Rational Acoustics



Matthieu Rousseau
May 31st, 2017, 07:06 AM
Hi,

I'm new here and also to measurement software. I read a lot on this site and other forum and I finally started measurement. I have questions relative to my first experience. This is not a "real life" one, I just look at a monitor speaker at home and I have really basics question relative to this measures but I can't find an answer on this forum or others. I know there is a lot of experienced guys here and I hope maybe one of you will have time to answer me.

954955956957958

All the measurement are on axis 1/48oct smoothed and 2sec average.

On the first picture I did a close up measure (30cm) of the speaker, good.
Pictures 2 & 3 I moved at 1meter and 1,5meter of the speaker.

On each measure I used the find delay function and inserted it. I noticed that when I move the microphone away from the speaker, the phase slope seems to increase. Why that ? As I inserted the delay I am supposed to see the same phase response, right? I consider this and thought that maybe the fact that I am in a small room result with 1 - very little time between the direct sound of the speaker and the reflections 2- Not a lot of level difference between reflections and direct sound. Also the more I put the microphone away from the speaker the more this is accentuated in the measure because I reduce level difference between direct and reflections. This can explain the mess in the phase response when I put more physical distance between monitor and measurement mic.
So I made another capture at 1,5meter but this time I decreased the FFT size to 1k , trying to exclude reflections of the measure. The result is on picture 4. And hurray ! the phases slope are more matching which make sense to me but I loose a lot of detail due to 1k FFT...

I understood that measurement is a lot of how to interpret the graph, can you tell me if my analyse is correct please ? Also when you are in real situation is this something to do to start taking a close up measure of the system giving us a "clean image" of the system that we can compare with the other measurements in the room and helping us to interpret the graphs ?

My last question concern picture number 5 where you can see all the measurements at once. The blue one and the pink one are respectively 30cm and 1,5meter from the speaker , MTW and 1k FFT. They match in the high frequencies. The green and light blue are 1 meter and 1,5meter both MTW, those two match in the high frequencies but differ from the 2 other measure. I don't understand that ...

Sorry for the noob side of this post, I searched on the web and can't find an answer. Hope someone could help me I will really apreciate !

Chris Tsanjoures
May 31st, 2017, 03:29 PM
Hi,

I'm new here and also to measurement software. I read a lot on this site and other forum and I finally started measurement. I have questions relative to my first experience. This is not a "real life" one, I just look at a monitor speaker at home and I have really basics question relative to this measures but I can't find an answer on this forum or others. I know there is a lot of experienced guys here and I hope maybe one of you will have time to answer me.

Welcome to the party Mattieu!




All the measurement are on axis 1/48oct smoothed and 2sec average.

The MTW gives you 48th oct resolution from 60Hz up and 1 Hz resolution up to 140 Hz, leaving the smoothing at None would give you higher resolution. Take for example, a measurement at 32k FFT and MTW, at 1/48 oct smoothing they look exactly the same. However, using a single size FFT takes away any of the benefits of the MTW (the ability to tell which signals are reflections and which are noise, as each MTW time constant is chosen to give the best determination of signal to noise in its effective range).


On the first picture I did a close up measure (30cm) of the speaker, good.
Pictures 2 & 3 I moved at 1meter and 1,5meter of the speaker.

On each measure I used the find delay function and inserted it. I noticed that when I move the microphone away from the speaker, the phase slope seems to increase. Why that ? As I inserted the delay I am supposed to see the same phase response, right? I consider this and thought that maybe the fact that I am in a small room result with 1 - very little time between the direct sound of the speaker and the reflections 2- Not a lot of level difference between reflections and direct sound. Also the more I put the microphone away from the speaker the more this is accentuated in the measure because I reduce level difference between direct and reflections. This can explain the mess in the phase response when I put more physical distance between monitor and measurement mic.

So I made another capture at 1,5meter but this time I decreased the FFT size to 1k , trying to exclude reflections of the measure. The result is on picture 4. And hurray ! the phases slope are more matching which make sense to me but I loose a lot of detail due to 1k FFT...

This is likely do to the positioning of the mic being somewhere in between the sample of the previous measurement. Please note that the difference in phase you are seeing is less that a .002 ms offset, and using the +/- delay buttons would vary the slope wildly, while only adjusting in .02ms increments.

The MTW takes choosing the appropriate FFT size out of the equation for you, as we've already chosen FFT sizes to have the appropriate time constants for their operation ranges, largely based on the past 2 decades of experience developing dual channel FFT based analysis. Pg.7 of the user guide goes into these concepts in greater detail, definitely an interesting read as a beginner user: http://www.rationalacoustics.com/download/Smaart-v8-User-Guide.pdf


I understood that measurement is a lot of how to interpret the graph, can you tell me if my analyse is correct please ? Also when you are in real situation is this something to do to start taking a close up measure of the system giving us a "clean image" of the system that we can compare with the other measurements in the room and helping us to interpret the graphs ?

For a single speaker, phase is showing you that systems frequency response relative to the reference signal. In other words, what frequencies from the reference signal are on time with those from the measurement signal, which frequencies arrive before and after the reference signal. For a single system, asking if the phase is 'correct' is sort of the wrong question. Take the first image, if you want the analyzer to be timed to the energy at 8-10KHz then, yes, the phase response is correct. If you want the analyzer to be on time with the energy at 800Hz then, no, the phase response is not correct - at 800Hz the measurement signal lags the reference. (see Phase Response, in the User Guide, pg 102)


My last question concern picture number 5 where you can see all the measurements at once. The blue one and the pink one are respectively 30cm and 1,5meter from the speaker , MTW and 1k FFT. They match in the high frequencies. The green and light blue are 1 meter and 1,5meter both MTW, those two match in the high frequencies but differ from the 2 other measure. I don't understand that ...

Basically, you are somewhere in-between a sample and are seeing that difference. The resolution of the delay finder changes at .02ms increments, and you are probably in between that - ie some sub-sample delay value, ry adjusting the delay increment and see if you get it closer, or just move the mic a little bit and re-track/find the delay.


Sorry for the noob side of this post, I searched on the web and can't find an answer. Hope someone could help me I will really apreciate !

No worries, again, welcome to the party! absorb as much info as you can from as many people as you can and just when you start to get it you will realize how much more you need to learn. I swear this stuff is a rabbit hole that has no bottom!

Matthieu Rousseau
May 31st, 2017, 05:53 PM
Many thanks for your detailed response Chris !

Timo Beckman
April 29th, 2018, 07:55 AM
Hi Matthieu maybe this helps a bit?

https://www.youtube.com/watch?v=s9rV31fC1Yw&t=258s