Bruce's sounds the clearest of these three to me (but he was also using the best recording gear). What kind of Room / setup / speakers / treatment do you have?
Listening Room Intelligibility Test
- Thread starter Art Noxon
- Start date
Regarding the photo from WFB Expert, a number of posts back.....a self photo with styrophone drivers strapped to his ears...This is so funny, I can't get the image out of my head. Just wanted to let you know...It needs a life of it's own......Art
Last edited:
Now there's an undeniable truth. I wish I could play, but I don't have a functioning mic in the house, much less a calibrated one.
Tim
Three high quality systems sounding so different, hmmm -- I could use a rude word, but I'll probably get my wrists slapped, so I'll do it this way: d!$^@*^ion ...
Frank
Steve williams
Site Founder, Site Owner, Administrator
I totally agree. The amount of great info here is amazing.
Stay tuned for some more additions
You guys are so creative, listening to each others MATT test. What a neat way to exchange and compare data sets. I never thought of swapping test runs. But then, you guys are such a lively bunch, you are already pushing the edge of the art... This medium of communication allows people to work things out in their own space, as they have time, and contribute. And it seems to me each contribution, no matter how "off track" it might seem, actually does lead to yet another piece of the story being told. This is just great. I listened to one test, because the techs had gone home today and then I find out you guys are way ahead of me, swapping tracks.
It starts at 28 Hz, and climbs steady at 16 Hz per second to 780 Hz and reverses. You should be able to estimate the frequency of the problem area.
Usually the articulation part is captured by mics, as long as there is sound there. The level depends on the mic roll off. Lot of recording circuits are for vocal and they heavily roll off anything below 80 Hz.
You can use RadioShack sound meter as a mic preamp and set the meter to C weighting and plug into the "audio out" RCA port. this is line level. Oh, be sure to switch gain so that the needle is wiggling and not pegging much throughout the test. Radioshack is linear amp for this monkey business for what ever scale you set it on, you record liinear for plus or minus 15 db. That's a linear 30 db range. Pretty good for most of this work. I forget LF roll off of that meter. Probably on instruction sheet.
If you record low frequency part of test and it seems to be too quiet compared to when you actually were there and doing the recording, then your meter is set on A weighting or the mic is a vocal mic or the meter has a built-in LF roll off.
I'll listen to all the tests soon and try to catch up with your listening.
I'll also capture them, process and print them. I might try to publish a few copies here if the owner is OK with that. Let me know who want their data published. Maybe you want to see it first privately on email before you declare for all to see. also, when I see the priintout, I'll look for evidence of weird data taking. like a high noise floor, or clipping, things like that. .................. art
It starts at 28 Hz, and climbs steady at 16 Hz per second to 780 Hz and reverses. You should be able to estimate the frequency of the problem area.
Usually the articulation part is captured by mics, as long as there is sound there. The level depends on the mic roll off. Lot of recording circuits are for vocal and they heavily roll off anything below 80 Hz.
You can use RadioShack sound meter as a mic preamp and set the meter to C weighting and plug into the "audio out" RCA port. this is line level. Oh, be sure to switch gain so that the needle is wiggling and not pegging much throughout the test. Radioshack is linear amp for this monkey business for what ever scale you set it on, you record liinear for plus or minus 15 db. That's a linear 30 db range. Pretty good for most of this work. I forget LF roll off of that meter. Probably on instruction sheet.
If you record low frequency part of test and it seems to be too quiet compared to when you actually were there and doing the recording, then your meter is set on A weighting or the mic is a vocal mic or the meter has a built-in LF roll off.
I'll listen to all the tests soon and try to catch up with your listening.
I'll also capture them, process and print them. I might try to publish a few copies here if the owner is OK with that. Let me know who want their data published. Maybe you want to see it first privately on email before you declare for all to see. also, when I see the priintout, I'll look for evidence of weird data taking. like a high noise floor, or clipping, things like that. .................. art
Steve williams
Site Founder, Site Owner, Administrator
Bruce B
WBF Founding Member, Pro Audio Production Member
This is a purpose-built mastering room. Speakers are Evolution Acoustics MM3, Pass Labs XA-100.5 amps, Crane Song Avocet pre and Playback Designs MPS-5. Here is a pic.
Room is about 22 x 13 x 10.5
Mic was about 9' 9" from speakers, 37" from floor and volume was about 85dB
Attachments
Last edited:
With the 3 test results already posted it would be very valuable to know what the actual volumes at the mic positions were, roughly, and how far from the speakers were the results recorded.
