We got books, blogs, forums, "mountains of content - some better, some worse", but the most important learning style is missing from the web. Learning with interactive tools that were designed around a specific problem.
It's time to go back to the roots and restart software development that has been abandoned for a long time. "Noise Perception Modeler" and "Speaker Directivity Modeler" provide an insight into the new design philosophy (starting with simple...).
High-end mythology (anti-rationalism)
Audio technology is surrounded by an irrationality and subjectivity that is rarely seen in other areas of life. In the mid-70s hifi became a realm of marketing lies, scams, big promises, self-perpetuating ignorance and lack of accountability. Audiophile magazines constantly promoted the notion of "hearing is more accurate than instruments", which is not only a nonsense, but a straight way to self-deception. Moreover, audio myths have no negative consequences that would prevent their spread (similar to astrology). The lack of negative consequences and people's varying susceptibility to emotional and social control (e.g. fear of exclusion, possession of special things, cult of personality) are perhaps the most important factors.
It is somewhat ironic that the plethora of misconceptions about audio formats, DACs, cables or vinyl records clearly shows that the ear is not a precision or measuring device. The McGurk effect is probably the most embarrassing discovery about how our senses can deceive us (video about the effect).
Our auditory memory evolved for communication and surviving, not for testing amplifiers or cables. (Nor have our eyes evolved to test 2k/4k displays...) If we compare hearing and vision then we can see that the ear's analytical ability is much worse than the analytical capability of vision. We can't separate one part of the music from the rest like we can separate one part of the image from the rest.
The rational way - multidimensional audio
Audio technology requires "multidimensional" thinking (~multilevel framework). Audio measurements, models, concept of audio fidelity should be evaluated or analyzed on three levels:
- Characteristics of signals
Behavior of analog and digital signals (signal theory, signals and systems)
- Characteristics of hearing
Relationship between thresholds and signals, auditory masking, noise perception, this level helps identify relevant parameters and related measurements
- Characteristics of music, instruments, real sound sources
Frequency spectrum of instruments, rise time and signal rate of change, peak value and RMS, peak value and amplitude of main harmonics, Attack-Decay-Sustain-Release (ADSR) system, etc.
The first two are of critical importance. The third is essential for clarifying the following questions:
- power calculations (peak RMS, average RMS)
- relationship between cone excursion and sound pressure level (design software calculate with a sinusoidal signal, not a complex signal)
- interpretation of nonlinear distortion above 10 kHz (signal levels above 10 kHz are much lower than at 500 Hz, harmonics have the highest level between 50 Hz and 500 Hz)
Csaba Horváth is the editor, designer, software developer of tonestack.net. He is interested in loudspeaker design, audio compression, psychoacoustics, guitar amplifiers & amp modeling and software development. In his spare time, he patches his YouTube video player (Here's to the crazy ones).