[e_resolu]

I was wondering if this method would be effcicient enough to have an empirical knowledge of the frequency response of headphones or at least for doing a comparison between headphones?

I imagine playing a music file with a very thin frequency starting from 20Hz to 20kHz over a period of time of let say 5mn. Listening to it should help at determining at what time (then what frequency) you start and stop earing something or if you ear a constant sound level over the frequency band.
That is something used for doing medical diagnostic of earing problems.

Of course it will characterise not only the headphone but also the source and the listener frequency capability but since the last two remain the same you should be able to compare between headphones.

Does anyone know how to generate such a file?

 

[ADD]

It sounds like what you are asking for is a program to generate a sine wave sweep of your specification.

You can download the demo version of the NCH Tone generator here:

http://www.nch.com.au/

You can then specify the start and end point of the tone sweep and how long it lasts and set a specific level in relation to the 0 dB reference level. I used it to calibrate the notch filtering on my IEMs.

I won't comment too much on your actual proposal or methodology, but you would really need an extended audiogram test result of your own hearing to know whether frequency response characteristics are as a result of your own hearing or the equipment's doing, as well as accurately calibrated source equipment of course.

For example, when I ran tone sweeps between 4 and 8 Khz, there was a very significant bump at 5.9 Khz as well as one at 8 Khz. But because I had an accurate audiogram that showed the exact same high sensitivity bump at 8 Khz, I only had to adjust for the bump at 5.8, where my hearing is completely flat. If I had adjusted for the 8 khz bump, then that would not have reflected the characteristics of my natural hearing.

 

[e_resolu]

Thanks you "Add" for the link and for your contribution to my newbie thinking

I finally found the file I was looking for, plus other useful ones: http://www.dogstar.dantimax.dk/testwavs/

There is an mp3 file (192 kpps) that start from 20Hz to 20kHz by 1/3 of octave every 2 seconds.

I checked rapidly my N73 + X3 + my ear response. I heard two peaks at 2Kz and 12 Kz, and a hole at 14kHz (don't have an audiogram to compare)

I saw on the etymotic site that their target is to have a bump at 12 kHz http://www.etymotic.com/ephp/er4-ts.aspx
(so I should only correct for the 2KHz bump)

 

[qazwsx]

Quote:

Originally Posted by e_resolu /img/forum/go_quote.gif Thanks you "Add" for the link and for your contribution to my newbie thinking I finally found the file I was looking for, plus other useful ones: http://www.dogstar.dantimax.dk/testwavs/ There is an mp3 file (192 kpps) that start from 20Hz to 20kHz by 1/3 of octave every 2 seconds. I checked rapidly my N73 + X3 + my ear response. I heard two peaks at 2Kz and 12 Kz, and a hole at 14kHz (don't have an audiogram to compare) I saw on the etymotic site that their target is to have a bump at 12 kHz http://www.etymotic.com/ephp/er4-ts.aspx (so I should only correct for the 2KHz bump)

A 192 kbps file ain't going to 20KHz.

 

[b0dhi]

In a 5 minute sweep the only thing you will notice would be relatively sudden peaks/dips. If all you want to do is compare headphones, you can record a sine sweep using any decent microphone. You can then compare the frequency spectrum of each recording and the differences will be obvious.

 

[Vul Kuolun]

Doing this test will give you a curve that actually should be a mixture between your individual hearing curve and the (negative) FR of the used phone:

http://www.phys.unsw.edu.au/jw/hearing.html

 

[ericj]

Quote:

Originally Posted by qazwsx /img/forum/go_quote.gif A 192 kbps file ain't going to 20KHz.

Why not? mp3 is encoded in the time domain, not in the frequency domain.

Usually the encoder has a bandpass filter on it to strip out frequencies that can't generally be reproduced or heard, but you can usually turn it off.

 

[e_resolu]

Quote:

Originally Posted by Vul Kuolun /img/forum/go_quote.gif Doing this test will give you a curve that actually should be a mixture between your individual hearing curve and the (negative) FR of the used phone: http://www.phys.unsw.edu.au/jw/hearing.html

Yeeeeessss, better accuracy for an audiogram with this method. And now I know better what is 30Hz and what is 12kHz !!!!

I obtained the same pattern than the average ones shown on the reference graph but flatter. Mabe I'm deficient in the mids or super efficient in the lows and highs !!!!