Hi, (already posted in PC Thread but maybe better here if soundcard and mic not guilty !)
I hope to measure my speaker but I can't. Just want to see big default : > + 3 db peak after 2 khz.
I use ARTA, Tracker Pre usb soundcard checked by ARTA, EMC-8000 mic without calib file. ASIO Creative driver and the USB2 out of my Windows XP Laptop.
I was advised to use sweep sine with peak hold to low the problems of room with simple Pink noise measurement. I use the sweep sine wave of audiocheck web site.
Is there a protocol or can i be advised for the setting up of ARTA :
Windows: Uniform, Hanning, flat top ?
Averaging : peak hold : I suppose
Smoothing function : on ?
Frequency & FFT : ?
Which set up for the windows graph : 100 DB with 40 DB to have 5 DB from line to line ?
Is there somebody who can show me a typical SPA curve without measurement (before Rec) because mine is very very non flat... and I d'ont understand if it's the noise or a normal thing at - xx DB (no sound). How much DB amplitude between the 2 curves: silence and with generated noise ?
Whith a pink noise or a sweep sine (are they non periodic noise ?) how much DB levelfor a correct measurement : 60 DB ?
I flatened the curve above 1k hz with my own calib file after measured my room in silent mode at 4 am
!
Sorry if basics questions but need help please, i found many papers about measurement but none for beginners with ARTA and none showing the right set up to expect trusted measurement. Maybe my bad, not efficient with googling in english
I hope to measure my speaker but I can't. Just want to see big default : > + 3 db peak after 2 khz.
I use ARTA, Tracker Pre usb soundcard checked by ARTA, EMC-8000 mic without calib file. ASIO Creative driver and the USB2 out of my Windows XP Laptop.
I was advised to use sweep sine with peak hold to low the problems of room with simple Pink noise measurement. I use the sweep sine wave of audiocheck web site.
Is there a protocol or can i be advised for the setting up of ARTA :
Windows: Uniform, Hanning, flat top ?
Averaging : peak hold : I suppose
Smoothing function : on ?
Frequency & FFT : ?
Which set up for the windows graph : 100 DB with 40 DB to have 5 DB from line to line ?
Is there somebody who can show me a typical SPA curve without measurement (before Rec) because mine is very very non flat... and I d'ont understand if it's the noise or a normal thing at - xx DB (no sound). How much DB amplitude between the 2 curves: silence and with generated noise ?
Whith a pink noise or a sweep sine (are they non periodic noise ?) how much DB levelfor a correct measurement : 60 DB ?
I flatened the curve above 1k hz with my own calib file after measured my room in silent mode at 4 am
!Sorry if basics questions but need help please, i found many papers about measurement but none for beginners with ARTA and none showing the right set up to expect trusted measurement. Maybe my bad, not efficient with googling in english
You might find this tutorial that I put together awhile back helpful:
Making gated-impulse frequency measurements using ARTA
-Charlie
Making gated-impulse frequency measurements using ARTA
-Charlie
THANKS A LOT GUYS,
I made a lot of measurement with my strange own flated mic calib in home. Made today a passive notch at 7400 (3 to 4 db bumps measured before)... not very happy with it, loy of peak energy has diseapered but a lot of fastness and energy in bass... understand more and more the critcism choice of tweeter !
I have to read your docs for an acurate how to and measure again!
I made a lot of measurement with my strange own flated mic calib in home. Made today a passive notch at 7400 (3 to 4 db bumps measured before)... not very happy with it, loy of peak energy has diseapered but a lot of fastness and energy in bass... understand more and more the critcism choice of tweeter !
I have to read your docs for an acurate how to and measure again!
Hi Charlie,
I read and re read your how to. It's very clear. I understand i have to use hanning windows to measure my speakers.
But I have a little problem because I don't use wire from my sound-card to the preamp. I work with my squeezebox and wav made by :Sine Sweep - Full Spectrum (20 Hz - 20 kHz) ; very usefull you can create on line anothers than the generics like this one going 20 hz to 20 khz
How to understand your how to in my case :
- is the sequence length 128k a multiple of the 0-22khz lenght ? Why only 48000 sampling rate ?
- Can I use the audiocheck .wave instead of the self generated Artas' imput signals ? Is this swept sine equivalent of the log-frequency sweep in Arta (your how to)
- Do I have to use Impulse response in Arta with AUdiochecknek instead of Spectrum analyser mode ? (don't really understand difference with impulse and spectrum here).
I see that in high frequency Pink noise give nearly the same result in my room than Swept sine, are they equal in quality when speaking of frequency > at 1 khz in home rooms ?
You talk about a calib mic, do you think it's a valid way to calib yourself a mic in your room the night when you silence : that's what I made witout calib file : i made my own frd file to have a very flat curve >500 hz (laptop with no fan and no noise inside and outside during the captures).
Sorry, a lot of questions... I hope it can help another people too.
