I have made a RPI-DSP using the tutorial of Richard Taylor and the patient help of @chalielaub and @phofman. http://www.diyaudio.com/forums/pc-based/316737-ecasound-dsp-startup-difficulties.html#post5305706
I have tested different configurations and currently am working on a 3 way system with three class D amplifiers. For the bass I use the Hypex UCD400 OEM and for mid and high I use two TDA7498. I have a calibrated microphone and use Arta to measure the frequency response. Using the LADSPA biquads of Richard Taylor I can configure the crossover and realize a more or less flat response. I used to position the microphone about 80 cm in front of one of the loudspeakers but noticed a big difference with the response at the listening position. So now I measure at listening position using both speakers. Gated measurements are not possible in this way. So the response is affected by reflections in the room and the frequency and phase response show a rather erratic graph (if no smoothing is used).
Until now I haven’t implemented corrections for phasing and I am uncertain how to deal with this. I have tried to use the FRD respons blender and minimum phase detector of Charlie Laub but the procedure is rather complicated and I got problems with different excel versions. I tried another experiment just by measuring the delay of the response in Arta. I discovered that the start of the (white noise) impulse as shown in Arta is at different moments of time for each speaker. For the mid and high this difference is small, but the woofer seems to be delayed considerably. I tried to delay the mid and high by adding the difference by using the following -el:delay_0.01s,0.013,1 This line of code is used by Richard Taylor but I can’t find documentation of this. Anyway, I tried it but nothing changed in the measurements of Arta. So given my setup, what would be the right and most convenient way to deal with phasing and delay?
I have tested different configurations and currently am working on a 3 way system with three class D amplifiers. For the bass I use the Hypex UCD400 OEM and for mid and high I use two TDA7498. I have a calibrated microphone and use Arta to measure the frequency response. Using the LADSPA biquads of Richard Taylor I can configure the crossover and realize a more or less flat response. I used to position the microphone about 80 cm in front of one of the loudspeakers but noticed a big difference with the response at the listening position. So now I measure at listening position using both speakers. Gated measurements are not possible in this way. So the response is affected by reflections in the room and the frequency and phase response show a rather erratic graph (if no smoothing is used).
Until now I haven’t implemented corrections for phasing and I am uncertain how to deal with this. I have tried to use the FRD respons blender and minimum phase detector of Charlie Laub but the procedure is rather complicated and I got problems with different excel versions. I tried another experiment just by measuring the delay of the response in Arta. I discovered that the start of the (white noise) impulse as shown in Arta is at different moments of time for each speaker. For the mid and high this difference is small, but the woofer seems to be delayed considerably. I tried to delay the mid and high by adding the difference by using the following -el:delay_0.01s,0.013,1 This line of code is used by Richard Taylor but I can’t find documentation of this. Anyway, I tried it but nothing changed in the measurements of Arta. So given my setup, what would be the right and most convenient way to deal with phasing and delay?
Attachments
Last edited:
The download files for Charlie Laub's active crossover design tool includes a well written tutorial that covers phase adjustments very well:
the Active Crossover Designer web page
He also has his own LADSPA filters available (including a delay filter 'mTap') here:
LADSPA-plugins
the Active Crossover Designer web page
He also has his own LADSPA filters available (including a delay filter 'mTap') here:
LADSPA-plugins
Could it be these are Single-channel ARTA measurements? If so, look here, because trouble might already start there:
Arta and phase measurment
First of all, the timing of the measurements needs to be correct.
Good look and keep us posted.
Eelco
Arta and phase measurment
First of all, the timing of the measurements needs to be correct.
Good look and keep us posted.
Eelco
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.