[b0ssMax]

Very very interesting. I will try this out once i figure out how to do this via mac.

Thanks Lunatique for putting this together.

 

[Solrighal]

I use Sennheiser HD 650 as my favourite headphone. Do yourselves a favour and install the Sonarworks Reference 3 plug-in. It just works.

 

[Music Alchemist]

  I'll admit to being a little confused by some of the posts in this thread. For example, the quote above. Do you mean compensate for an individual's hearing or for human hearing in general? This question might sound like semantics but it really isn't, it makes a substantial difference!
 
If we are talking about an individual's hearing, *maybe* there is some benefit to compensating for a particular deficiency that an individual might be experiencing, say a loss of sensitivity of one or both ears in a particular frequency band, due to some illness or hearing damage.
 
If on the other hand we are talking about compensating for human hearing in general, say compensating along the lines of the Fletcher-Munson curves, then this is something we/you should definitely NOT be doing! Sure, if you are listening ONLY to individual sine waves or sweeps, then by all means compensate for loudness contours if you wish but, if you are listening to commercial music, film or other content then you are listening to a mix created for human beings and by human beings (who obviously have human hearing). In other words, compensation for the Fletcher-Munson contours has already been built-in to the audio mixes to which you are listening. Applying EQ to compensate for a loudness contour is, in effect, compensating for the second time and is obviously going to take you a long way away from neutral. The only caveat I would make to this statement is if one listens to music at a very low level, in which case boosting the bass a little is not such a bad idea.
 
I would also add that most commercial recording studios are not designed with a flat/neutral frequency response, most apply a "house curve". House curves vary from studio to studio but they usually include some amount (up to about 6dB) of bass boost. This is to compensate for the fact that most consumer music systems usually have a bass boost. The exception to this is dubbing theatres (where theatrical films are mixed), which are usually quite flat, with the exception of slight bass and treble roll-offs (the x-curve). However, I mention dubbing theatres only as a point of interest, it does not affect/concern consumers.
 
G

 
I'm surprised you missed my point entirely when I explained it so clearly.
 
All I meant was that the raw measurements of a headphone (whether it's neutral or otherwise) are not going to look like a straight line. It's going to follow a curve. For example, here are measurements of the STAX SR-207 with the diffuse-field and Harman curves superimposed.
 

 
You can see many other raw headphone measurements here:
 
http://www.innerfidelity.com/headphone-data-sheet-downloads
 

  I use Sennheiser HD 650 as my favourite headphone. Do yourselves a favour and install the Sonarworks Reference 3 plug-in. It just works.

 
+1!
 
Bear in mind, however, that they use a proprietary compensation curve, so it won't sound the same as Harman and diffuse-field.
 
Here are my quick instructions for installing the plugin in foobar2000:
 

Download and install the free trial of Sonarworks Reference 3 Headphone. (Or purchase it.)  
Download and install this VST adapter in foobar2000.
 
Go to Components, VST plugins and add the Sonarworks plugin.
 
Go to Playback, DSP Manager and activate the Sonarworks plugin. Then click Configure Selected.

 

[gregorio]

  I'm surprised you missed my point entirely when I explained it so clearly.

 
1. I was responding to the thread in general rather than specifically your post, which I too thought I made clear.
 
2. You appear to now be talking about having to compensate for the frequency response of the headphones, rather than having to "compensate for human hearing".
 
G

 

[Music Alchemist]

  You appear to now be talking about having to compensate for the frequency response of the headphones, rather than having to "compensate for human hearing".

 
Please read my post and the preceding conversation again, then. I thought someone was implying that a headphone's measurements would look like a flat line on a graph when he was talking about flat...but by flat, he just meant neutral. I was contrasting electronics (which can measure as a literal flat line on a graph when they are neutral) with headphones (which never measure as a literal flat line on a graph). I also mentioned that if you manage to make a headphone follow a compensation curve perfectly, then the compensated (not raw) measurements will look like a flat line.

 

[gregorio]

  I also mentioned that if you manage to make a headphone follow a compensation curve perfectly, then the compensated (not raw) measurements will look like a flat line.

 
Exactly, so you are talking about compensating for headphones' frequency response, not compensating for human hearing. What am I missing/misunderstanding?
 
G

 

[Music Alchemist]

  Exactly, so you are talking about compensating for headphones' frequency response, not compensating for human hearing. What am I missing/misunderstanding?

 
Wow. It's like you don't even read my posts.
 
Like I said in my last post: "I was contrasting electronics (which can measure as a literal flat line on a graph when they are neutral) with headphones (which never measure as a literal flat line on a graph)." The raw measurements of headphones do not measure as a literal flat line (like electronics do) because you have to compensate for human hearing. That is what I meant. Compensated measurements have nothing to do with that, since they are just comparing how well something follows a compensation curve.