I read and re read your how to. It's very clear. I understand i have to use hanning windows to measure my speakers.
But I have a little problem because I don't use wire from my sound-card to the preamp. I work with my squeezebox and wav made by :Sine Sweep - Full Spectrum (20 Hz - 20 kHz) ; very usefull you can create on line anothers than the generics like this one going 20 hz to 20 khz
How to understand your how to in my case :
- is the sequence length 128k a multiple of the 0-22khz lenght ? Why only 48000 sampling rate ?
- Can I use the audiocheck .wave instead of the self generated Artas' imput signals ? Is this swept sine equivalent of the log-frequency sweep in Arta (your how to)
- Do I have to use Impulse response in Arta with AUdiochecknek instead of Spectrum analyser mode ? (don't really understand difference with impulse and spectrum here).
I see that in high frequency Pink noise give nearly the same result in my room than Swept sine, are they equal in quality when speaking of frequency > at 1 khz in home rooms ?
You talk about a calib mic, do you think it's a valid way to calib yourself a mic in your room the night when you silence : that's what I made witout calib file : i made my own frd file to have a very flat curve >500 hz (laptop with no fan and no noise inside and outside during the captures).
Sorry, a lot of questions... I hope it can help another people too.
You can't use some other test sweep with ARTA. You MUST playback the signal generated by ARTA, in real time (not by recording it and then replaying it later) while the computer is "listening" for the response by microphone. ARTA is recording the amplitude as a function of time and then processing that into the frequency response.
Read the ARTA manual regarding the sequence length. The 48kHz sample rate is not mandatory, but the value you use will be limited to your soundcard's maximum rate. The total time that is recorded by ARTA is equal to the number of samples (the sequence length) divided by the sampling rate. The appropriate and/or necessary settings depend on how long you want to record the response from the microphone and the delay before the sound arrives at the microphone (including delay of the signal as it travels within your sound chain). For instance I used a wireless audio unit at one time to send the signal over to the amplifiers, and this added several milliseconds of delay to the beginning of the sampled time period during which there was no signal detected.
I don't quite see how you can come up with the mic calibration file just by doing something in a silent room. The calibration file is what the microphone will "measure" when no correction is applied to a uniform SPL sound source. The "correction" is the inverse of that curve. The only way I know to obtain a calibration on your own is to compare the measured response of your microphone to the response measured by a known, calibrated mic when the measurement is done at the exact same position in space on the exact same signal.
Read the ARTA manual regarding the sequence length. The 48kHz sample rate is not mandatory, but the value you use will be limited to your soundcard's maximum rate. The total time that is recorded by ARTA is equal to the number of samples (the sequence length) divided by the sampling rate. The appropriate and/or necessary settings depend on how long you want to record the response from the microphone and the delay before the sound arrives at the microphone (including delay of the signal as it travels within your sound chain). For instance I used a wireless audio unit at one time to send the signal over to the amplifiers, and this added several milliseconds of delay to the beginning of the sampled time period during which there was no signal detected.
I don't quite see how you can come up with the mic calibration file just by doing something in a silent room. The calibration file is what the microphone will "measure" when no correction is applied to a uniform SPL sound source. The "correction" is the inverse of that curve. The only way I know to obtain a calibration on your own is to compare the measured response of your microphone to the response measured by a known, calibrated mic when the measurement is done at the exact same position in space on the exact same signal.
THANKS, it helps a lot...
I was hoping that a silent room near 0 DB > 1000 hz was a 0 db uniform SPL sound source". Then seing for example +1 or 2 db above zero DB line is the natural rising of the mic ?! the 1 khz naturaly overlay the zero db line as predicted in all the test with ECM8000.... above and below I put correction to flat the curve of the mic, saying myself it will be true only above 1 khz because the room at low hz...
If i understand what you wrote, I can't use in SPA mode an external .wave sweep sine or others rec when Rec switch is on with Arta ! Don't undersdtand why audiocheck makes such .wave ? Is it just with ARTA or have I been missanderstood ? Do you talk only about the impulse method ? SPA option is not accurate ?
I was hoping that a silent room near 0 DB > 1000 hz was a 0 db uniform SPL sound source". Then seing for example +1 or 2 db above zero DB line is the natural rising of the mic ?! the 1 khz naturaly overlay the zero db line as predicted in all the test with ECM8000.... above and below I put correction to flat the curve of the mic, saying myself it will be true only above 1 khz because the room at low hz...
If i understand what you wrote, I can't use in SPA mode an external .wave sweep sine or others rec when Rec switch is on with Arta ! Don't undersdtand why audiocheck makes such .wave ? Is it just with ARTA or have I been missanderstood ? Do you talk only about the impulse method ? SPA option is not accurate ?
Last edited:
If you need to do FR measurements with playback of pre-recorded signals instead of wiring the soundcard to the amp, you may have a look af SynRTA (freeware): http://www.libinst.com/SynRTA.htm
--
--
- Status
- Not open for further replies.