Demystifying the "flat curve"?

Aug 11, 2004 at 7:00 AM Thread Starter Post #1 of 67

lindrone

King Canaling
Joined
Aug 25, 2003
Posts
3,887
Likes
27
By now, we all know how Ultimate Ears has been advertising its "flat curve" as the holy grail of sound for quite a while now. However, the fact remains that no other headphone manufacture even claims that flat curve is the way to make headphones. Today I went to my audiologist to pick up my new box from Sensaphonics, and I talked to her a bit about why other manufacturers (not just IEM's, but just sound equipments in general) don't seem to be concerned with this flat curve idealism.

Well, I learned something that I didn't know before. Apparently that human ear doesn't preceive sound level being the same at all frequencies, so if we do indeed get a flat frequency response in our perception of sound, we're actually not getting frequencies that are flat, but those that has been adjusted against our sound perception to become flat. I tried doing some more research on the web to see if there's more information about this... and this is all I managed to come up with:

http://www.termpro.com/articles/weight.html
http://www.termpro.com/articles/hearing.html

Two articles written by the same guy. Here's the quote to take away from the article if you didn't read it all the way through:

"In order to appreciate the need for weighting networks, it is first necessary to acquire a basic working knowledge of the way we perceive sound. Most significant is the fact that our ears are not equally sensitive to all frequencies, and although they may appear to work very well, their frequency response is actually far from "flat." To complicate things even further, this variance in hearing sensitivity is more pronounced at low Sound Pressure Levels (SPLs) than at high ones."

Basically, what it comes down to, is if you were able to measure a flat curve coming directly out of the equipment, it will not be perceived as flat by our ears/brains. I guess additional equipments needs to be used to simulate the how our ears perceive sound in order to get a true, perceived flat response.

So this brings me to the question, when Ultimate Ears conduct these measurements, are they using any sort of a weighted measurement? Or are they simply taking measurements in some sort of a chamber directly out of the equipment? In fact, does Headroom do any sort of a weighted measurement when conducting their frequency response tests?

Anyway, I'm just very curious as to how speaker/headphone manufacturers measure their equipment, and perhaps the way that we've been measuring our equipment has been wrong all along? Technical measurement is something I'm not very familiar with (not just reading the charts, but actually conducting the test environment). Can anyone chime in on this?

I was also told that Sensaphonics accounted for the way that ears responds to different frequencies when they designed their headphone. I'm sure that other headphone manufacturers does as well, could this be why we don't get a measured flat curve coming out of our speakers/headphones?
 
Aug 11, 2004 at 7:10 AM Post #2 of 67
Also, perhaps, each person has variations in which frequencies are emphasized. So, what may be "flat" to one person is different from another, because each person hears things differently.
 
Aug 11, 2004 at 2:40 PM Post #4 of 67
The problem is... none of these companies publish how they conduct their tests. So if it's simply a microphone picking up the sound from the earphone in a little enclosed cup mechanism or something, then the "flat curve" coming out of the earphone will invariably sound crappy to our ears, because we don't perceive frequency flatly at all.

Anyone have any idea on how companies really conduct these tests?
 
Aug 11, 2004 at 3:17 PM Post #5 of 67
Music performed -> flat response mic -> flat response reproduction -> ears with their non-flat response

Music performed -> ears with their non-flat response

Obviously flat measured response from sound reproduction is closer to actual music performance than sound reproduction with non-flat response.
 
Aug 11, 2004 at 3:42 PM Post #6 of 67
Once again we are reminded not to rely solely on specs to judge equipment. Always audition!
biggrin.gif
 
Aug 11, 2004 at 3:43 PM Post #7 of 67
Is live music really flat though? I guess that's another question. I really don't know what type of equipments that musician use in a live performance, but I know that mics are made to have flat response. However, if you're talking about the various amps and soundboards that all the other equipments are connected into and then tweaked, do they still maintain that the sound coming out of the speakers are flat?

Furthermore, there's always the acoustic effect of the arena they're playing in, so even live classical performance shouldn't really be completely flat either, since the acoustic of the concert hall plays a part in it, doesn't it?

I know for sure mastered recordings really aren't flat, because producers has to mix for crappy, mediocre equipments anyway.
 
