REW Beta Release Frequency response compensation problem

Antonio Di Motta

Active Member
Thread Starter
Joined
Feb 8, 2018
Messages
103
Location
Italy
Hi John, I was testing REW 5.20 beta 25 on Ubuntu 18.04 PC system and external Sound Blaster SB1240 sound card on USB bus,
I connected the cables between output and input and after calibrating at 48Khz sampling rate I verified the correct calibration by tracing the response curve from 2 Hz to 22Khz.
I noticed that on very low frequencies up to 10Hz there are slight oscillations as if the answer was not compensated correctly.
For completeness, I always performed the impedance calibration with the relative curve with the Rsense standard values = 100 RLeads = 0 Rinput = 20K with the loop between input and output.
I attach the files so that you can view them as soon as you have some time ..

thanks a lot
 

Attachments

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
2,131
If you look at the soundcard calibration you will notice a large right hand window is used to properly capture the low frequency response. If you increase the right hand window of the measurement the ripples, which are due to the window curtailing the response, will disappear. Soundcard calibration is not used for impedance measurements.
 

Antonio Di Motta

Active Member
Thread Starter
Joined
Feb 8, 2018
Messages
103
Location
Italy
Sorry, but there is something I don't understand, during the calibration there are no settings that can be changed. What are you referring to ?
Only when tracing the Response curve is the Setting Length parameter that can be driven but does not affect the low frequency measurement.
I attach the files with the different settings ..
 

Attachments

Antonio Di Motta

Active Member
Thread Starter
Joined
Feb 8, 2018
Messages
103
Location
Italy
Fixed, the setting you recommended works perfectly,
I entered the value of the Right Side parameter at 1731 directly at the Analysis Preference tab level.
Are there any contraindications to insert it directly on the Preference or should it be selected for each curve from the IR Window panel?

I enclose the curves to highlight the differences.

Thanks again John
 

Attachments

John Mulcahy

REW Author
Joined
Apr 3, 2017
Messages
2,131
There is no problem with using a larger right hand window default for measuring electronic equipment, for acoustic measurements it may introduce some extra noise if it captures a lot of the noise floor. Looking at the impulse response with a dB axis allows a quick visual check of where the noise floor sits for the measurement.
 

Antonio Di Motta

Active Member
Thread Starter
Joined
Feb 8, 2018
Messages
103
Location
Italy
The measurements with a passband so extended towards the low frequencies but also towards the high frequencies (30 / 40Khz) I use them exclusively for the measurement of the pass band of the electronics.
For acoustic measurements I do not go down beyond 20Hz and therefore the setting of the Right Side parameter could easily remain the default one. Your software has lots of parameters that can be changed to
make a tunning with respect to the use you want to do.
This is one of the reasons that could require the addition of a third measure in addition to Impedance and SPL so as to be able to differentiate the tunning of the parameters with respect to the response measure
of an electronic system or with respect to the response measure of an acoustic system.
You John, you certainly don't need my advice, so everything you make is a great resource for us to use for our favorite hobby.

Thanks
 
Top Bottom