Difference between microphones and ears

[gerhardroux]

Dear All,

To what can the differences between the sound "heard" by microphones and the human ear be attributed? For instance, I always hear more mids in a recorded signal than listening to the original signal at the position of the microphone.

For arguments' sake let us accept that the recording chain is neutral and is not introducing any resonances and delayed signals or any other forms of coloration. (Although technically not possible, I believe modern equipment can be neutral enough to be negligible in this instance)

Some guesses from my side:
a) The ears difference in sensitivity at different levels. (Flecher-Munson) and the inability to accurately match the playback level to the original level.
b) Head related transfer function. The sound source is limited to a stereo (or surround) source that provides different auditory clues than the original sound.
c) The influence of spatial clues on the brain's ability to interpret frequency information can be more than expected.

Is there some transfer function that I'm just not aware of?

Regards,
Gerhard Roux

[psykostx]

Originally Posted by gerhardroux

Dear All,

To what can the differences between the sound "heard" by microphones and the human ear be attributed? For instance, I always hear more mids in a recorded signal than listening to the original signal at the position of the microphone.

For arguments' sake let us accept that the recording chain is neutral and is not introducing any resonances and delayed signals or any other forms of coloration. (Although technically not possible, I believe modern equipment can be neutral enough to be negligible in this instance)

Some guesses from my side:
a) The ears difference in sensitivity at different levels. (Flecher-Munson) and the inability to accurately match the playback level to the original level.
b) Head related transfer function. The sound source is limited to a stereo (or surround) source that provides different auditory clues than the original sound.
c) The influence of spatial clues on the brain's ability to interpret frequency information can be more than expected.

Is there some transfer function that I'm just not aware of?

Regards,
Gerhard Roux

I think you know way more than you think you do!

a) turn the fletcher-munson eq curve upside down and thats about what most mics pick up and speakers output (without crossovers/mid-cuts), not to mention also that your brain does filter a lot of mids out of real sounds as it accounts for room accoustics automatically and emphasizes sounds that may be important to you as a living organism trying to keep yourself that way ( the little mixing engineer inside your head )

b) Only thing I can say here is why you shouldn't mix bass in headphones. You can feel well past your hearing range in both directions. Some headphones you can actually "hear" ultrasound emission, which is supposed to be filtered out. Throw those away. Nothing is more fatiguing to your ears, and you can't even hear it. Your ears will hurt and ring even at low volumes after very short periods of time. A lot of people will argue this post, but you can also hear the oscillation of a tv scan tube if you listen. It sounds very close to minor tinnitus from a show except it goes away when you shut the tv off or throw the headphones out...

c) All I can say is. You betcha! Take a tape recorder into your bathroom and sing away..."I don't remember being almost comepletely unintelligable"...
This is because reflections are really not much quieter than the source sound in a small space, but your brain recognizes their "image" almost instantly and drastically reduces your perception of them so you can hear those predators coming...

[psykostx]

PS: If you are wondering why you can't A/B a recording and real life until they match, Its because of dynamic range. You actually can do this, its just the amount of energy it takes to run a playback system with that much dynamic range, loudness, AND detail isn't really economical. I mean, in order to reproduce drums through a speaker system in absolutely full detail at full loudness with abosolutely inaudible noise floor, you would need to have a signal that could peak at the level of a jet engine, yet produces no noise. Not cheap to say the least. There are a few systems out there though that can!

You will never own one though, don't worry. You wouldn't even be able to get that much electricity routed to your house. In fact, if you played some dance music through it you probably wouldn't have a house left.

[gerhardroux]

Ok, even if you can't match the exact SPL the 96dB dynamic range of a 16 bit recording should accurately render the dynamic content of the signal. And with something like a piano, it is not that difficult to produce the same SPL in the control room than at the position of the microphone, which takes me back to my original question.

Some smart person must have given this transfer function (frequency difference between source and playback) a collective name, or is it seen as part of the frequency response characteristics of a reproduction system?

[Bob Olhsson]

Unless the recording or playback room is completely non-reverberant, you are always listening to two rooms. The microphone needs to be tiny or you're getting diaphragm diffraction effects. Loudspeaker diffraction is another common source of distortion. In fact that's where the extra upper mids come from in many speakers that measure flat.

[Ethan Winer]

Originally Posted by gerhardroux

I always hear more mids in a recorded signal than listening to the original signal at the position of the microphone.

As Bob said, all rooms have a very strong affect on the response at any given location. If you move a microphone even a few inches the response is very different. The root problem is comb filtering caused by reflections. This creates a complex series of peaks and nulls in the response. The response is whatever it is at the microphone, but when you listen you have two ears. So nulls heard by one ear are not nulls in the other ear and vice versa.