Aug 11, 2004 at 4:08 PM Post #8 of 67
Quote:

Originally Posted by lindrone
Is live music really flat though?


The frequency response is dependant on the specific reference point of measurement. Obviously, the farther you move from the sound source the less high treble energy remins(air moleculs attenuate high frequencies), the more left of right you move in relation to a stage(if close to the stage or pit) the more predominant some instruments will be as opposed to others, etc. So really you can only refer to 'flat' as far as the specific point in space the mics are placed. Concerning close-miced recordings as ar done in studios; I doubt these can ever sount completely 'accurate' due to the un natural frequency response of recording the objects so close. SIgnificant equalization would be needed in order to obtain a natural sound.



Quote:

but I know that mics are made to have flat response.


Sadly, many mics have significanlty non-linear response. In studio recordings, it's not uncommon for vintage mics to be used which have terribly deviant response because someone likes this coloration. Classical music recordings on location seem to more often than not use microphones that are linear, though.

Quote:

Furthermore, there's always the acoustic effect of the arena they're playing in, so even live classical performance shouldn't really be completely flat either, since the acoustic of the concert hall plays a part in it, doesn't it?


Yes. The farther back from the actual performance in a contained environment like a concert hall, the more dominant the reflected sound will be.

Quote:

I know for sure mastered recordings really aren't flat, because producers has to mix for crappy, mediocre equipments anyway.


I suppose this depends on the objective/mindsets of the people actually doing the work. Seems this is usually true, though. Not even considering the ignorant levels of dynamic compression that usually turns out on studio mastered recordings....

-Chris
 
Aug 11, 2004 at 4:15 PM Post #9 of 67
Quote:

Originally Posted by breez
Music performed -> flat response mic -> flat response reproduction -> ears with their non-flat response

Music performed -> ears with their non-flat response

Obviously flat measured response from sound reproduction is closer to actual music performance than sound reproduction with non-flat response.



You are correct, but you have to consider the perspective in which the music was captured. Listen to some stereeo speakers, now turn your head so that one of your ears is directly facing a speaker. A sudden drastic raise in upper midrange/treble is now present. The ear acts as a filter, relative to the ange to source. With a perfect flat linear recording, the speaker is essentially outputting the sound where the mics picked it up, your head/ears still get to filter the sound. With a headphone, their is no opportunity for this to occur since it is directly outputting into your ear.

-Chris
 
Aug 11, 2004 at 4:18 PM Post #10 of 67
Quote:

Originally Posted by breez
Music performed -> flat response mic -> flat response reproduction -> ears with their non-flat response

Music performed -> ears with their non-flat response

Obviously flat measured response from sound reproduction is closer to actual music performance than sound reproduction with non-flat response.




That is almost correct, but the way I understand this it is:

Music performed -> flat response mic -> SPEAKERS with flat response reproduction -> ears with their non-flat response

-or-

Music performed -> BINAURAL response mic -> HEADPHONES with flat response reproduction -> ears with their non-flat response

So according to this speakers should have the response as flat as possible while headphones playing non-binaural recordings should have a curved resopnse to make up for the fact that sound waves don't have to pass through your head and interact with ears and so on. At least that's my theory
smily_headphones1.gif
 
Aug 11, 2004 at 4:29 PM Post #11 of 67
Quote:

Originally Posted by HiWire
Once again we are reminded not to rely solely on specs to judge equipment. Always audition!
biggrin.gif



yeah, too bad I can't audition the UE10 and the Sensa without actually buying both.
frown.gif
 
Aug 11, 2004 at 4:51 PM Post #12 of 67
Quote:

Originally Posted by WmAx
Sadly, many mics have significanlty non-linear response. In studio recordings, it's not uncommon for vintage mics to be used which have terribly deviant response because someone likes this coloration. Classical music recordings on location seem to more often than not use microphones that are linear, though.


Well said, Chris. The only reason why I said that I know mic have flat response, is that I know some mics are made that way, and Shure definitely market a lot of their high-end mic as being having a flat response. No doubt there's a lot of recordings that's done with non-linear mic.

In comparison, I'm just completely unaware of any speakers used in concert situations that has flat response. Of course, I have not done any research or looked at any graphs for any speakers. I just don't know if such a thing exist... as much (or as little) as headphone that has flat curve exists.