I do note in Audacity that the waveforms are dramatically, dramatically different: we can't be in Kansas any more, Toto ...
Frank
I do note in Audacity that the waveforms are dramatically, dramatically different: we can't be in Kansas any more, Toto ...
Frank
I figured as much. That's way cooler than mine, which is just your average living room, more or less:
The mic was ~12 ft from the speakers (on the couch), the sound level was around 75-80 db or so (I didn't measure it, though).
Here is how my system currently measures.
Art, I don't mind if you post results on mine here. It would be a great learning experience.
Doug, I don't follow the 2 frequency response curves: is the blue trace just an averaged response, or measured using very coarse increments? The red trace looks very savage, + and -15dB limits; is that a reasonable, realistic room response? If it is, don't know about people complaining about my miserable HT fella! ...
Frank
Doug, I don't follow the 2 frequency response curves: is the blue trace just an averaged response, or measured using very coarse increments? The red trace looks very savage, + and -15dB limits; is that a reasonable, realistic room response? If it is, don't know about people complaining about my miserable HT fella! ...
Frank
Pink is the raw response, blue is the same thing with 1/3 octave smoothing. It is typical to see a very jagged raw response when measured from a single position. If I were to average multiple raw traces across the listening area, the raw response would look a lot better (I have done this in the past). The 1/3 octave smoothing better approximates the response of our ear/brain in terms of how we perceive the sound. This is why you generally see smoothed curves when people report frequency response.
Excellent explaination.
People have 2 ears, each located in different positions. Our ear have directionality features where they tend to highs more from front than back, due to ear shape, the pinna. Our head tends to shadow one ear and the other from side to side sound in mid and highs. And then we move around, always averaging the tonal color we hear over a wide variety of positions. And additionally, we have the Haas effect at work, which is adding the history of early reflections with the direct, to creating an overall percieved sound level/spectrum. What the human hears is so much more processed, averaged over distance and over time, compared to the simple snapshot view of the world seen by the microphone. The human is not an electrical circuit and listening is not what a mic/electronic does. The closest we get to electronics is a battery.
When the MATT is run, the top outline of the MATT test shows how loud the sound gets during frequency sweep. It is what the blue line shows. It will be intereting to compare the two versions of the same loudness curve.
We are processing the 3 MATT recordings that were submitted. We should have results posted today for the 3 rooms. Thanks for being brave.
Art Noxon
Acoustical Engineer
Invented TubeTraps, Pres of ASC
www.acousticsciences.com
People have 2 ears, each located in different positions. Our ear have directionality features where they tend to highs more from front than back, due to ear shape, the pinna. Our head tends to shadow one ear and the other from side to side sound in mid and highs. And then we move around, always averaging the tonal color we hear over a wide variety of positions. And additionally, we have the Haas effect at work, which is adding the history of early reflections with the direct, to creating an overall percieved sound level/spectrum. What the human hears is so much more processed, averaged over distance and over time, compared to the simple snapshot view of the world seen by the microphone. The human is not an electrical circuit and listening is not what a mic/electronic does. The closest we get to electronics is a battery.
When the MATT is run, the top outline of the MATT test shows how loud the sound gets during frequency sweep. It is what the blue line shows. It will be intereting to compare the two versions of the same loudness curve.
We are processing the 3 MATT recordings that were submitted. We should have results posted today for the 3 rooms. Thanks for being brave.
Art Noxon
Acoustical Engineer
Invented TubeTraps, Pres of ASC
www.acousticsciences.com
We have data. the first one is Full Test File which shows all 3 MATT test print outs. The second is LF Range, which shows the performance foir all 3 rooms in Low Frequency range, 28 to 100 Hz.
If someone wants to post these so they show up visually somewhere, please do that. I'm not sure how I can to that right now, I tried the img thing, it didn't work.. for me. I'll let you guys look at them and listen to them as you look at them and then I'll jump in with some observations.
First observation is to notice how extremely articulate Bruce's studio is, between 15 and 20 db, nearly full bandwidth. And below 150, it's about 10 dB articulation. We see ome level shifts, one at 135 Hz and one at 50 Hz. My guess is that we have a 1at and 3rd harmonic "partial cut" or suck out, due to a rear wall bounce that has not been weakened enough yet. Also the bass levels seem a little hot below 100, probably becaue he's fighting the two partial cancels.
