Frequency response cliff after 10K

**TheaterFan** · **09-20-08, 08:24 PM**

Guys,

I am attempting full range EQ of my audio system. Each side has a sub with two drivers and a planar high end. In terms of the high end, I am getting the graph below, with the response falling off the table after 10K. Thinking my drivers were bad, I swapped in a pair of excellent monitors I have which have a very extended high freqency response. I got a very similar radical drop off graph.

I have worked two days on this now, and ready to admit I need help. What am I doing wrong or what do you guys think is happening? The microphone is linear, and the sound card calibration appeared quite good also.

Thanks.

**brucek** · **09-20-08, 08:34 PM**

Quote:

The microphone is linear

Can you explain this statement and the type of microphone you're using?

I assume you did a full range measurement of a loopback cable to ensure it is completely flat?

brucek

**Wayne A. Pflughaupt** · **09-20-08, 08:49 PM**

Quote:

TheaterFan wrote:

The microphone is linear...

Do you have a calibration file for the mic?

Regards,
Wayne

**TheaterFan** · **09-20-08, 08:56 PM**

Yes. It is a DBX RTA-M, and I installed the calibration file. It is shown above in the trace. It is up like

I also saw very similar drop off with my RS digital meter, so it looks not to be the mic. The mic is the solid line and the sound card calibration the dotted line below.

**TheaterFan** · **09-20-08, 09:42 PM**

Bruce,

I think the sound card is hitting the wall at 10K. I am going to try and recalibrate it and see what happens.

**brucek** · **09-20-08, 11:16 PM**

Quote:

I am going to try and recalibrate it and see what happens

Again, as I mentioned above, after the calibration and saving of the soundcard cal file, you need to take a measurement of the loopback cable (before removing it) to prove that the file is valid by measuring a flat response for the cable measurement. (be sure the meter cal is cleared for that measurement).

brucek

**Anthony** · **09-23-08, 10:21 AM**

Another member here -- BoomieMCT had similar problems when he used REW with his wife's mac laptop. Apparently that laptop's soundcard had a nice recording cliff at 10k as well, which is why he makes me bring my laptop over whenever he EQ's his subs

You might be experiencing a similar problem. He never found a solution to it -- just seemed to be a problem with the laptop soundcard.

**brucek** · **09-23-08, 11:02 AM**

Quote:

You might be experiencing a similar problem.

His soundcard calibration file would say otherwise. It extends quite nicely to 20KHz in the graph...

brucek

**Anthony** · **09-23-08, 11:07 AM**

Sorry, for some reason that second graphic didn't load when I read this post earlier.

And your soundcard calibration was the same ports you used to wire in the mic / mic-preamp? Normally that's line-in to line-out, but if you use the mic-in for the measurements, but calibrate on line-in, that might explain the discrepancy. Something to double-check anyways.

If not, are you using a mic preamp? Is it in the loop when you calibrated (if that's even possible).

**brucek** · **09-23-08, 11:26 AM**

Quote:

Is it in the loop when you calibrated (if that's even possible).

Not the accepted practice. The soundcard is a line level device, not really designed to plug directly into a mic input (low level). The small response difference in the mic stage of a line level mixer is considered insignificant when creating a soundcard cal.

brucek

**Anthony** · **09-23-08, 11:31 AM**

I didn't think it was possible, as I went through that trying to calibrate my M-Audio mobile pre (where the mic preamp is built into the sound card and can't be separated).

I'm just wondering if that could be a source of error in the chain (as could anything else that is in the signal path, but not in the calibration path).

**brucek** · **09-23-08, 11:55 AM**

Quote:

I'm just wondering if that could be a source of error in the chain

Yeah perhaps, but it could only be easily checked by substitution of another unit, I suppose.

Generally the mic stage of a preamp/mixer has very good upper frequency response (i.e. Xenyx 802 = 10Hz-150KHz /-1dB). The usual reasoning being including the preamp in the soundcard cal is to offset low frequency dropoff of the line stages (i.e for sub use) and to check that the EQ stages aren't wacky (like the 802 sometimes is) and to flush out any faults.

brucek

Advertisement