[jaakkopasanen]

How high was the voltage you used for the mics at the Behringer? There is probably a sweet spot for the primo capsules between 3-10V regarding signal to noise ratio. The other question is how constant is the build quality of the primo capsules or the Sound Professionals SP-TFB-2 ?

I'm using RODE VXLR+ adapters which provide 5 V. Can't say anything about the mic consistency and quality control. But at least the Primo's have flatter bass response than SP-TFB-2. That doesn't really matter a lot for this application though since the heapdhone compensation corrects the mic frequency response also.

 

[musicreo]

This was unexpected because specs wise the EM 258 is a better capsule.

Actually the specs are very similar.
Primo Em 258:

• Signal to Noise ratio 74dB
• Sensitivity -32 dB
• Max Input Sound Level : 115 dB S.P.L

Sound Professionals SP-TFB-2

• Signal to Noise ratio 75dB
• Sensitivity -32 dB
• Max Input Sound Level : 115 dB S.P.L

 

[jaakkopasanen]

Actually the specs are very similar.
Primo Em 258:

• Signal to Noise ratio 74dB
• Sensitivity -32 dB
• Max Input Sound Level : 115 dB S.P.L

Sound Professionals SP-TFB-2

• Signal to Noise ratio 75dB
• Sensitivity -32 dB
• Max Input Sound Level : 115 dB S.P.L

Those are specs for the more expensive master series model MS-TFB-2. I have the cheaper SP-TFB-2 which have 17 dB of more noise and 10 dB smaller sensitivity.

 

[Speedskater]

Seems like the wrong test, to measure signal-to-noise.
I would use mid-band pink noise to set the reference level.

 

[jaakkopasanen]

Seems like the wrong test, to measure signal-to-noise.
I would use mid-band pink noise to set the reference level.

Could you elaborate why is this a wrong test to measure signal to noise ratio in impulse response? Or actually this is peak to noise ratio to be more exact.

 

[Speedskater]

With pink noise or a 1 kHz tone, you can set a full volume reference level. This is the long standard protocol for measuring signal-to-noise.

 

[jaakkopasanen]

With pink noise or a 1 kHz tone, you can set a full volume reference level. This is the long standard protocol for measuring signal-to-noise.

The signal level played on speakers was the same for all measurements. The recorded signal level differences are caused by microphone sensitivity and mic pre-amp gain. The noise level of Zoom H1n changes depending on the gain and the best noise performance can be achieved at position 5.5 or 6 of the gain wheel. This is what I used for the measurements done with Zoom H1n. Additionally I don't think the recording level affects signal to noise ratio because changing the mic gain amplifies the transient signal and the noise similarly. This is in effect the same as normalizing the reconstructed impulse responses digitally.

PNR is obtained as the noise power of a normalizedRIR sample, wheren nn[k] is the noise that is observed in a 0 dB normalized RIR.

this is equation 3 from A note on the definition of signal-to-noise ratio of room impulse responses, 2012 paper by Csaba Huszty and Shinichi Sakamoto. This is precicely what I did, although I have to admit that the noise part calculation in this case was done visually so give or take a dB because of that.

EDIT: I realized that my decay plots are amplitudes and the PNR equation in that paper uses powers. So the PNR numbers from the graphs should be halved.

 

[jaakkopasanen]

I did new measurements today. This time I had speakers a bit closer together and this improved the phantom center image. Room acoustics management was used during all speaker measurements and while this was only two paramteric filters at around room mode frequencies, it had very positive impact on the results. Now there is barely any need for reverberation management to remove the bass ringing. There is still a bit left but it's very minor. I strongly suggest using room eq during measurement until the frequency dependent reverberation management is done.

This time I made sure I don't have channel balance issues in Volume2 and now the channel balancing in Impulcifer is not needed. The trend balance method does basically nothing audible, the balance is so good out of the box. I did headphone compensation measurements 4 times and each time the frequency response changes a lot above 6 kHz. Headphone compensation remains a problem. The more I read about it and study it, the more difficult it seems to do right.

Webcam was above me this time and that really made it a lot easier to place the room measurement microphone to the same location where the binaural mics where. This resulted in a lot better tonality for surround channels. I wore a hat (tube scarf actually) beacuse otherwise my hair hides the pinna making it impossible to locate the ear canal entrance from the webcam ghost images. Still the pictures don't quite show the location of the ear canal entrance but I think I got it estimated well enough. You'll see in the picture that the room measurement mic is placed a bit towards the center of the head from the left ear pinna. There are two pictures in the side panel for center measurements (pics 2 and 3) because I wanted to test FC measured with FL vs FR. Pictures 4 and 5 are for surround channels.

I think all in all this is my favorite so far and therefore I updated the demo recordings to the new one.

 

[musicreo]

Great to hear that you still can optimise your results.
When i look at the image I see that you not only rotate your head but also change the head position for the surround measurements (I draw the line in the image to show what I mean). In some of my measurements I'm pretty sure that it got worse when my head moved out of the sweetspot.

 

[castleofargh]

I did new measurements today. This time I had speakers a bit closer together and this improved the phantom center image. Room acoustics management was used during all speaker measurements and while this was only two paramteric filters at around room mode frequencies, it had very positive impact on the results. Now there is barely any need for reverberation management to remove the bass ringing. There is still a bit left but it's very minor. I strongly suggest using room eq during measurement until the frequency dependent reverberation management is done.

