record the output of a soundbox look strange in excessphase and rt60.

[user44455555]

Hello

I think there is no feedback, because distortion is very low and FR is very linear. I choose estimate IR and create minimum phase. attach is the mdat file too . maybe a setting is not good or REW problem ?

 

[John Mulcahy]

The RT60 curve is from the octave band filter delay, there are options to reduce it - see the RT60 graph help.

 

[user44455555]

The RT60 curve is from the octave band filter delay, there are options to reduce it - see the RT60 graph help.

I choose third octave this give more delay in low. I choose time reversed filtering it reduce to 200 ms at 50 hz and then filter order 6 then i get 120 ms. have the time reversed filtering any disadvantages ?. At least good to see that it is not only my room that increased that so much.
is it possible to add for settings a compensation. measure a theoretic perfect signal as reference example displayValue50hz=result50hz - reference50hz. displayValue60hz=result60hz -reference60hz ?.

same maybe work for the phase ?.

 

[user44455555]

or a better idea . is it possible when do a calibrate soundcard that this is the reference and the result of the soundbox loopback is subtract from the measurement automatic ?.

 

[John Mulcahy]

is it possible to add for settings a compensation

No, the effect of the filters depends on the signal.

same maybe work for the phase ?

No, phase is affected by the position of t=0. Adjust the t=0 position or try estimate IR delay to alter it. Minimum phase generation needs to use suitable low and high frequency tails for best accuracy, but note that DAC reconstruction and anti-alias filters are usually not minimum phase.

 

[user44455555]

No, phase is affected by the position of t=0. Adjust the t=0 position or try estimate IR delay to alter it. Minimum phase generation needs to use suitable low and high frequency tails for best accuracy, but note that DAC reconstruction and anti-alias filters are usually not minimum phase.

I do always estimate IR delay and it shift 0,1 sample. do you think this is correct ? . the step response look strange is this because the sounddevice is very bad or is that normal look ? . it is a focusrite scarlett which is test good

the impulse response

 

[John Mulcahy]

Yes, it's normal.

 

[user44455555]

then the question is how can get correct results in phase and RT 60 ?. if there is no exact way then i think it is more near the truth how much phase correction need when subtract in my case 67 degree when i want see real phase at 50 hz

 

[John Mulcahy]

You can compensate for the soundcard's phase response using the soundcard cal measurement. It is not meaningful to look at the excess phase for most soundcards as they do not have minimum phase reconstruction filters, most are linear phase.

 

[user44455555]

It is not meaningful to look at the excess phase for most soundcards as they do not have minimum phase reconstruction filters, most are linear phase.

this explain alot. I do soundcard cal then the phase get better, but excess phase go more minus. And when i measure a speaker i compare and i see now that in bass the excess phase go in that value lower as the excess phase from soundcard go lower. so better dont use excess phase for speaker measurement when have a soundcard with linear filter which can see on high minus excess phase. I look at excess phase in the past because that give nicer results on speakers so i assume it can better reduce room influence. Seem the best is to look for timing to use the spectrogram this look linear on soundcard even if not calibration file for soundcard use. So i assume soundcard have then no influence on this.

maybe a excess phase aproximate setting for soundcards with linear filter can help

 

[user44455555]

I do measure now from 2 hz instead 20 hz of soundbox loopback. its low level on a 16 bit UMC22 bwehringer USB from 2 hz let the excess group delay at 40 hz much more valid. for rt60 or rt60 decay this not help.

from 20 hz