Multiple sweeps may not work when the input and output devices are different - the OS may not maintain sample-accurate synchronisation between the device audio streams, so the synchronous averaging fails. There is a warning note about that in the help.
REW freq. resp. errors at high samples/ multiple scans
Hello,
Well, I made it through the intimidating signup process. :sweat:
I just started using REW 5.13 today and it's been good fun, but I did find a problem. I didn't simply use the default settings for the frequency scan, I had to crank it up to the max samples and multiple scans to see if there was a noticeable improvement, but there's actually a problem with doing this for full-range measurements. The first 5 plots were from scans I did on my DIY sub/ mid-bass "half" speaker with 512 k or 1 M samples and 1 or 2 scans. I used a UMIK-1 and a Music Streamer II+ USB DAC with an old Win XP laptop. The results are very closely repeatable with the same settings, as expected. But using 2 scans at 512 or 1 Mpt causes some big dips in the midrange. It also caused some broadband changes of about 5 db in the distortion plots (I didn't save those plots).
Then I also tested a second setup- same DAC but everything else different: headphones, hanging on a USB webcam mic, fast desktop PC Win 7. The last plot shows the effect of default 256 kpts (only some small dips) and 512 kpts (worse) and 1 Mpt (bad dips) all with only 1 scan. I listened carefully and I'm sure it isn't actually producing these dips in the output. I'm curious if others can repeat this problem with a non-USB DAC and/ or non-USB mic. :dontknow:
Well, I made it through the intimidating signup process. :sweat:
I just started using REW 5.13 today and it's been good fun, but I did find a problem. I didn't simply use the default settings for the frequency scan, I had to crank it up to the max samples and multiple scans to see if there was a noticeable improvement, but there's actually a problem with doing this for full-range measurements. The first 5 plots were from scans I did on my DIY sub/ mid-bass "half" speaker with 512 k or 1 M samples and 1 or 2 scans. I used a UMIK-1 and a Music Streamer II+ USB DAC with an old Win XP laptop. The results are very closely repeatable with the same settings, as expected. But using 2 scans at 512 or 1 Mpt causes some big dips in the midrange. It also caused some broadband changes of about 5 db in the distortion plots (I didn't save those plots).
Then I also tested a second setup- same DAC but everything else different: headphones, hanging on a USB webcam mic, fast desktop PC Win 7. The last plot shows the effect of default 256 kpts (only some small dips) and 512 kpts (worse) and 1 Mpt (bad dips) all with only 1 scan. I listened carefully and I'm sure it isn't actually producing these dips in the output. I'm curious if others can repeat this problem with a non-USB DAC and/ or non-USB mic. :dontknow:
Attachments
1 - 1 of 3 Posts
1 - 1 of 3 Posts
