Room Measurement Tutorial for Dummies Part 2

[amirm]

So, what's the verdict from you who are using Dirac Live?

It works and is simple to use. The drawback is the cost. Stereo version for PC costs $600 from what I recall.

 

[RayDunzl]

why didn't you limit that range in the "Range" tab (lower right corner of your graph)?

That controls the zoom of the display, not the filter calculation.

 

[Krunok]

It works and is simple to use. The drawback is the cost. Stereo version for PC costs $600 from what I recall.

Ok, good to hear. I was anyhow planning to purchase Dirac Live upgrade for miniDSP 2x4HD and that one costs $200, which seems acceptable as you said it works well.

 

[Krunok]

That controls the zoom of the display, not the filter calculation.

Aha. And you can't limit the filter calculation via settings? In that case you can maybe delete the samples you don't need from the measurement file.

 

[RayDunzl]

And you can't limit the filter calculation via settings? In that case you can maybe delete the samples you don't need from the measurement file.

Sorry, I have reached (maybe already overreached) the limit of my knowledge of this subject.

---

Side note: One reason I didn't buy Dirac for the OpenDRC-DI because it prevents experimentation with other tools like rePhase due to the different firmware that gets installed in the device. I don't know the details.

 

[Krunok]

Sorry, I have reached (maybe already overreached) the limit of my knowledge of this subject.

---

Side note: One reason I didn't buy Dirac for the OpenDRC-DI because it prevents experimentation with other tools like rePhase due to the different firmware that gets installed in the device. I don't know the details.

I would expect such software to have feature like that, it's probably there but not on a very intuitive spot.

Were you able to solve that curve with the AcourateDRC?

 

[Krunok]

@RayDunzl, Looking at the rePhase again, but this time with my first glass and a cigar in hand, I'm thinking you could try to enter the values directly in those "Hz" fields under the sliders. As you want to solve the issue that lies between 45 and 55 Hz you may try to enter values 42..58 under those 17 slides and see what happens. You may as well try to play with their Q values as well. I'm pretty sure 4096 taps should be enough to do it, even at 96kHz.

 

[RayDunzl]

Were you able to solve that curve with the AcourateDRC?

It doesn't solve that phase issue, but does a good job otherwise.

I haven't tried convolving the Acourate Filter with a filter with a manual adjustment, to see what happens.

 

[Krunok]

I haven't tried convolving the Acourate Filter with a filter with a manual adjustment, to see what happens.

I was referring to manually adjusting with the rePhase, not Acourate.

 

[RayDunzl]

I'm pretty sure 4096 taps should be enough to do it, even at 96kHz.

The 4096 taps is the total available, allocable per channel on the 2x4HD.

I would want to spread them over 4 channels, so 1024 each, leading to the rePhase display above which shows little effect available at the 50hz area.

 

[pos]

I you are doing phase corrections only you can try setting windowing to "rectangular" and centering to "energy" and see if it helps.

 

[RayDunzl]

I you are doing phase corrections only you can try setting windowing to "rectangular" and centering to "energy" and see if it helps.

Blue = request, Red = result

For the 2x4HD, phase only:

 

[pos]

I was talking about this situation you were describing here, supposedly with an openDRC:

Back in the real world, I want to twist the phase in the area of 45 to 55Hz +90 degrees in one channel, and -90 degrees in the other, as an experiment to see what happens to my deep hole in that area.

Tweaking windowing algorithm and centering is not a silver bullet that will solve situations requiring many more taps like the one in the screenshot above, but it can help in real world scenarios when you are simply a bit short on taps. "optimization" option can also be used to reduce ill effects on the magnitude response.

 

[Krunok]

Blue = request, Red = result

For the 2x4HD, phase only:

View attachment 15928

I suggest you try to use the rest of the 13 sliders. You may try to enter freqeuncies manually, for example setting 10 of them around 50Hz and 3 around 300Hz to add weight to the points that need more correction.

 

[Krunok]

Guys, why is this phase graph so odd? This was taken with my notebook mic when the notebook was practically between the speakers. Is it normal for such lousy measurement?



It's got nothing to do with rePhase as it looks the same in REW with which I took the measurement:



So, what am I doing wrong here?

 

[pos]

You need to measure one speaker at a time to avoid comb filtering effects between the two correlated sources.

 

[Krunok]

You need to measure one speaker at a time to avoid comb filtering effects between the two correlated sources.

Ok, I'll give it a try.

 

[Krunok]

That was right speaker only.

I measured left speaker with notebooks mike as close as I could get it (app 1m) and here is what I got (var smoothing):

 

[amirm]

Guys, why is this phase graph so odd? This was taken with my notebook mic when the notebook was practically between the speakers. Is it normal for such lousy measurement?

That sudden drop is not normal unless that is what you have inserted in the audio path.

 

[pos]

Krunok, what you show here looks like the materialization of a pure delay that has not been compensated for.
If you set the frequency scale to linear instead of logarithm you should see a strait line up high (ie constant group delay).

In that case you need to set the temporal offset of the impulse so that its t=0 is at (or slightly before) the main peak.

You can also adjust the time offset directly in rephase, in the measurement tab, but having a visual display of the impulse helps there.
I suppose REW can also auto adjust this offset for you.