Rational Acoustics



Arturo
March 10th, 2011, 01:10 AM
It seems that there hasn't been a whole lot of action going on here the past few days so I thought I'd post something new and try to stir up some discussion. I've been taking measurements of my living room listening system recently in order to get comfortable with Smaart 7. So far so good.

I've attached a couple of photos to illustrate my system and to show my measurement microphone position as well.

The first screen shot I've attached is of a TF measurement of the left loudspeaker. The second is of the right. Each had slightly different delay times, just FYI.

The first thing I noticed was the LF bump at about 125Hz. I'm not too worried about that at the moment. The second thing I noticed was the phase wrap and the corresponding magnitude dip at 300Hz. I thought about this for a second and decided that it's really a problem at 600Hz and the first cancellation is occurring at 300Hz. So I typed 600Hz into my audio calculator and found the distance to be 1.86'. I thought about it a bit and decided that the wall behind the loudspeakers was to blame.

Next, another wrap very close to the first at about 450Hz. I typed 800Hz into my calculator to get 1.4' of distance. Could this be the back wall as well? I don't know.

Finally, and the most perplexing issue, is a persistent phase wrap beginning at about 4kHz and continuing on to 20kHz. I did a search for this type of behavior with no results.

Thoughts? Ideas? Am I doing this wrong? Please let me know. Thanks for reading.

Ferrit37
March 10th, 2011, 02:09 AM
Hey arturo,
first thing is to turn on the coherence trace, this is the third leg in the measurement triangle.
It can be used as a confidence meter in so much as it shows "what is the likelihood that the data is valid"?
if you see poor coherence, a dip in the magnitude, and a wiggle in the phase response, all at the same point, this is a good clue that it's "bad" data at this point.(probably caused by a reflection)

Secondly take an impulse measurement, and you should be able to identify the different arrival times of the reflections. if you move the loudspeaker with respect to it's boundaries you should see these dips move.
The different delay times between speakers could be attributed to the mic not being exactly equidistant from each source. Turn on Delay tracking and you should be able get them pretty close.

The "phase-wrap" , you can go into settings and show phase as unwrapped, (it can be a little unstable) but what you should see is a phase response that descends from left to right. This is typical of these type of systems (google: group - delay)

An interesting experiment is to take a spectrogram(spectrograph?) measurement of one speaker whilst moving the microphone around, you should be able to get a sense of the polar pattern. In the horizontal, check out where the HF tails off and in the vertical, on axis, see if you can spot the "lobing" from driver offset at X-over point.

Also place a speaker on the floor and take a ground-plane measurement to look at "semi-anechoic response"
Finally place two speakers side-by each measure, then move one and watch the comb filter change.

all these thibgs will give you a "taste" of what to expect/see when aligning systems.

Arturo
March 10th, 2011, 10:44 AM
Thanks Ferrit.

I toggled the coherence back on. Thank you. I wasn't sure why it wasn't on there then I looked at the options menu again. Oops!

I've also unwrapped the phase response measurement and am in the process of learning more about group delay.

I'm going to take IR measurements today and compare the results to the TF measurements.

Also, I'll give the other experiments you suggest a try as well.

Thanks for all of the information.

Ferrit37
March 10th, 2011, 04:03 PM
Hey arturo,
this is also very good : http://www.excelsior-audio.com/Publications/Phase_Response_&_Receive_Delay.pdf
:)

Arturo
March 11th, 2011, 11:26 AM
Ferrit,

Thanks for the article.

One portion stood out to me and it was "we have removed the correct propogation delay when the phase response in the frequency region well above the cutoff frequency of the DUT approaches a horizontal line".

So I started doing some more reading and was confused when I read some instructions given by Harry concerning alignment of a two-way system. It reads "Turn on the HF only. Set your delay in smaart. Adjust the smaart delay to get the flattest overall phase trace through the center of the passband".

Is this because the "frequency region well above the cutoff frequency" of the HF in the DUT presumably would be near 20kHz and we can't see phase information beyond that point?

Also, I took IR measurements of each of my loudspeakers yesterday and the T=0 was identical to the delay in the TF. I believe the paper by Charlie Hughes is explaining how to use group delay to set the delay for a TF measurement but at the moment that part is a bit over my head.

Thank you for pointing me in the right direction.

Ferrit37
March 11th, 2011, 03:02 PM
Hey Arturo,
I've attached a doc that pretty much lays out interpreting the phase trace.
enjoy :D

Arturo
March 20th, 2011, 12:53 PM
Thanks Ferrit.

I'm going to read that right now.

Kip Conner
July 24th, 2011, 05:39 PM
Hey Arturo,
I've attached a doc that pretty much lays out interpreting the phase trace.
enjoy :D

This is a great doc, I had to explain to someone today why their phase trace was sponsored by the letter "N"...

Question- is there something after Langston's pic that shows the polarity flip? There are two blank pages and it seems that there is something there if you try to highlight. Maybe not?

Arthur Skudra
July 24th, 2011, 06:07 PM
This is a great doc, I had to explain to someone today why their phase trace was sponsored by the letter "N"...

Question- is there something after Langston's pic that shows the polarity flip? There are two blank pages and it seems that there is something there if you try to highlight. Maybe not?
Maybe pages 11 and 12 are out of polarity with the first 10? Or maybe the coherence threshold on those two pages is set too high! :D Sorry, couldn't resist!!!