers give compared to a centre mono speaker and whether the compensated stereo speaker would be preferred over the non-compensated. Has that been done?

To my knowledge no. If Mr. Toole replies, he can give us more info on that.

I do think compensating for spaced stereo speakers is still not going to be right because the compensation for center phantom images will mean the response is off for left and right sources.

 

[Blumlein 88]

That the engineers go for a desired sound that includes the above said compensation.

I think andreasmaaan will agree that engineers are working by ear. They aren't sitting down to compensate via some formula. If they intuitively alter EQ for a preferred sound then it might happen to be compensating for any difference in spaced stereo vs actual center mono speakers.

 

[Thomas_A]

To my knowledge no. If Mr. Toole replies, he can give us more info on that.

I do think compensating for spaced stereo speakers is still not going to be right because the compensation for center phantom images will mean the response is off for left and right sources.

That is correct, and that is why you need to make a compromise and not make the full compensation. Much of the information is within the centre phantom image though. It is compensation within +/- 1 dB so I think that means the experiment is rather hard to do (many trials etc).