Damn you Bennett!

Ivan Beaver

Ivan Beaver

Graduate Student
Jan 11, 2011
2,300
10
0
Atlanta GA area
OK Bennet-now you have gone and messed things up.



Why do you make me go and THINK! and question things. Everything was nice and happy until you just had to go and compare DSP's now didn't you!



And you think you want to get data on the differences between DSP's. Well how about the differences between measurement platforms-HUH What about THAT for a project
icon_lol.gif
icon_lol.gif




Kemper Watson is bringing over a couple of his DBX DSP's for me to measure and give you the data.



So to get ready I setup my Smaart rig and did something stupid. I did a loopback test. Yeah I know-why do something stupid like that
icon_rolleyes.gif




I am using a Motu8pre as my interface.



I have done loopback tests with this unit before on Systune and all was fine, but never bothered to do it with Smaart V7. I just ''assumed'' it was correct. You know what happens when you do that now don't you.



So anyway I am getting an odd freq response. It is flat up to about 2KHz and then starts a rise of 1.5dB at 18Khz.



So my first thought is that the interface is not flat
icon_mad.gif
. So I did a loopback test using Systune-without touching anything on the interface or setup and it is flat to 20Khz.
icon_confused.gif




I went back to Smaart and it was still off.



But that is not all. Systunes phase response is around 2° with a 30° rise at 18Khz.



Smaart V7s phase response is around -178° with a 30° rise at 18Khz.



So the shape is the same-but they are out of polarity.
icon_eek.gif




As the old saying goes- a Man with a meter is sure-A man with 2 meters is sure of nothing.



So what is correct?
icon_confused.gif




Yes I have already posted a question to Rational.



So realize that the curves I will be sending you will have a slight error on the top end-that is not a result of the DSP settings.



So take that data and - well do whatever you want with it
icon_lol.gif
icon_lol.gif




Just adding to the confusion.



Why does this stuff keep getting harder and more mysterious? Is anything real? Where's my RTA and I will call it a day.
icon_lol.gif
 
Re: Damn you Bennett!

I've noticed a couple things using SMAART with a USBPre.



One is that the USBPre (the older version) outputs tend to act a little funny when loaded down. Rational recommends using a Y-cable for the reference signal, so that if the output gets loaded funny and changes response it won't matter much to the measurement.



The other is that SMAART doesn't have very fine delay resolution compared to high frequency phase. So when I measure a digital device the phase trace often goes well away from 0 degrees at high frequencies, because there isn't a delay compensation setting that exactly matches the test device latency.
 
Re: Damn you Bennett!

Scott Helmke wrote on Sat said:
I've noticed a couple things using SMAART with a USBPre.



One is that the USBPre (the older version) outputs tend to act a little funny when loaded down. Rational recommends using a Y-cable for the reference signal, so that if the output gets loaded funny and changes response it won't matter much to the measurement.



The other is that SMAART doesn't have very fine delay resolution compared to high frequency phase. So when I measure a digital device the phase trace often goes well away from 0 degrees at high frequencies, because there isn't a delay compensation setting that exactly matches the test device latency.
Click to expand...



It's because the minimum delay increment is related to the sample rate. You can't get any smaller than one sample.
 
Re: Damn you Bennett!

i have notice a similar rising HF phenomonen.

i use a small mixer for routing and phantom,to a 2 chan usb line in pre.

i had both output chans of the pc fed into a stereo input on the mixer,and used auxes to return back to the pc.

i found a similar rising hf response until i unplugged the Right leg of the stereo input.it seems somehow that the summation of the l/r signals to the aux induced the rising response.

both channels tested flat individualy.

maybe i am missing something simple here,so if anyone can enlighten me, i would be greatful



cheers,,dave
 
Re: Damn you Bennett!

David Stiles wrote on Sat said:
scott,,i have notice a similar phenomonen.

i use a small mixer for routing and phantom,to a 2 chan usb line in pre.

i had both output chans of the pc fed into a stereo input on the mixer,and used auxes to return back to the pc.

i found a similar rising hf response until i unplugged the Right leg of the stereo input.it seems somehow that the summation of the l/r signals to the aux induced the rising response.

both channels tested flat individualy.

maybe i am missing something simple here,so if anyone can enlighten me, i would be greatful



cheers,,dave
Click to expand...



Some cheap sound sources may use a single DA converter to generate both outputs and alternates between Left and Right. This could make a mono signal have a time offset between left and right channels. I can't immediately see how this would cause a rising trend. It seems more like a falling trend would result as the equivalent phase shift created by the fixed time offset becomes more significant at higher frequency.



Combining identical sine waves in phase is +6dB, +/- 90' (total) is only +3dB (or 3dB less than in phase), and +/-180' is a null. So I would expect a gradual roll-off as effective phase shift increases.



Perhaps it is something else?



JR
 
Top Bottom