Tech Tuesday: Deep Dive into Room Correction

[Maurizio]

Just got a updated WiiM Home App. In the release notes I see this!

"External Microphone Calibration File: Import calibration files for improved Room Correction accuracy."

But no further instructions seems to be available.View attachment 17324

You must do as @slartibartfast says. Connect a microphone and you will have the option available.
I just did a try with my IMM 6C and now I have the proper calibration file imported and, I hope, working in the right way This on Android, still no update for IOS App

 

[hgo58]

You must do as @slartibartfast says. Connect a microphone and you will have the option available.
I just did a try with my IMM 6C and now I have the proper calibration file imported and, I hope, working in the right way This on Android, still no update for IOS App

Yes. This was only to point out that it's available but no documentation has been released yet.

 

[Maurizio]

Yes. This was only to point out that it's available but no documentation has been released yet.

Mhhhh do you think that some documentation will be released? I’m not so confident

 

[canard]

the possibility of cal mic has just appeared on the very latest wiim home under android.... (on the other hand the multi chirp function still gets stuck, bug .....)
to be tested....


(on the other hand still no range differentiating the corrections gain into negative or positive...see a max "q" definable in ls/hs mode)

 

[hgo58]

Mhhhh do you think that some documentation will be released? I’m not so confident

I would expect that. The file formats needs to be documented as a minimum.

 

[slartibartfast]

I would expect that. The file formats needs to be documented as a minimum.

I am sure for all the common mics the calibration file is used as it comes from the manufacturer.

 

[slartibartfast]

the possibility of cal mic has just appeared on the very latest wiim home under android.... (on the other hand the multi chirp function still gets stuck, bug .....)
to be tested....


(on the other hand still no range differentiating the corrections gain into negative or positive...see a max "q" definable in ls/hs mode)

In practice the algorithm seems to use much reduced positive gain anyway. It doesn't appear to try to fill dips completely.

 

[hgo58]

I am sure for all the common mics the calibration file is used as it comes from the manufacturer.

Maybe?

I do however see documentation as an important part of a software release snd functionality. Maybe an occupational hazard?

I know most people don't read it anyway

 

[canard]

In practice the algorithm seems to use much reduced positive gain anyway. It doesn't appear to try to fill dips completely.

It's not very explicit ;-)
(a limitation of the "q" in ls/hs to 0.7 would be welcome in rc mode)

 

[slartibartfast]

That's exactly what I think is one core problem at the moment and the reason that we see things like e.g. this:

View attachment 17318

A +4 dB "Peak" filter with a Q of just 0.1! What this really does is simply boosting almost the entire frequency range chosen for correction (33 Hz to 3300 Hz in this case). Let's watch this isolated from the other filters:

View attachment 17319

This is from a stereo "calibration". I'm pretty sure that an incorrectly determined average SPL is the root cause. In particular it looks like the typical silly peak between 8 kHz and 16 kHz found with most smartphone measurements might have been taken into account, although it is far outside the correction range.

I still have to think about how independent L/R "calibration" can be affected under different conditions. How likely is a full range channel imbalance of e.g. 2 dB? I agree that this should be taken care of first, but how should automaiRC approach this? A sort of imbalance over a wide frequency range (but not necessarily the full range) is probably more likely.

I compared a room correction from WiiM with one from REW using the B&K target. Blue is REW, Red is WiiM. You can see the WiiM RC has filled the dips slightly but only by 2-3dB. The low frequencies are noticeably lower in the WiiM curve indicating that balance between low and high frequencies isn't quite right. I used a range of 20-700Hz. Not bad though.

 

[harkpabst]

I compared a room correction from WiiM with one from REW using the B&K target. Blue is REW, Red is WiiM. You can see the WiiM RC has filled the dips slightly but only by 2-3dB. The low frequencies are noticeably lower in the WiiM curve indicating that balance between low and high frequencies isn't quite right. I used a range of 20-700Hz. Not bad though.

View attachment 17353

Not bad, indeed.

Are there actually positive gain peak filters right at the dip frequencies ~160 Hz and ~230 Hz? If so, what is the gain setting? Even if it is possible to partially fill theses dips, that would be undesirable, if the gain is any higher than what was achieved here.

 

[slartibartfast]

Not bad, indeed.

Are there actually positive gain peak filters right at the dip frequencies ~160 Hz and ~230 Hz? If so, what is the gain setting? Even if it is possible to partially fill theses dips, that would be undesirable, if the gain is any higher than what was achieved here.

 

[harkpabst]

View attachment 17354

I would conclude that filters 7 and 8 are mostly wasted, then.

Adding 4.4 and 3.0 dB of gain to shove a portion of that value from an 18 dB and 10 dB dip doesn't look like amplifier power well spent to me. Not as bad as had feared, though.

 

[Bismarck]

I compared a room correction from WiiM with one from REW using the B&K target. Blue is REW, Red is WiiM. You can see the WiiM RC has filled the dips slightly but only by 2-3dB. The low frequencies are noticeably lower in the WiiM curve indicating that balance between low and high frequencies isn't quite right. I used a range of 20-700Hz. Not bad though.

Which microphone?
With calibration?

 

[slartibartfast]

Which microphone?
With calibration?

UMIK-1 with calibration.

 

[Wiimer]

UMIK-1 with calibration.

Has the WiiM microphone calibration been reflected correctly, is everything OK?
Thanks

 

[slartibartfast]

Has the WiiM microphone calibration been reflected correctly, is everything OK?
Thanks

Probably but how could I be sure?

 

[Wiimer]

Probably but how could I be sure?

Would you like to compare the results of the measurements before and after calibration?

I will try later if have time.

 

[canard]

Would you like to compare the results of the measurements before and after calibration?

I will try later if have time.

observe read the values of your cal... if as very probably an impact at the top of the bandwidth... observe by measuring just between like 8k and 20khz with or without the cal correction of the microphone...
you will have more chance to see if an impact or not of this file...
(The irony is that correcting room issues is generally the most useless in this fr ..but where it eventually allows us to observe possible impacts of this cal as well...logical)
but it would probably require a display much more precise than what Wiim Home displays I fear...

 

[slartibartfast]

but it would probably require a display much more precise than what Wiim Home displays I fear...

Exactly. It would be easy if the results could be exported then imported into REW.