As proof, look at the two responses below. These were measured in a bedroom size space with no acoustic treatment. The only difference between these two measurements is the microphone was moved four inches to one side. An adult's ears are more like six inches apart.

So there's your answer. A microphone listens at only one location, so the sound quality is highly colored. But your years are in two locations, and the sum/average of both locations is closer to a "normal" response.

--Ethan

Source:

[psykostx]

Originally Posted by gerhardroux

Ok, even if you can't match the exact SPL the 96dB dynamic range of a 16 bit recording should accurately render the dynamic content of the signal. And with something like a piano, it is not that difficult to produce the same SPL in the control room than at the position of the microphone, which takes me back to my original question.

Some smart person must have given this transfer function (frequency difference between source and playback) a collective name, or is it seen as part of the frequency response characteristics of a reproduction system?

O I get it, so they build those huge walls of speakers in laboratories for nothing! Can a 16-bit file "waveform" ("staircaseform") that includes a pin drop, a snare drum, and a gunshot, all in full detail without any compression and with negligible noise, be played back at the original spl without distortion on studio monitors? Then why would you think that the ASTOUNDING dynamic intricacies of a piano could be fully represented and played back in 16-bits at original spl full detail without any distortion or noise through studio monitors?

Not trying to be mean. If you had anything to criticize about my post, it should have been my answer to B, in which I missed the point of your post entirely. I was tired . And yes, a broken set of headphones, and/or a broken headphone amp is very very bad for you! I may have been off topic but I am not making this sh&t up!

Mr. Winer, I'm not surprised that you downplay psycho-acoustics when it comes to room acoustics. So every room has comb filtering that is different at every point in that room, I understand with your hearing two rooms point but why does it still sound different even on headphones...theres much less combing with open ear cans and no phase cancellation between L/R channels going in that case? You have to give some credit to psychoacoustic phenomenon even though Gehrard should buy some RealTraps anyways!

[Ethan Winer]

Originally Posted by psykostx

I'm not surprised that you downplay psycho-acoustics

Not at all. There are many factors. Bob mentioned microphone limitations due to diaphragm size, and there are a dozen other reasons why it's difficult for a microphone to capture exactly what we hear. But I'm convinced that the main reason the OP notices more mids is as I stated.

Gehrard should buy some RealTraps anyways!

Indeed.

Though that's not why I posted to this thread.

--Ethan

[psykostx]

Originally Posted by Ethan Winer

Indeed.

Though that's not why I posted to this thread.

--Ethan

Ahah I put that in there so you wouldn't think I was accusing you of salesmanship... What I meant is that you are absolutely obcessed with room acoustics and it is contagious! In a good way!

[gerhardroux]

Thank you very much Bob, Ethan, and SafeandSound. Your answers helped me a lot.

[gerhardroux]

Originally Posted by psykostx

If you had anything to criticize about my post, it should have been my answer to B, in which I missed the point of your post entirely.

I had no criticism of your post. I asked another question. I appreciate your responses.

Regards,
Gerhard Roux

[gerhardroux]

I would guess that the following source is a good starting point explaining this phenomena:

Olive, S.E. (1986). The preservation of timbre: Microphones, loudspeakers,
sound sources and acoustical spaces. In: AES 8th International Conference,
pp. 127–149. Washington.

Any more ideas?

[Gutbucket]

This strikes me as one of the most profound questions concerning recording that one can ask.

I've come to believe more and more strongly over time that the biggest difference between ears and microphones is the presence of a brain attached directly and inseparably from the first.

A microphone is a relatively simple transducer. Human hearing is an entire system which encompasses all of the aspects mentioned previously in this thread and many, many others. The incredibly complex, non-linear processing that goes on in the process of human hearing before we even become conscious of a sound is staggering.

There are also some books and papers listed on Stan Linkwitz's site that directly address this question. There are also some interesting papers on David Giesinger's site that address this somewhat indirectly.

[gerhardroux]

Thanks for the reply Gutbucket! Thank you very much for the info about Linkwitz's papers, I'll follow it up. I know Griesinger's work - he published extensively in the AES Journal - good stuff!.

To summarize my (very limited) conclusions after the input of all these kind people:

Although the same pair of ears and same brain is listening to the original and reproduced signal (thus same psychoacoustic principles apply to both sources) there is just too much limitations in microphones and speakers to accurately reproduce the original spatial image, timbre and resonances. For instance, the sound radiation of a guitar soundboard can only be reproduced perfectly by a vibrating guitar soundboard and not by a speaker which got fed a signal from a microphone that only samples air pressure variations at a certain place with a certain directivity.

Therefore, too much aspects influence this "transfer function" to limit it and give it a name.

Benade also wrote about this, his article:
Benade, A.H. (1985). From-instrument to ear in a room: Direct or via recording. Journal of the Audio Engineering Society, vol. 33, nr. 4, pp. 218–233.