The heart of Magic TRIck

The heart of the system is the ADAU1701 chip. It’s a complete single-chip audio system with a 28-/56-bit audio DSP, ADCs, DACs, and microcontroller-like control interfaces. Signal processing includes equalization, cross-over, bass enhancement, multiband dynamics processing, delay compensation, speaker compensation, and stereo image widening. This processing can be used to compensate for real-world limitations of speakers, amplifiers, and listening environments, providing dramatic improvements in perceived audio quality.

The chip has got two ADC inputs, four DAC outputs and up to 8 digital inputs and outputs. I use 4 digital outputs in I2S format to send the signal to additional DACs.

Some program parameters, like signal gain and delay, are to be adjusted by the user in real time. They are transferred from a microcontroller to the ADAU1701 using the I2C interface. Continue reading

Coping with room modes

I have all my Magic TRIck v4 components measured and ready to work. Now it's time to connect all components and check their actions in the set.

In a room, it is the position of the loudspeakers and the listener, and the room itself, that have a major effect on how the system sounds, especially at low frequencies. What is heard at low frequencies is not likely to be anticipated from an examination of anechoic data, therefore only measurements can describe what is happening in a room.

By applying multiple subwoofers one can achieve fairly even low frequency response in many places of a room. My goal is to get the best sound in the sweet spot and the more even bass coverage in the room will be a bonus. Continue reading

Center it right

My Magic TRIck v4 system has the Trifield decoder implemented and therefore it reproduces the ordinary stereo program through three main full-range speakers: left, right and center. The theoretical and optimal model assumes that the listener is equidistant from these three main speakers. However from practical reasons the center speaker is often somewhat closer to the listener than the left and right speakers. This still an acceptable scenario as long as the delay and gain of the center channel are properly adjusted. Continue reading

Blanda Matt speakers

Magic TRIck in version 4 is a three channel stereo reproduction system, thus I need three main speakers to reproduce the main frequency range. Because the woofers will cover the bass range up to 250 Hz, the main speakers don’t have to be big. I want to use small full-range drivers to cover the whole band from 250 Hz. Therefore I don’t have to make passive crossovers for them and the drivers can be driven directly by the power amplifiers, preserving the consistency of the active loudspeaker system. Small drivers also have the wide dispersion up to several kHz, hopefully gently narrowing it above, which is desirable for uncolored sound reproduction in a room.

There are many good quality small full-range drivers suitable for this task, however some of them are somewhat rugged in treble region. My choice is 3.5” glass fiber cone TG9FD full-range drivers from Peerless (Vifa). They have a reasonable maximum linear excursion and smooth frequency response. They are also well regarded in many DIY projects. The drivers will receive nice IKEA salad bowl enclosures of a high WAF. Continue reading

Multi-sub trick

This is one of these magic tricks that have been applied in my audio system since the very first version. The idea is to reproduce the lowest audio frequencies with multiple separated loudspeakers.

In every room the reproduced sound is reflected by the room boundaries (floor, ceiling, walls, big furniture), and at certain frequencies two or more reflected waves superimpose themselves resulting in many dips and peaks in sound level (room modes). The effect depends on the sound source and the listener position.

At higher frequencies this is not a big problem: there are many modes relatively close one to another, giving only limited variations in sound level, and a human brain copes with the effect quite well. But below the Schröder frequency (between 150 and 300 Hz for a typical living room) the modes become a problem. Although there are fewer modes in this range, they are more spaced apart and give very perceptible dips and peaks in frequency response. A particular dip or peak can be mitigated by moving the sound source or the listener to another position, but it usually worsens the situation at another frequency. The difference in sound level is dominated by the room and not by the loudspeaker anymore. This commonly leads to the typical sensation of "booming" or "one-note" bass or even no bass at all at certain frequencies (Markus Mehlau). Continue reading

Loudness for Magic TRIck

One can build a great audio system with a smooth and flat frequency response and, as far as other parameters are also great, it will sound beautiful and real, however with a real sound pressure level only. A very loud whisper and a quiet organ concert don’t sound right, as their sound level is not adequate to our real experience. Yet there are many occasions when I don’t want to listen in real levels.

Every time the volume knob is not in its best position (regarding the sound characteristic and recording or mastering level) the audio system introduces distortions. The sound can be absolutely clear, but its tonality is perceived badly, as the sensitivity of human ear changes according to the sound pressure change. Too loud whisper can be corrected by lowering the volume, but setting the real level of a symphonic orchestra during late evening listening sessions is not an option. For bringing back the proper tonality some corrections to the sound must be done. Continue reading

Bass and tone

Many years ago I joined the audio puritans’ club (luckily I left it some time later). Even if I had some sonic controls close at hand in my audio gear, I didn’t used them because of the sound purity I wanted to enjoy. So what? Sometimes the music played in my car or through mediocre headphones gave me more joy. One of the most useful gauges I had been using for many years is the tiny knob adjusting the amount of bass in my desktop PC audio system. Maybe it isn’t high-end, but it always give the chance to make the sound right.

This is why Magic TRIck v3 was equipped with bass and tone controls, which proved to be highly useful. The next revision of Magic TRIck will have some even more controls. Continue reading

Trifield decoder

Magic TRIck will soon change into 3 main channel system for stereo program reproduction. Such systems have been known for over 80 years and they were first introduced by the film industry to make central sounds in a stable relationship with the screen. The first 3-channel decoders were however rather simple.

Much better decoders can be made nowadays by using modern electronics. Such a system is capable of subjectively enhanced realism as well as the improved stability of central images. One of the big hidden gains of 3-speaker stereo is its lower listening fatigue and artificiality as compared with 2-speaker systems. Continue reading

Where are you going, Magic TRIck?

Magic TRIck is the active loudspeaker system consisting of speaker set driven by eight-channel power amplifier. I built this system several years ago and it replaced my previous main living room audio system with the fairly famous Concertino tube amplifier.

Each speaker is driven by its own one of eight LM3886 power amplifiers. These amplifiers get the signal from two Analog Signal Processor boards - multi-channel active filters. The idea is well-known from many Siegfried Linkwitz projects.

Continue reading

Measuring Stratobaffle B200

Several additional Stratobaffle B200 measurement results are presented in this entry. The measurements were carried out in a medium-size living room (area 27 square meters;  height 2.6 m). The speaker was elevated, so that the driver was at the height of 1.3 m. The distance from the driver, as well as the microphone, to the nearest boundary object was not less than 1.3 m. The distance from the microphone to the driver was roughly 1.2 m. All measurements were spatially and time averaged by using the moving microphone procedure (in a +- 5 deg window). Continue reading