Notice Doug, less articulate in general but fairly consistant at about 8 to 10 dB. Typically stronger articulation above 500 Hz. See sound level variations. He ha almost no gargling. Quite a roll off starting at 70 Hz.
Notice Gary. He has a noise floor. Look to the left of the LF range printout. that low frequency noise is the HVAC or something noise. Notice how even the lower edge of his articulation envelope is? It looks like Bruces studio. But Bruces is real and Gary's articulation is limited by the noise floor. He probably would look like doug if he turned the AC off. Doug also ha stronger articulation and level starting around 600 Hz. He lolds level and reasonable articulation down to 40 Hz but we can't see much below that because the noise floor dominates the lower sound levels. He could fix this by playing the suytem +10 dB and then he'd pull the data out of the mud. Oh, Gary does have a premature roll off in level starting at 300 and steady down to 75 Hz. He probably has a sub that picks up the bottom end. I'm not sure what would cause that roll off. woofers phase reversed? curious. EQ in system? Possibly the speakers are playing in front of an opening to another room and the bass is not loading properly. Copy/paste whole line, including the .pdf part even if it isn't underlined and blue. donno what happened there.
http://www.acousticsciences.com/articles/MATT_comparisons_Full Test File.pdf
http://www.acousticsciences.com/articles/MATT_comparisons_LF range.pdf
Remember, we haven't begun to analyze exactly what is "wrong" anywhere and how to fix it. For that, a set of photos and good floor plan, and side view is needed. For now, worrying about analysis is outside our learning curve. But it's enough for now to know that it can be done.
Art Noxon
Invented TubeTrap, Pres of ASC
www.acousticsciences.com
If someone wants to post these so they show up visually somewhere, please do that. I'm not sure how I can to that right now, I tried the img thing, it didn't work.. for me. I'll let you guys look at them and listen to them as you look at them and then I'll jump in with some observations.
First observation is to notice how extremely articulate Bruce's studio is, between 15 and 20 db, nearly full bandwidth. And below 150, it's about 10 dB articulation. We see ome level shifts, one at 135 Hz and one at 50 Hz. My guess is that we have a 1at and 3rd harmonic "partial cut" or suck out, due to a rear wall bounce that has not been weakened enough yet. Also the bass levels seem a little hot below 100, probably becaue he's fighting the two partial cancels.
Notice Doug, less articulate in general but fairly consistant at about 8 to 10 dB. Typically stronger articulation above 500 Hz. See sound level variations. He ha almost no gargling. Quite a roll off starting at 70 Hz.
Notice Gary. He has a noise floor. Look to the left of the LF range printout. that low frequency noise is the HVAC or something noise. Notice how even the lower edge of his articulation envelope is? It looks like Bruces studio. But Bruces is real and Gary's articulation is limited by the noise floor. He probably would look like doug if he turned the AC off. Doug also ha stronger articulation and level starting around 600 Hz. He lolds level and reasonable articulation down to 40 Hz but we can't see much below that because the noise floor dominates the lower sound levels. He could fix this by playing the suytem +10 dB and then he'd pull the data out of the mud. Oh, Gary does have a premature roll off in level starting at 300 and steady down to 75 Hz. He probably has a sub that picks up the bottom end. I'm not sure what would cause that roll off. woofers phase reversed? curious. EQ in system? Possibly the speakers are playing in front of an opening to another room and the bass is not loading properly. Copy/paste whole line, including the .pdf part even if it isn't underlined and blue. donno what happened there.
http://www.acousticsciences.com/articles/MATT_comparisons_Full Test File.pdf
http://www.acousticsciences.com/articles/MATT_comparisons_LF range.pdf
Remember, we haven't begun to analyze exactly what is "wrong" anywhere and how to fix it. For that, a set of photos and good floor plan, and side view is needed. For now, worrying about analysis is outside our learning curve. But it's enough for now to know that it can be done.
Art Noxon
Invented TubeTrap, Pres of ASC
www.acousticsciences.com
Last edited:
Steve williams
Site Founder, Site Owner, Administrator
Bruce B
WBF Founding Member, Pro Audio Production Member
Thanks, Art... very interesting. I think the roll-off is artifactual since, if anything, my system measures bass heavy. I am kind of thinking I should try to get a better measurement with my other mic, which is what I used for the Holm measurements. Would the sound level make much difference in the results? My room is quite live, with no special treatments. (I couldn't get to those links either, by the way).
