advertisement


Speaker/Room Measurement Witchcraftery

My conclusions:
- Overall I think It's quite hard to get meaningful measurements and they can be tricky to interpret.
- REW is very impressive, especially for the asking price!
- The ear is the final arbiter.
Interesting thread. Meaningful measurements are cheap and easy to get these days compared to earlier times but you won't get meaningful measurements if you don't make the effort to understand what the measurements represent and how to vary the gating, averaging, signal, etc... in order to measure what you want to know. In order to have a precise set of things you want to know (rather than the currently vague graphs of something or other that might be vaguely related to what one might hear) you first need to understand a bit about how speakers radiate sound, how speakers interact with the room and how we perceive sound.

When I first took in an interest in home audio in the early 70s the majority of hi-fi enthusiasts seemed to have this interest in "why" but it all but disappeared with the rapid growth of the audiophile thing in the late 70s that promoted a belief in magic (both wizardly and witcherly). Now the audiophile thing doesn't seem to have quite such a strong hold as it did perhaps 10 years ago it will be interesting to see if typical posters here start to take more of an interest in "why" rather than "what" and, for example, chip away at the gating, microphone positioning, averaging,... until you do come up with a more meaningful set of measurements to represent what the LS50 is doing in your room.
 
I'm very interested in understanding the why and how. But it seems that discussing it on a forum doesn't help with that (nothwithstanding the excellent help from those that guided me on how to use REW more effectively).
What this exercise has done is convince me even more that putting an accurate signal into a decent-sounding room is the right way to go. It's not difficult to get much better looking measurements, but I chose to post plots of my speaker in the position they have been used for any years and as heard by a couple of members here. I thought that would be more useful to the discussion. I also suspect I have been rather too modest and circumspect in my conclusion as well, which might give the impression that I don't know what I'm doing.

I have tried the gating function on REW and it did not give results that I thought were right. I decided not to post these. I will investigate this further with a pal who absolutely does know what he's doing (the inventor of the ZRB concept).

h.g., maybe you could offer some useful guidance on how to "measure what you want to know".
 
In light of h.g.'s comments I think I will revise my conclusion from:
My conclusions:
- Overall I think It's quite hard to get meaningful measurements and they can be tricky to interpret.
- REW is very impressive, especially for the asking price!
- The ear is the final arbiter.

To:
- Overall I think It's quite hard to get measurements that align really well with published ones, there are too many variables.
- The most enlightening measurement was my comparison of the theoretical room response at lf compared to my measured response. This shows clearly that I am putting a very accurate signal into the room.
- REW is very impressive, especially for the asking price!
- The ear is the final arbiter. LS50s + ZRBs sound superb, vastly better than the LS50s on ther own and certainly amongst the best hifi I've heard (e.g. considerably better than Kii3s).
 
I too would like to understand more about gating and how to go about implementing it correctly in REW.
 
In light of h.g.'s comments I think I will revise my conclusion from:
My conclusions:
- Overall I think It's quite hard to get meaningful measurements and they can be tricky to interpret.
- REW is very impressive, especially for the asking price!
- The ear is the final arbiter.

To:
- Overall I think It's quite hard to get measurements that align really well with published ones, there are too many variables.
- The most enlightening measurement was my comparison of the theoretical room response at lf compared to my measured response. This shows clearly that I am putting a very accurate signal into the room.
- REW is very impressive, especially for the asking price!
- The ear is the final arbiter. LS50s + ZRBs sound superb, vastly better than the LS50s on ther own and certainly amongst the best hifi I've heard (e.g. considerably better than Kii3s).
I think you have misunderstood REW’s room simulation, it just shows you the modal response of your room, based on the dimensions of your room, nothing to do with the accuracy of otherwise of your system.
Keith
 
I think you have misunderstood REW’s room simulation, it just shows you the modal response of your room, based on the dimensions of your room, nothing to do with the accuracy of otherwise of your system.
Keith

I completely understand it Keith. I think you misunderstood my conclusion.
Maybe I should remind you that you were the one who didn't know the difference between and mode and a node!
 
