Distortion in reproduction of music is one of the most complex and often misunderstood areas within the audio industry. The complexity of this subject make it hard for the bigger audience to understand what is really going on here… How ever for the people who want to know a bit more, I hope that this page can answer some questions within this area. Following text is just a brief overview of the topic and does not claim by any means to be a complete coverage of this complex but very interesting subject.
The most obvious example of a misused term is total harmonic distortion, THD. As it would be a measure of sound quality. Nothing could be more wrong. THD does not by any means take into account our perception of sound. In my view, THD is just a technical term that has minimal relevance in high-end audio.
Our hearing has a masking effect that is dependent of the frequency, the sound pressure level but it also have temporal characteristics. An effect of masking is that harmonic distortion close to the main frequency of a tone is much less audible within the masking curve of our hearing. On the other hand this also means that we are very sensitive to higher order distortion that are present outside the masking. As natural sound in most cases consist of a set of tones or harmonic components, our resulting masking curve for hearing in real life sound is build from all components of the sound we hear. As the sound change with time the masking curve follows this but with some lag as the masking has a decay curve as well.
In real life sound not only consist of a large number of tones and harmonics, it is very dynamic as well. And a tone or harmonic component that has a shorter duration than 200 ms start to get harder and harder to hear as the duration gets shorter. This is also relevant for the perception of micro details in music, but it relate more to acoustics and that is a bit outside of the scope here.
There are many kinds of distortion. Harmonic, intermodulation, linearity and non-harmonic distortion. All theses are sound level related type of distortions. But within the area of distortion we also need to include dynamic distortion and time related distortion. And in the context of our perceptual abilities, time related distortion is very audible. This has been addressed well in in some DAC designs, but it is very relevant in high-end amplifiers as well.
As voices, string and air flow instruments (instruments that can play a melody) often produce a rich harmonic content, a slight degree of added harmonic distortion does not by necessity mean that we loose perception of micro details and musicality. However the characteristics of the sound is affected by the distribution of such distortion. This means that the distribution in harmonics distortion can be used as a tool to change the sound from warmer sound to more analytical. It is generally preferred to have a successive falling level distribution of harmonic distortion components.
The harmonic content of sound is of most importance as it defines the timbre of an instrument. Instruments like pianos have a very rich harmonic content. The timbre we hear of a piano depend not only on the piano itself. How the piano is handled by the pianist, the recording room and what recording microphones are used and set up of those play significant roles. From a reproduction point of view, we have a responsibility to as accurate as possible recreate the harmonic content as it was recorded by the microphones. Many aspects in our feeling of musicality in home listening depend om how accurate micro details from voices and instruments are presented. Both in terms on amplitude and accuracy in timing.
Our listening tests has shown that we are sensitive to phase shifts at higher frequencies. This is natural as all what we hear in the real world is a spectrum of frequencies a any single point in time. If high frequencies are delayed more than others, our perception of the sound will be affected. This is the main reason why our amplifiers have a very high bandwidth. It’s not about the frequency range or a good looking specification, it’s about minimising phase shifts.
Summary
There are many potential design pitfalls that cause distortion to blur, mask or hide micro details in reproduction of music. This is especially relevant for high order complex and intermodulation distortion but not so much for reasonably low order harmonic distortion. Amplifiers with large amount of negative feedback in the design typically produce very low THD numbers that look good in the specification. Many of such designs create high order complex distortion patterns that have a negative effect in how we perceive micro details in recordings. Its very easy to design an amplifier with good specifications but much harder to design an amplifier where musicality is top priority. This require a much deeper understanding of electronics, discrete components, audio in general and last but not least our perception of sound.
So how what are the recommendations? How can you as an audiophile or music lover find the great components among all there is? Listening is of course the first and obvious answer. But all gear that you read about and are interested in may not be available for listening where you live. In this case its good to take advice from a reviewer that you think is inline with your values. Another advice is to read manufacturers web sites, product sheets etc. Are they really into what high-end audio is about or is it just marketing stuff? What do they want to achieve? Good looking specifications or musicality? You are always welcome to chat with us or our partners. We love to hear from you.
/Anders Hansson, Chief Designer