jerg
Headphoneus Supremus
- Joined
- Aug 11, 2010
- Posts
- 6,126
- Likes
- 864
Maybe try to re-measure a few sets of data, each with some minor variations in the earcup placement on your measuring instrument? That could uncover what is causing this.
Headphone CSD waterfall plots
- Thread starter purrin
- Start date
vid
Headphoneus Supremus
- Joined
- Jan 5, 2005
- Posts
- 2,063
- Likes
- 130
Quote:
Artificial ear and ear canal leading to an electret capsule mic. An MLS signal.
Quote:
I'll see about doing it later today.
Artificial ear and ear canal leading to an electret capsule mic. An MLS signal.
Quote:
I'll see about doing it later today.
stv014
Headphoneus Supremus
- Joined
- Jul 17, 2011
- Posts
- 3,493
- Likes
- 273
Using the programs from dsputils.zip (see description here), I have created a script for a fairly complete set of headphone measurements. I am not sure if anyone finds this useful, though, since the user friendliness and the documentation are lacking, but maybe someone will find it interesting:
hptest.zip
It uses the following input files:
- the test signals recorded from the headphone (the test does take a few minutes, so the sample could be optimized for shorter length). resample.txt needs to be set up with the correct start time and gain parameters so that the 1 kHz tone at the beginning of the file starts at exactly 1 second, and has an amplitude of 0 dBFS
- the impedance test file (a loopback recording using a device with a known resistive output impedance; the test signal consists of two MLS loops, the headphone load needs to be connected during the second one)
- global equalization IR file (it is applied to all measurements)
- frequency response equalization IR file (it is only applied to the FR graph)
The graphs created include:
- frequency response (from impulse response, equalized; it does not average the response from multiple positions, though, which would be a useful feature)
- CSD (left and right channel)
- 500 Hz THD spectrum at 90 and 100 dB SPL (note: for the SPL values to be correct, the test file should be played at a volume setting at which 0 dBFS = 105 dB SPL - the drivers need to be able to handle this, and a long sine sweep at 100 dB)
- 55 Hz + 2000 Hz 4:1 IMD spectrum at 90 and 100 dB SPL
- 15000 + 16000 Hz 1:1 IMD spectrum at 90 and 100 dB SPL
- THD vs. frequency at 90 and 100 dB SPL
- THD vs. level at 50 Hz, from 80 to 105 dB SPL
- group delay
- impedance vs. frequency
The package also contains some sample graphs. However, since I did not use a high quality microphone and dummy head or artificial ear, these are not particularly accurate, and are really only examples.
hptest.zip
It uses the following input files:
- the test signals recorded from the headphone (the test does take a few minutes, so the sample could be optimized for shorter length). resample.txt needs to be set up with the correct start time and gain parameters so that the 1 kHz tone at the beginning of the file starts at exactly 1 second, and has an amplitude of 0 dBFS
- the impedance test file (a loopback recording using a device with a known resistive output impedance; the test signal consists of two MLS loops, the headphone load needs to be connected during the second one)
- global equalization IR file (it is applied to all measurements)
- frequency response equalization IR file (it is only applied to the FR graph)
The graphs created include:
- frequency response (from impulse response, equalized; it does not average the response from multiple positions, though, which would be a useful feature)
- CSD (left and right channel)
- 500 Hz THD spectrum at 90 and 100 dB SPL (note: for the SPL values to be correct, the test file should be played at a volume setting at which 0 dBFS = 105 dB SPL - the drivers need to be able to handle this, and a long sine sweep at 100 dB)
- 55 Hz + 2000 Hz 4:1 IMD spectrum at 90 and 100 dB SPL
- 15000 + 16000 Hz 1:1 IMD spectrum at 90 and 100 dB SPL
- THD vs. frequency at 90 and 100 dB SPL
- THD vs. level at 50 Hz, from 80 to 105 dB SPL
- group delay
- impedance vs. frequency
The package also contains some sample graphs. However, since I did not use a high quality microphone and dummy head or artificial ear, these are not particularly accurate, and are really only examples.
vid
Headphoneus Supremus
- Joined
- Jan 5, 2005
- Posts
- 2,063
- Likes
- 130
Beyer DT 100 (400-ohm)
Measured without the artificial ear, but it looks close enough to the Golden Ears measurement and seems about what I hear.
Some actually do call these headphones neutral and accurate.
Measured without the artificial ear, but it looks close enough to the Golden Ears measurement and seems about what I hear.
Some actually do call these headphones neutral and accurate.
Reticuli2
500+ Head-Fier
- Joined
- Dec 20, 2007
- Posts
- 530
- Likes
- 14
Anyone have an REW measurement response file for the HD25 they could send me? I want to do a parametric eq on it.
jiiteepee
Headphoneus Supremus
- Joined
- Apr 5, 2005
- Posts
- 1,606
- Likes
- 62
Can you use these curves:
http://graphs.headphone.com/graphCompare.php?graphType=0&graphID[]=513&graphID[]=523&scale=20
+
http://jiiteepee.tripod.com/peqgui-10mc.html (read the documentation)
It's not perfect way but ...
http://graphs.headphone.com/graphCompare.php?graphType=0&graphID[]=513&graphID[]=523&scale=20
+
http://jiiteepee.tripod.com/peqgui-10mc.html (read the documentation)
It's not perfect way but ...
Users who are viewing this thread
