Calibration Of Soundcard

Bryan Rason

New Member
Thread Starter
Joined
Oct 3, 2020
Posts
4
Hello Everyone,

I calibrated my sound card but the high end looks odd.
Why wouldn’t it be a flat line?

As you can see it’s affecting the room measurements in the opposite way.

Plus phasing look weird.
 

Attachments

  • 15D2323D-DE14-44DC-9E65-BE9763F962CA.jpeg
    15D2323D-DE14-44DC-9E65-BE9763F962CA.jpeg
    337.1 KB · Views: 139
  • B8A7B92F-BFA0-4C19-9638-EDBF635B29A0.jpeg
    B8A7B92F-BFA0-4C19-9638-EDBF635B29A0.jpeg
    255.5 KB · Views: 153
  • 31BF90DF-366D-4C0F-852A-C78AA29E4CEF.jpeg
    31BF90DF-366D-4C0F-852A-C78AA29E4CEF.jpeg
    264.1 KB · Views: 142
What soundcard? Some soundcards have weird filters, but actually I would suspect cross-talk between inputs and outputs. This could be due to a ground loop or poor grounding due to oxydation.
 
I calibrated my sound card but the high end looks odd.
Please detail device, connection and settings.

While DanDan is correct that calibration is not required, it does not hurt to have a look-see.

35793
 
This graph looks like a THD+N vs. frequency measurement, which does not strictly require calibration, just the same as a frequency reponse graph does not require calibration. B;ut here the OP shows an issue with frequency response; considering the amplitude of the problem, I think he is right in assuming something is wrong. I would.
 
Soundcard calibrations typically show up with a +/- 3 dB scale, so I’m not sure what that is.

Regards,
Wayne
 
Measuring or Calibrating the Soundcard is intrinsically close to creating a feedback loop. Some have a 'direct monitor' path to eliminate latency for music recording purposes. If the DM and Thru the system paths are both active one will get serious comb filtering or even echo. The OP graph has a vertical scale of +/- hundreds of dB. Walk away, there is no need for SoundCard Calibration.
 
Can you elaborate on that? Particularly without knowing what soundcard it is?
It's a Focusrite Solo Gen 3.
The earlier graph is a loopback showing level and distortion as % of full level.

This graph is level with Y axis ±1dB.
35813


IMO, it's worth while to perform a loopback measurement at minimum to ensure there is nothing weird like bad jack, cable, hum, etc.
 
Fair enough Ian, but I disagree. Most people who run into trouble with SoundCard Calibration are new to REW. The process is quite prone to problems with feedback and other errors as we can see.
But to your point, SC Calibraion does not test the actual path in use by the mic so overall a bit of a red herring IMO. For those who need Loopback Correction for critical timing...... well, they are likely to be experienced users and likely to not run into newby red herrings.
 
Last edited:
At the risk of getting ousted, let me say your first answer is unworthy of a moderator.
Your second answer is much better since it proposes some possible explanation, but it came only after you've been challenged.
I believe there is nothing wrong in assessing the intrinsic performance of test equipment.
50 years of pro audio design including several years with Audio Precision allow me to affirm it.
 
I must confess to about 50 years in audio, 25 as a professional, before I ever used REW.
Equipment verification before use is SOP.
[@JLM1948 - LOVED my AP Portable c. 1993]
 
Best practice is per se performed by professionals. In Noise Control, for legal/paranoid reasons we Calibrate our SLMs before and after a test procedure. This Lab standard and environment proof equipment is very robust and extremely reliable. I have never seen such a Calibration fail or detect a problem of any kind. But best practice and fear of legal challenge......

A Sound Card is not really test equipment, nor are the connectors and cables of that type of medical/military standard.
So it might seem wise to test and verify the REW signal chain. Indeed I recommend it. Play back music and or test sweeps on the actual play back channels. If anything is wrong it will be audible. Listen to the actual input channel on headphones, and look for credible levels via meters. Tap the microphone to check if there is a feedback loop to the speakers. But testing the other unused channel, omitting the mic inpu, t strikes me as a bit obtuse. More importantly it simply does not test or verify the actual chain in use. It is a also very frequent case of confusion and error (IMO distraction) for REW beginners.
I believe the most helpful advice regarding SoundCard Calibration is to state that it is unnecessary. I may sometime add that anyone who has graduated to quite complex REW work, actually needing Loopback Correction, they would hardly need advice.
 
Last edited:
Hello Everyone,
I redid everything this week without the calibration. I clearly didn’t calibrate the soundcard right but the rest of it turned out well.
 
Best practice is per se performed by professionals. In Noise Control, for legal/paranoid reasons we Calibrate our SLMs before and after a test procedure. This Lab standard and environment proof equipment is very robust and extremely reliable. I have never seen such a Calibration fail or detect a problem of any kind. But best practice and fear of legal challenge......

A Sound Card is not really test equipment, nor are the connectors and cables of that type of medical/military standard.
So it might seem wise to test and verify the REW signal chain. Indeed I recommend it. Play back music and or test sweeps on the actual play back channels. If anything is wrong it will be audible. Listen to the actual input channel on headphones, and look for credible levels via meters. Tap the microphone to check if there is a feedback loop to the speakers. But testing the other unused channel, omitting the mic input strikes me as a bit obtuse. More importantly it simply does not test or verify the actual chain we want to use. It is a also very frequent case of confusion and error (IMO distraction) for REW beginners.
I believe the most helpful advice regarding SoundCard Calibration is to state that it is unnecessary. I may sometime add that anyone who has graduated to quite complex REW work, actually needing Loopback Correction, they would hardly need advice.
Definitely a newbie to REW anyway.
I’m just going to Not calibrate.
I’ve already got my room set up.
Again thank you for your time.
 
Last edited by a moderator:
At the risk of getting ousted, let me say your first answer is unworthy of a moderator.
Your second answer is much better since it proposes some possible explanation, but it came only after you've been challenged.
I believe there is nothing wrong in assessing the intrinsic performance of test equipment.
50 years of pro audio design including several years with Audio Precision allow me to affirm it.
Then why say anything ‍♂️
 
Back
Top