Home Theater Forum and Systems banner

Sound card calibration: stupid question

4K views 9 replies 4 participants last post by  AudiocRaver 
#1 ·
In reading about sound card calibration I see that we're supposed to loop the output to the input. The question is, is REW calibrating the INPUT or the OUTPUT? Furthermore, how can it possibly calibrate either based on this setup? If the output is garbage then there's no way to use that signal to calibrate the input, and vice versa.

I'm planning on using REW and the Dayton iMM-6 mic to tune my car's audio using a laptop or tablet. I'd have to use "mic" jack for the input and "headphone" for the output. Feel free to stop me here if I'm already going down the wrong path. :)
 
#2 ·
Nope, that seems right. I have the exact setup for the same use except I have a single-jack laptop that is just like the mic/smartphones so I had to get a "Y" adapter for the loopback. I think you will need the exact opposite "Y" adapter to combine output and line in for using the imm6. Those are cheap on Amazon if you didn't already get one. If your car system has an Aux input, that works out really well. In theory you normalize your laptop system, then the only difference will be your intervening car system so everything will show up, warts and all. One of the best ways to visualize this will be the spectrogram output. Look at your calibration result and it will be pretty smooth, but not absolutely perfect.

Although I am not sure exactly how it is implemented in REW, calibration is possible with an iterative process. By garbage sound I guess you mean that a signal at a given frequency is not actually that frequency or associated with a lot of noise or harmonics. That would be minimized by not overdoing the mic boost or sensitivity settings in the windows driver. If you look at the process in REW, you will see that the first step is determining what the maximum level output would be (the -12 to -6db level setting step) and there is also a check if overall levels are too low to be accurate or are too noisy. Between there is a happy medium that the software seems to be able to use to calibrate. Admittedly it is not ideal relative to an external setup, but it seems very functional. It is a bit different than the process in TueRTA, but I have used both and get indistinguishable results.
Newer sound cards, even though they exist in a noisy motherboard environment, are not that bad actually. If you have a reasonably new laptop, I wouldn't be too concerned.

In the end you can always check against an external sound source, or hook up to trusty speakers and see what you get. but that can also introduce it's own trouble.
 
#3 ·
In reading about sound card calibration I see that we're supposed to loop the output to the input. The question is, is REW calibrating the INPUT or the OUTPUT? Furthermore, how can it possibly calibrate either based on this setup? If the output is garbage then there's no way to use that signal to calibrate the input, and vice versa.
REW sends a flat signal to the input and analyzes what comes from the output. So it doesn’t matter which one is deficient. Not that it matters in your case, because...


I'm planning on using REW and the Dayton iMM-6 mic to tune my car's audio using a laptop or tablet. I'd have to use "mic" jack for the input and "headphone" for the output. Feel free to stop me here if I'm already going down the wrong path. :)
Happy to help. The mic requires phantom power, so you won’t be able to use the laptop’s input. You’ll need an outboard USB audio interface – plenty of options listed in the REW Soundcard Database. The good news is that such devices have flat response so a calibration is not needed. :T

Regards,
Wayne
 
#4 ·
#6 · (Edited)
#7 ·
Thanks for the advise everybody. I'm excited to see this all in action for the first time. I know that if it makes my wife roll her eyes that it's going to be AWESOME.

Since the iMM6 is a little new to the process I'll try to remember to share my process and results with the group.
 
#8 ·
In reading about sound card calibration I see that we're supposed to loop the output to the input. The question is, is REW calibrating the INPUT or the OUTPUT? Furthermore, how can it possibly calibrate either based on this setup? If the output is garbage then there's no way to use that signal to calibrate the input, and vice versa.
Very few even think to ask that question. It is a bit puzzling how it can work.

The calibration process works because REW measures a closed loop. The calibration process removes the variables - amp, speakers, room - from the loop and compensates for any remaining variations - input & output of audio interface are the main ones, but the calibration process can include other elements it you have reason to. REW does not know where those variations are, only that they are somewhere in the loop.

When you make a regular measurement, REW then simply tells you what in its loop is different from what was in the calibration loop.
 
#9 ·
AH! I get it!

So there's an important point that should be made, being that one needs to use REW to generate all the test tones through the output of the PC in order for the calibration to be valid and useful. I was planning on using a pink noise CD to do my testing but apparently that would bypass results of the calibration (which is what confused me in the first place).
 
#10 ·
For absolute highest accuracy, that is correct. You will find cases, however, where someone will break that rule when using Pro or high quality gear that they are confident they can trust. For instance, using an external pink noise source, as you suggest, or some other signal source, can be done with accuracy if you are confident that the output frequency response of your audio interface and the output frequency response of the device providing that signal are both very flat within the critical range of your measurements. But generally speaking, it is best to stick with the loop that has been calibrated.

When using a USB mic with REW, it is not possible to run this kind of calibration properly, since the input stage of the sound card or audio interface will not be in the loop when the USB mic is in use. We generally recommended not doing the calibration in that case, assuming that a line output is likely to have flatter frequency response than a mic input , and you will be better off without the calibration compensation when the USB mic is in use. Of course, the mic should always have its own calibration curve.
 
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