This time I made sure I don't have channel balance issues in Volume2 and now the channel balancing in Impulcifer is not needed. The trend balance method does basically nothing audible, the balance is so good out of the box. I did headphone compensation measurements 4 times and each time the frequency response changes a lot above 6 kHz. Headphone compensation remains a problem. The more I read about it and study it, the more difficult it seems to do right.



Webcam was above me this time and that really made it a lot easier to place the room measurement microphone to the same location where the binaural mics where. This resulted in a lot better tonality for surround channels. I wore a hat (tube scarf actually) beacuse otherwise my hair hides the pinna making it impossible to locate the ear canal entrance from the webcam ghost images. Still the pictures don't quite show the location of the ear canal entrance but I think I got it estimated well enough. You'll see in the picture that the room measurement mic is placed a bit towards the center of the head from the left ear pinna. There are two pictures in the side panel for center measurements (pics 2 and 3) because I wanted to test FC measured with FL vs FR. Pictures 4 and 5 are for surround channels.



I think all in all this is my favorite so far and therefore I updated the demo recordings to the new one.

For that kind of test, I'd suggest to use a much simpler chair for the measurement. Anything with a very small acoustic footprint.
And I'd also advise to completely turn your body, making sure that your shoulders are always aligned with your head, so the impact of the torso is the right one.

 

[jaakkopasanen]

Great to hear that you still can optimise your results.
When i look at the image I see that you not only rotate your head but also change the head position for the surround measurements (I draw the line in the image to show what I mean). In some of my measurements I'm pretty sure that it got worse when my head moved out of the sweetspot.

It's very hard to keep the center of the head in the same position between the measurements. I'm looking directly to TV when doing the FL, FR measurement so when I'm doing surround channels, I don't see where my head is relative to the first measurement. I did separate room measurements for all of the positions, left and right ear separately. This ensures that the frequency response ends up being the same even if I move the head out of the sweet spot. Or at least in theory that is...

For that kind of test, I'd suggest to use a much simpler chair for the measurement. Anything with a very small acoustic footprint.
And I'd also advise to completely turn your body, making sure that your shoulders are always aligned with your head, so the impact of the torso is the right one.

I'm not sure I quite understand what you mean by "that kind of test" or "simpler chain for the measurement". Could you elaborate a bit?

You are right with your second point. Unfortunately that's not really possible on sofa and this time I didn't want to reorganize my room. I intend to have a measurement session where I optimize the speaker placement and all the furniture for best possible acoustics but I haven't got there yet. That would of course also include full rotation of the body.

 

[Joe Bloggs]

It's very hard to keep the center of the head in the same position between the measurements. I'm looking directly to TV when doing the FL, FR measurement so when I'm doing surround channels, I don't see where my head is relative to the first measurement. I did separate room measurements for all of the positions, left and right ear separately. This ensures that the frequency response ends up being the same even if I move the head out of the sweet spot. Or at least in theory that is...


I'm not sure I quite understand what you mean by "that kind of test" or "simpler chain for the measurement". Could you elaborate a bit?

You are right with your second point. Unfortunately that's not really possible on sofa and this time I didn't want to reorganize my room. I intend to have a measurement session where I optimize the speaker placement and all the furniture for best possible acoustics but I haven't got there yet. That would of course also include full rotation of the body.

Simpler "chair"

 

[jaakkopasanen]

Simpler "chair"

I know, but the sofa just happens to be there in the listening position.

 

[castleofargh]

It's very hard to keep the center of the head in the same position between the measurements. I'm looking directly to TV when doing the FL, FR measurement so when I'm doing surround channels, I don't see where my head is relative to the first measurement. I did separate room measurements for all of the positions, left and right ear separately. This ensures that the frequency response ends up being the same even if I move the head out of the sweet spot. Or at least in theory that is...


I'm not sure I quite understand what you mean by "that kind of test" or "simpler chain for the measurement". Could you elaborate a bit?

You are right with your second point. Unfortunately that's not really possible on sofa and this time I didn't want to reorganize my room. I intend to have a measurement session where I optimize the speaker placement and all the furniture for best possible acoustics but I haven't got there yet. That would of course also include full rotation of the body.

I only mentioned this in case it had skipped your mind. If you don't bother with something by choice, then all is well

 

[johnn29]

Re-arranging your room to get Impulcifer recordings done is something that I've got a lot of experience in! I can barely get the UMIK placement right with only two positions since I use my surround system. It is a right pain though. I never thought about buying a property articulating mic stand - I only use a monopod and sit on the floor. It would be much more comfortable to sit in a chair and use an articulating mic.

Re: responses above 6Khz - will that also apply to something like the Sony XM3 that has a "calibrate" button. Once you put your headphone on, it's supposed to measure the wearing condition (glasses, hair, placement etc.) to get an identical FR each time you reseat. I'm not sure how how it goes to though, or if it only cares about bass regions.

Re: virtual center is what you get best from Impulcifer compared to real life. Always bang in the sweet spot. But like in real life, I prefer applying Dolby Prologic to stereo signals via ffdshow audio with Impulcifer.