I got the ECM8000 working (turns out I had to deactivate USB power management to force the phantom power stay on).
FWIW, here is an updated version recorded at ~85 db > LINK
I got the ECM8000 working (turns out I had to deactivate USB power management to force the phantom power stay on).
FWIW, here is an updated version recorded at ~85 db > LINK
Last edited:
garylkoh
WBF Technical Expert (Speakers & Audio Equipment)
Many thanks Art. This is illuminating and interesting. I hope that others will post their results. I think that looking at the charts, and listening to the tests with your explanation also gives us a better understanding of what to listen for in a room.
You are right - mine is an industrial space with an industrial HVAC running outside the room. Unfortunately, there is also road works going on outside my building this whole week - so the low frequencies and the noise are likely also contributed by the heavy machinery. I'll have to come back one evening to make another recording. That is when I normally work on design anyway.
The speakers are my little 2-way monitors. No sub or woofers.
http://www.genesisloudspeakers.com/g7p1p.html
You are right - mine is an industrial space with an industrial HVAC running outside the room. Unfortunately, there is also road works going on outside my building this whole week - so the low frequencies and the noise are likely also contributed by the heavy machinery. I'll have to come back one evening to make another recording. That is when I normally work on design anyway.
The speakers are my little 2-way monitors. No sub or woofers.
http://www.genesisloudspeakers.com/g7p1p.html
When we have a mic or preamp with a roll off, the modulation effect is not changed, just how loud the modulation gets can change between loud to quiet. Loudness and intelligibility are different things. They are independant variables. You can change loudnes and not change intelligibility or you can change intelligibility and not change loudness. You might have fun doodling and showing yourself how these change can take place. show also how sound level can remain the same and intelligibility goes to zero. Loudness and intelligibility are independant of each other. Isn't that intereting. OK, maybe I've just crossed the line..sorry
Imagine what happens to intelligibility when compression schemes are introduced in an effort to produce a more uniform sound level. Loud sounds get turned down and quiet sounds get turned up... that means the difference between loud and quiet is reduced, which means intelligibility is reduced.
Live concerts are so noisy and so difficult to mic that the audio signal is seriously compressed in an effort to get more of the music delivered above the noise floor without over loading during peaks.
See how easy it is to understand the impact of compression once you understand how intelligibility works. Without it, the connection between compression and intelligibility becomes very vague.
Thank you Bruce for saving the day. I'll have techs try to figure out why the lnks didn't work. I know I had a hard time pasting them into the page here. The link kept changing right before my eyes.
Art Noxon
Invented TubeTraps, Pres of ASC
www.acousticsciences.com
Imagine what happens to intelligibility when compression schemes are introduced in an effort to produce a more uniform sound level. Loud sounds get turned down and quiet sounds get turned up... that means the difference between loud and quiet is reduced, which means intelligibility is reduced.
Live concerts are so noisy and so difficult to mic that the audio signal is seriously compressed in an effort to get more of the music delivered above the noise floor without over loading during peaks.
See how easy it is to understand the impact of compression once you understand how intelligibility works. Without it, the connection between compression and intelligibility becomes very vague.
Thank you Bruce for saving the day. I'll have techs try to figure out why the lnks didn't work. I know I had a hard time pasting them into the page here. The link kept changing right before my eyes.
Art Noxon
Invented TubeTraps, Pres of ASC
www.acousticsciences.com
For fun, go to Gary's MTF, that's Modulation Transfer Function. Start on the far right. We are going to count tone burst peaks. The first peak in fron the right is under the 2 of 20. That's 1. Now 1 thru 6 are clean peaks. 7, 8 and 9 are scrambled peaks, not spikes as before. And then we see 10, a double peak. And if we continue we count over to 15 and there is another double peak. We talked about this before, I think. Anyway, here is the perfect example of a sound cancel reflection effect. Remember, time here moves from left to right. The left hand peaks of 10 and 15 are the rise in the loudness of the direct sound. All of a sudden a reflection hits the mic and instead of the peak continuing to grow, it gets partially cancelled. The pressure of the reflection is in time with but out of phase with the direct signal, and quieter. Still, the out of phase collision of the two wave trains gives us a dip in the sound level. And as long as the tone is on, the cancel reflection effect stays there, canceling. Then the tone turns off and what do you hear? Not the speaker, you hear the tail end of the sound canceling reflection whipping by your head until it also runs out of sound. I just love that. Art
Similar threads