 

[gregorio]

   
The raw measurements of headphones do not measure as a literal flat line (like electronics do) because you have to compensate for human hearing. That is what I meant.

 
I think I get what you meant: You are talking about compensating for measurements taken from inside a plastic/gel head, rather than compensating for human hearing?
 
G

 

[Music Alchemist]

  I think I get what you meant: You are talking about compensating for measurements taken from inside a plastic/gel head, rather than compensating for human hearing?

 
No.
 
Electronics like amps and DACs measure as a flat line when they have a neutral frequency response.
 
Transducers like headphones do not measure as a flat line when they have a neutral frequency response, and also they never measure as a flat line either way.
 
If the raw measurements of headphones did ever measure as a flat line, they would not sound neutral to us; in fact, they would sound awful.
 
On another (UNRELATED!) note, if you do not understand the difference between raw and compensated measurements, here is an example.
 
First we have the actual raw headphone measurements represented by the blue and green lines. The dotted line on the graph is the Golden Ears compensation curve superimposed on it. Actually, I may be mistaken. It says the measured frequency response is "smoothed 1/3 oct" so these may not even be the raw measurements. But it's just an example to illustrate my point.
 

 
Now we have the compensated measurements below. It starts with the compensation curve (in this case the proprietary Golden Ears curve) as the neutral reference (thus shown as a flat line) and then shows how the headphone measurements align with it. If the headphone measurements (whether before or after EQ) followed that particular curve perfectly, then the headphone measurements would look like a flat line on the graph—but only for the compensated measurements, never the raw ones.
 

 
I suspect that you are confusing various meanings of the word compensated without getting the context each time.

 

[gregorio]

  Transducers like headphones do not measure as a flat line when they have a neutral frequency response ...

 
Mics and speakers are transducers like headphones and they do measure flat (in an anechoic chamber) when they have a neutral frequency response.
 
G

 

[Music Alchemist]

  Mics and speakers are transducers like headphones and they do measure flat (in an anechoic chamber) when they have a neutral frequency response.

 
Missing the context again. -_- Forget I used the word transducers. Just pay attention to what I am saying about headphones.

 

[briskly]

 
I think I get what you meant: You are talking about compensating for measurements taken from inside a plastic/gel head, rather than compensating for human hearing?

No.

Electronics like amps and DACs measure as a flat line when they have a neutral frequency response.

Transducers like headphones do not measure as a flat line when they have a neutral frequency response, and also they never measure as a flat line either way.

If the raw measurements of headphones did ever measure as a flat line, they would not sound neutral to us; in fact, they would sound awful.

These types of measurements include the frequency response of the ear canal and the diffraction of the outer ear. If you had noticed from my previous post, there was a measurement of a headphone's response from inside the headphone paired with the subjectively determined response of the same headphone.

The sound input before entering the ear should still be flat.

 

[RRod]

These types of measurements include the frequency response of the ear canal and the diffraction of the outer ear. If you had noticed from my previous post, there was a measurement of a headphone's response from inside the headphone paired with the subjectively determined response of the same headphone.

The sound input before entering the ear should still be flat.

 
For a headphone not necessarily, in the sense of them being like speakers at 90 degrees. If you want a signature like speakers of smaller azimuth, then you'd need a non-flat response away from the ear.

 

[Music Alchemist]

These types of measurements include the frequency response of the ear canal and the diffraction of the outer ear.

 
Yes, when I talk about raw headphone measurements, I am referring to the type taken by InnerFidelity with a dummy head, as that seems to be the industry standard.
 

If you had noticed from my previous post, there was a measurement of a headphone's response from inside the headphone paired with the subjectively determined response of the same headphone.
The sound input before entering the ear should still be flat.

 
As for this post and what you have added, the information is not specific enough for me to fully understand.
 
Elaboration is appreciated, especially pertaining to your last statement.

 

[castleofargh]

Music Al was only talking about how a measurements without compensation would not help getting a neutral sounding signature.
but after finding a more or less proper compensation curve (be it diffuse field, the goldenears one, whatever sonawork uses, etc),they are all still based on one model with no guaranty that the random guy will fit inside that model. so the last step of all EQ will still depends on the listener himself.
 
I can only once again past this vid as it's IMO the clearest thing I've seen (even if a little long) about the problems/limitations that exist in any headphone measurement:

 
the question for this topic IMO is to find out how close a sine sweep can get us? to me it's a powerful tool to get rid of massive spikes, but not an easy thing to use to get neutral sound as it will tend to lead more toward equal loudness contour than perceive neutral outside world sounds . but again it depends a lot on our own experience of sine sweeps. if one can get used to listening to it on a neutral source, then he can do a nice job at EQing a headphone.