Home Theater Forum and Systems banner
1 - 6 of 6 Posts

·
Registered
Joined
·
3 Posts
Discussion Starter · #1 ·
I would like to use REW to plot the amplitude response of an audio processing device (BSS Audio SoundWeb). Because it is digital, and the audio passes through A/D and D/A converters and there is some amount of delay.

After calibrating the sound card with an analog loopback, I run an SPL measurement on the SoundWeb

When the SoundWeb is programmed for flat response, I still see notches in the passband (for example at 100Hz).

The phase response is crazy, but I expect that due to the processing delay.

Is there a technique for ensuring the correct amplitude response for a system that includes delay?
 

·
Premium Member
Joined
·
6,711 Posts
If you are seeing notches when the response should be flat it is probably due to some monitoring being active, producing a feedback loop. One way to test that is to connect the device between the left output and the right input, that cross connection can break a feedback loop depending on where it is happening - best to track down the source of the monitoring if you can, though.
 

·
Registered
Joined
·
3 Posts
Discussion Starter · #3 ·
Thanks for the reply. If I disconnect the analog input, the signal level drops to below -65dB. The signal is -15 for the level check. I don't think there is any input monitoring of the output signal going on.

I have attached mdat files for a simple analog loopback around the audio interface, and looping through the SoundWeb with a "straight" through configuration.

In each case, there are notches, but at different places.
 

Attachments

·
Premium Member
Joined
·
6,711 Posts
Glitches like that are typical of interruptions in either the signal generation or capture - try listening to the sweeps and/or look at the scope plot after making a measurement to see if there are gaps. You may find you can get clean results using shorter sweeps (256k is fine for nearly all purposes) or by using ASIO drivers (you need the beta version of REW for that, set the ASIO control panel for high latency/large buffer).
 

·
Registered
Joined
·
3 Posts
Discussion Starter · #5 ·
Thanks - that was it. I installed the beta and used the Alesis ASIO drivers, and also changed the length to 256K. The plots were smooth as expected. I did notice that about 10% of the time I ran a measurement or level test, I got a "No soundcard input data" error. Clicking OK and trying again usually fixed it. Sometimes there was a related error where the measurement would report "headroom -0dB" in red and get stuck. It takes a few times of canceling and re-starting the measurement to get it to run correctly again.

If there is any further information I can provide to help with the beta, let me know. I generated the soundcard_debug.txt file, but it does not seem to include the ASIO information yet.
 

Attachments

·
Premium Member
Joined
·
6,711 Posts
Try setting the ASIO driver latency (or buffer size) setting to the maximum available, that may fix the occasional problems you have seen.
 
1 - 6 of 6 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top