About
Passion for audio combined with expertise and honesty.
Because what you want need and what you get are two completely different things...
Last edited: October 1, 2026
Tonestack.net is a technical website focused on loudspeaker design, digital audio processing, analysis of audio measurements and many other things related to "sound reproduction" or "audio engineering". It's a great resource for those who want to learn new methods, walk new paths and also to move away from formulaic answers.
The site philosophy: "If your experiment needs statistics, you ought to have done a better experiment." - Ernest Rutherford. The biggest mistake in science is reducing scientific research to calculating probabilities, scientific knowledge to correlations, and ignoring causality, the "mechanism" or model. This leads to a "calendar creation science", where everything is calculated, but little is understood. (An illustration of the problem: we can create a working calendar without understanding anything about the Solar System, just like the ancient Egyptians did... The modern version is the mindless application of data analysis tools and formulas in the name of science.)
My obsession with audio technology started with loudspeakers. Later my interest shifted to guitar amplifiers, guitar amplifier modelling, lossy audio compression, headphone measurements and concert hall acoustics (perception of surround sound). Fortunately, I also have a good understanding of music, which helped me a lot. (Playing an instrument can add such a new dimension to life that turntables and tube amplifiers can never provide... These old technologies are fascinating, but also outdated and should be left in the museum.)
Behind every successful project there are many failed attempts and a lot of struggle. A few stops from the "the long and winding road" follows.
Initially I thought that high-order crossover filters in loudspeakers are a source of audible time smearing. My original idea was that if I built a loudspeaker with time-aligned drivers and transient optimized crossover, I would have the best loudspeaker on the planet... Transient optimized loudspeakers are not very common as they require special drivers, true first-order crossover filters and acoustical center alignment between the drivers (usually in the form of an angled baffle). Fortunately, I dropped the project in time. After running a number of tests with software-emulated crossovers and doing some background research (a classic: Group delay distortions in electroacoustical systems by Blauert & Laws), I let go of the idea of transient response optimized "linear phase" crossovers. (12th order Linkwitz–Riley at 2 kHz can be detected with pulses, but not the 4th and 8th order.)
Measuring headphones on a flat plate fixture is a great example of how simplicity can be a solution to complex problems. Flat plate measurements provide a good prediction of how headphones sound and the relevance of the response is not worse than that of dummy head measurements. Unfortunately, there is a lot of noise in this research field. There are plenty of preference studies without any attempt to understand headphone acoustics or headphone-ear acoustics. Scientific models are discarded, target curves are evaluated without modeling the acoustics of headphones, whereas studies filled with p-values are created similarly to products manufactured on an assembly line.
Many people confuse high fidelity with high-end audio. High fidelity means lack of audible coloration ("transparency"), whereas high-end audio is an unfair business based mainly on people's insecurity and gullibility. If we look under the hood of a high-end audio product or development, we will find either rather dumb solutions or the lack of real development wrapped in an extraordinary design (sold with a dumb marketing). This is especially true for overpriced "audiophile grade" DACs, amplifiers, streamers... As for headphones and loudspeakers, there may be exceptions that can be labelled as high-end and not nonsense.
One of the greatest misconceptions in audio is related to blind testing. Though confirmation bias and other types of self-deception are real problems and should be avoided, the idea that blind testing is the only way to eliminate bias is certainly wrong and very limiting. People believe in nonsense because they don't know how to brake down a complex problem into smaller, easily verifiable tasks, and not because of the lack of blind tests. In addition, without proper knowledge blind testing can be either misleading or may give us an illusory knowledge. Understanding the science behind audio technology does not require blind tests, just a reasonable understanding of hearing, signal behavior (analog, digital) and the complexity of real life signals and sound sources.
Latest:
Headphone measurements: what outer ear measurements can tell us
Headphone measurements: with or without ear?
Lossy audio compression: principles, methods, misconceptions 🔊 🎧
How much noise and distortion can we hear in complex signals? 🔊 🎧
Speaker Driver Simulation With Room Response (online simulator)
NextLevelFun - A Huge Collection Of Music Videos 
