Once again we continue reading the definitive tale of TubeTrap creation, evolution, function, and usage. This week we expand upon the Modulation Transfer Function, and learn how its use in high end audio rooms informs us of the musical accuracy contained therein.

We hope you gain a new appreciation for the importance of dynamics in your playback chain after reading this excerpt. To quote Claude Debussy, “The music is not in the notes, but in the silence between them”.

Enjoy!

Read the entire paper

Modulation Transfer Function

It turns out that when people talk they create on the average, world wide, about 8 separate sound level fluctuations per seconds. For this speech testing they used two frequencies, one was a 2 octave wide noise centered at 200 Hz and the other was 2 octaves wide centered at 2k Hz. Great Scott! We were doing the same test except that we were using gated pure tones instead of gated noise. I stayed up all night and wrote a short white paper and in the morning I asked Dr. Peutz to read it. Sure enough, he confirmed to my delight that we were actually running narrow band MTF testing, narrow band intelligibility or articulation testing in Hifi rooms. We named our version of playback intelligibility MATT. Musical Articulation Test Tones and it has become one of the audio acoustic reference test signals in the industry. See Stereophile Test CD #2, Track 19. Later I discovered that the same kind or research about sonic pulse rate for music had been done and again, musical sonic events occur worldwide average at 8 times a second, the pattern of dynamic pulses for music is on the average 8 Hz. What I had stumbled into was a, maybe the, test for musical intelligibility.

Something else was very interesting about this test. It was the only test we ever tried that produced results which were in sync with our customer’s enthusiasm for having added TubeTraps, at no small expense and inconvenience, to their systems. Instead of showing them ½ dB adjustments in the frequency response curve, we were showing them 3 to 10 dB improvements in their musical articulation response curve.

In other words, adding TubeTraps to the room allowed the room to change sound level more rapidly, to be more responsive to musical dynamics. It’s not how loud the music was during a pink noise test or frequency sweep that mattered, it was how quickly it could respond to changes in loudness. Sorta like dynamic headroom in amps, except this was the acoustic side of the situation.

When we ran a MATT test in someone’s room before it was trapped, they might display dynamic acoustic headroom of, say, 3 to 5 dB between 30 and 700 Hz at an 8 Hz tone burst rate. After TubeTraps were loaded into the room we would easily measure 6 to 10 dB dynamic acoustic headroom across the board. Finally we had a test that made sense. When we ran the test, the customer always wanted the low headroom bandwidths fixed first, not suck-outs, modes or room boom. Garbled sound was worse than anything else. And even more, they knew where these problem areas were because they could pull albums out and play the passages that would also demonstrate where the rapid dynamics were garbled.

Wow, the audiophiles already knew all about this type of garbled sound problem. I’m the one who didn’t know about it. They knew what was wrong with their system all the time. When we talk about it, they usually admit that they thought this was an amp problem. I wonder how many good amps were returned, and new amps bought because people were trying to fix a garbled passage in an album?

This kind of disconnect between the seasoned participants and upstart engineers is typical. I had long before learned to always believe whatever the audiophile or recording engineer was saying, even if I couldn’t understand it. They were talking about their experiences at the edge of perception, where no words exist to describe what they see. But in this case, no one had ever mentioned it, or more likely, they did, and I just didn’t understand what they were saying to me at that time.

Psychoacoustics

Studying this idea of dynamic headroom brings some interesting understandings to light. Dynamic headroom is about how many dB of sound level difference a person can hear. If we just start playing music very quietly say at 30 dB and the volume control is rotated to 100 dB we have experienced a sound level difference of 70 dB. This is about absolute loudness and it’s not what we are talking about. However, if we make this sound level change quickly, loud..quiet..loud..quiet…say 8 times a second we are talking about hearing rapid relative loudness changes.

If we have a set of 60 dB tone bursts, whose electronic signal level changes between 20 dB and 80 dB at a rate of 8 tone bursts per second, at the listening chair the acoustic signal might only be changing between 75 and 80 dB. We are only hearing 5 dB tone bursts. However, if we put headphones on and we apply 60 dB modulation, and compare it to 50 dB, 40 dB 30 dB and even 20 dB modulation, we can’t tell the difference between any of these tests. Modulation between 40 and 80 dB sounds like modulation between 60 and 80 dB. Only when we hear modulation changes less than 20 dB can we recognize differences. We can hear the difference between 5 and 15 dB modulation.

Once again I thought we had the explanation nailed for what TubeTraps really did right. And as usual, to some degree, we had, but there was still that nagging issue of cleaning up low end articulation leads to improved treble detail. We sell hadn’t figured that part of the equation out. And so for this, let’s turn to the music theory classroom, where the outline for the sound level changes of a single musical event is defined. Where the dynamic nature of a sonic event is defined.