I've been reading a lot about THD, but it all seems to be about whqt it is not how it can affect music reproduction. Can anyone explain it in those terms?
Engineers do harmonic distortion measurements to assess the quality of audio amplifiers ..objectively. If the numbers are low, some people will say it sounds good. But some people enjoy a particular kind of distortion such as in guitar amplifiers.
Visit this page and scroll down to see a few graphs.
https://www.diyaudio.com/community/threads/howto-distortion-measurements-with-rew.338511/
If the right side of a graph which shows the 1khz noise and harmonics has odd numbers it might indicate an annoying sounding amplifier.
1) Just for starters it confuses what you are hearing, it is something which was not there before, it´s not part of what Music was recorded.
2) to boot: in general it´s nasty ugly.
3) imagine looking at a nice picture through scratched dirty glass.
.
2) to boot: in general it´s nasty ugly.
3) imagine looking at a nice picture through scratched dirty glass.
.
THD measurements are important when validating an engineering design.
For example A DAC employing a high-end DAC IC that measures .005% THD+N (-86dB) would be a fail.
Amplifiers that use primitive vacuum tubes, especially in the output stage, are usually not expected to measure well in terms of THD.
For example A DAC employing a high-end DAC IC that measures .005% THD+N (-86dB) would be a fail.
Amplifiers that use primitive vacuum tubes, especially in the output stage, are usually not expected to measure well in terms of THD.
THD is defined as a ratio of the sum of power of all harmonics vs the power of fundamental.
If your amp is perfectly linear, signal out equals signal in. No additional harmonics, zero distortion.
No such amp exists. Even Halcro had some distortion.
I heard terribly sounding amplifiers with barely measurable distortion. And great sounding amplifiers with significant distortion. Type of harmonics matter. There are two camps, and they will never agree.
Last edited:
So in this picture I snagged from the thread link, does that huge spike at around 1khz going to sound horrible?
Also, I’ve been thinking about this after seeing a video about measuring THD of a speaker driver.
Looks like something X would produce. Extremely clean. No sign of 60Hz or 120Hz power supply. Wish my stuff was this clean.
Ok, still confused. In the graph, does that spike at 1khz represent a loud sound or a massive garble or what???
Ok. You start out by applying a 1kHz test tone. Extras appear at 2kHz, 3kHz etc and these are the distortion components.
More than 99% of the original 1kHz test tone is still there. What you want to see is where the harmonics are in relation to it.
More than 99% of the original 1kHz test tone is still there. What you want to see is where the harmonics are in relation to it.
As simply as I can:
You apply a sine wave (the 'Fundamental frequency', often 1khz - the large spike you are seeing.) to your device under test and measure the sine wave that is output.
If the measured output matches the input precisely, then there is no distortion (never happens). You would see one spike at 1khz corresponding to the signal you applied and nothing else.
What you will actually see are peaks of varying magnitude at different frequencies.
Those that are a positive integer multiple of the fundamental frequency are called 'harmonic frequencies'. For example your graph above shows 1KHz Fundamental, then 2nd and 3rd order harmonics at 2KHz and 3KHz respectively.
If the even harmonics peak higher than the odd (that is 2KHz and 4KHz etc are higher than 3KHz and 5KHz etc) then the amp is said to sound warm or pleasant. If the odd harmonics are higher, the amp is often described as clinical or harsh or sometimes unpleasant.
This is a generalisation and subject to eternal argument, but a basic understanding is the goal.
You will usually also see spikes around 50Hz with another lower one at 100Hz. This is your mains noise or hum.
The percentage THD is effectively how your output signal differs from the input - 0.000% would be no distortion, 0.001-0.01% is a reasonable solid state or opamp, 1%-10% is reasonable for a tube or guitar amp, 10%+ is usually awful.
Again in terms of sound reproduction, a little distortion can 'colour' the sound and give character to an amp's voicing. Too much, or the wrong type can make an amp sound unpleasant and often indicates a problem.
Hope that helps in some way.
It may help to realize that the difference in sound between, say, a piano and a flute playing the same tone is due to the harmonics produced by the instrument.
So you can appreciate that if an amp or DAC adds harmonics to the sound, it 'colors' the sound.
Jan
So you can appreciate that if an amp or DAC adds harmonics to the sound, it 'colors' the sound.
Jan
I know exactly what harmonic distortion is... but I have no clue what "whqt" is.
Electronic harmonic distortion is caused by resonances introduced by the electronic circuit.
Music is composed of signals with composed of a fundamental frequency and a number of harmonics.. the "harmonic envelope" if you will. This is specific to an instrument and a way of playing.
Adding extraneous harmonic resonances to the signal changes the sound, the tone, of the instrument.
Worst case it might make a grand piano sound like an upright honky tonk piano. Or a Stradivarius sound like something you got for free at the County Fair. Or Beverly Sills sound like Lady Gaga.
Last edited:
Thank you avtech, for a simple but very informative and useful answer/explanation! Much appreciated.
No. It is not caused by resonances. It is caused by non-linearities in devices and/or circuit.
Jan