Any info on the ZRB?
What amp is powering it?

Cheers
Matt

Hi Matt, sorry I cannot disclose details because it is being commercialised... the amplifier has some special characteristics and works with a compensation system that corrects the response of the drive units. There are many advantages: box tuning is not relevant, can use several small drivers, 10X reduction is driver distortion, lf rolloff can be programmed as required.

I doubt it will be called ZRB when it becomes available, but I will flag it up as soon as I can.
 
Hi Matt, sorry I cannot disclose details because it is being commercialised... the amplifier has some special characteristics and works with a compensation system that corrects the response of the drive units. There are many advantages: box tuning is not relevant, can use several small drivers, 10X reduction is driver distortion, lf rolloff can be programmed as required.

I doubt it will be called ZRB when it becomes available, but I will flag it up as soon as I can.

Ok thanks for the info.
I was just curious(after people’s comments about phase)incase you were using an inuke dsp amp as these have a phase reversal somewhere between the input and output, meaning you have to correct this wherever you choose in order for the inuke to operate in-phase with the rest of the system.

Cheers
Matt
 
I followed the instructions here:
https://www.minidsp.com/applications/acoustic-measurements/loudspeaker-measurements

But I was not happy with the results, I need to work out where the variation came from.

Thanks. Based on this info then it seems gating is of most use in loudspeaker design applications (or if you want a more accurate indication of how your loudspeaker measures without the interaction of the room it's in). However it says that gating is unsuitable for low frequency measurements ("...any frequency below about 1 / (window length) cannot be represented...") Besides, if you're trying to balance subs with mains, then surely you want the room's contribution to be INCLUDED in the measurements, as this is ultimately what your ears are going to hear?
 
"...any frequency below about 1 / (window length) cannot be represented...

Ah I think that must be the explanation, because it was at lf that I thought the result was wrong.

Somehow my brain must have gated :oops: out that sentence when I read the instructions.
 
I didn't try DSP as such... it was the 10 band software graphic equaliser in Volumio, that I tried. I was not able to improve the sound even though the centre of the bands fell quite nicely for my purposes e.g 31Hz to reduce the main room mode.
However this is a very blunt instrument, so I would not like to draw conclusions.
 
graphic equaliser in Volumio, that I tried. I was not able to improve the sound even though the centre of the bands fell quite nicely for my purposes e.g 31Hz to reduce the main room mode.

While the center frequency may coincide, the width (Q) of the room and the filter certainly will not. So the result will be far from flat.

An equaliser can only cancel a room mode when it is fully matched. Which means that you need 1) a parametric and 2) accurate measurement.
 
I understand that there is now a parametric EQ plug-in available for Volumio. I have no idea how well (or not) it works or how stable (or not) it is.
 
Werner, yes i totally agree - hence my comment about a "very blunt instrument".

Lee, is that different to the 10 band graphic equaliser function I was using? This seems to be called "Volumio Simple Equalizer" and has 10 frequency bands of unknown (to me) Q.
 
S-man, I'm afraid your sub are out of phase. the dip of 15db at 45hz is probably simply caused by the mains being out of phase from the subs. Maybe its really your room but it would take 2 minutes to make sure...

To:
- Overall I think It's quite hard to get measurements that align really well with published ones, there are too many variables.
yes, in a untreated room, its impossible.
the only variable is the front end + the room. The room is the biggest problem to acheive the flat FR of the speakers at the listening position. especially if you have neutral front end.

frequency response measurements only show a small part of in room performance; it doesnt show the even arguably bigger problem of a room which are decay (ETC) of frequencies. arguably, a good ETC response is even more important the a good FR.

the only way to have even decay at all frequencies at the listening position is with serious room treatment: very deep bass traps and early reflection panels at all points. that could be your next step, I found it to make a amazing improvement.
 


advertisement


Back
Top