Also, in a studio recording environment, when there's just a mic next to the artists mouth/instruments, all the sound cues are added through the production methods, right? Things like spatial separation and soundstage are all a part of manipulation done through the production... am I correct in assuming this?
 
Aug 11, 2004 at 5:03 PM Post #13 of 67
Quote:

Originally Posted by lindrone
In comparison, I'm just completely unaware of any speakers used in concert situations that has flat response.


I was referring to non amplified concerts(classical, opera, etc.). In case of amplified concerts, their is nothing natural about such presentations. SIngers are close miced, sound effects and soundstage are all produced through the effects rack/mixing board.

Quote:

Also, in a studio recording environment, when there's just a mic next to the artists mouth/instruments, all the sound cues are added through the production methods, right? Things like spatial separation and soundstage are all a part of manipulation done through the production... am I correct in assuming this?


Correct. All components are close miced in specially built acoustic isolation rooms. The recording is 'produced' later using these discrete recorded tracks.

-Chris
 
Aug 11, 2004 at 8:16 PM Post #14 of 67
There are two things to consider when talking about “flat response”. One is that due to the Flecter-Munsion effect (see http://www.webervst.com/fm.htm), people’s hearing is not flat. The degree of this varies with volume, from person to person and with age for any given person. The thing to realize is that no attempt to “fix” this should be made, because it is not a problem. Everything a person hears is with their own hearing profile, and to that person anything that deviates from that is not flat. So if a person listened to an instrument, such as a piano for example, in a large anechoic chamber so that any reflected sound did not reach the listener (just to eliminate a complication), a recording and playback of that same piano would need to be completely flat in order to sound the same as the live piano. Note that this has nothing to do with our individual ears' frequency response; two people with very different hearing would have the same requirement of a flat recording and playback to sound flat to them. So this is a goal for many people…recreating the original performance (high fidelity). But for reasons discussed below as well as mentioned by other people in this thread, this is not easy to do. For other people the goal is just to have something sound good…but what defines that is up to the individual and may have little to do with high fidelity. That’s fine with me too, we don’t all have to have the same goals and therefore the same equipment to achieve those goals.

The fact that we do not listen in anechoic chambers is the other consideration. Sounds get absorbed and reflected on the way to us, and the amount of this varies with frequency and distance and of course the surroundings. In a room, sounds can also get reinforced at low frequencies. These effects don’t get naturally compensated in a close miked recording, but often get compensated in playback in a normal room with loudspeakers. Headphones of course eliminate these compensations, so they often have their own equalization by way of a non-flat frequency response curve.
I had season tickets to the San Jose Symphony before they went bankrupt, and one year I did a lot of experimenting with seating through exchanges. The sound was quite different in Row 3 vs Row 9 vs Row 17, not only in the volume and the amount of reflected vs direct sound, but also in terms of perceived frequency response. Of course they are all “correct” even though different. Which one I preferred is a personal preference, just as which headphone I might prefer, maybe because it matches the sound of my seating preference.

There is a lot of variation in the way recordings are made, so this is another wildcard to consider. Some audio equipment have tone controls to help with this, although they often make things worse.

The bottom line is that there is no one single correct solution. We have to deal with the source material that is out there, and make it sound the best we can to our own preferences. If someone tells me that a certain headphone sounds "flat", I'm not sure what that means unless I know that person's taste. But if they say that headphone A is brighter than headphone B and I have heard either A or B then that can be useful. All this because we just want to enjoy the music as much as possible.
 
Aug 11, 2004 at 9:49 PM Post #15 of 67
So, in your opinions, what do you guys think is the value of playback equipments that claim to have flat curves? I mean, I understand the need for flatness from a source perspective, you want to make sure as much purity as possible is preserved from the recording before it reaches the amp & speaker/headphone stage. However, if there are so many different situations that one has to account for, flatness in the final playback equipment doesn't seem to matter all that much.

I've always had the believe that curves don't say anything about the way any particular equipment really sounds like. In the end it comes down to sound signature that can't be respresented by some frequency measurements alone.

Do you guys think that in general, it's more important to just arrive at the sound that you want, and frequency response charts really doesn't mean anything? Or is it still important to produce equipments that gets as close to flat response as possible?
 

Users who are viewing this thread

Back
Top