[sackley]

I have a friend and fellow audiophile that thinks THD is the most important thing in a system. Is that true? I can't imagine that it is the tell all story for amps and preamps. I hope someone can help me.

 

[Beagle]

No, I think THC is the most important thing in your system

 

[chych]

THD is like... SPL, or Freq response....

It doesn't mean much... as long as it is <1% you are fine, and usually they are like <0.05%. Not like it makes a difference to your ears anyway.

 

[sackley]

Yeah, THD as in Total Harmonic Distortion and that is what I thought. So, when looking for a preamp or an amp, what should I look for statwise?

 

[coolvij]

The only thing is that there are a ton of mini-systems...i think my mom will be getting one soon. Some have THD's of <1%, others, <10%. The latter will sound CRAPPY!

Stating the obvious,

- vij

 

[skippy]

i remember reading somewhere that low THD used to be a big selling point. amp designers would use lots of negative feedback to lower THD to obscene levels, but introduced other problems (TIM and such). a low THD doesn't tell much about an amp.

 

[sackley]

This is good info. Duncan, what is TIM? I need to make sure that I have plenty of info so I can more convincingly shoot his arguement about the joy of THD down.

 

[AndreYew]

THD+N as a measurement is basically useless. THD+N vs. frequency is only slight less useless. The reason is that it is basically impossible to predict, aside from really pathological cases (super high THD+N or incredibly low THD+N), the audible properties of a box from THD+N measurements because the same THD+N number can be derived from very audible phenomena as well as very inaudible phenomena. By crunching all the measurements into a single-number, you're throwing away all the information that may tell you anything about audibility.

THD+N is called THD+N because noise is indistinguishable from distortion in the measurement. This should already tell you something about how bad it is.

--Andre

 

[lini]

sackley: TIM is transient intermodulation, another kind of distortion spec that describes how clean the signal runs through the zero line (ground level). It has something to do with the linearity of a transitor's amplification - that's where class A operation comes into play: If you add some additional - uhm... what's the English equivalent for Ruhestrom?

- constant current, you lift the transistor into the linear part of its amplification curve, but thereby you sacrifice quite a lot of power. Some amps can be switched from regular class B into class A operation - the Marantz PM8000 for example. This can be helpful, if you want to use your amp for both personal Hi-Fi use and as PA-workhorse for parties....

Nevertheless, it's yet another spec: If you don't exactly know how each manufacturer measures, it will only serve as a hint - quite similar to S/N-ratio, THD, wattage... Nowadays even cheap amps have very good specs (on paper, at least), so the specs won't tell you a lot, anymore. Usually I find it more revealing to have a look inside the amp, because the types and dimensions of parts as well as the board layout and cabling can tell you quite a lot if you have a little experience. So better trust your ears - and maybe your eyes, too.

Greetings from Munich!

Manfred / lini

 

[blr]

I agree that giving a THD as a single figure is meaningless.
What is needed is a total harmonic distribution function.
This is for 2 reasons:
First we are more sensitive to higher order HD and second we perceive even and odd harmonics differently. A classical example is an older tube amp design that can put up to 1% THD without sounding harsh or distorted. Thats because most of it is a second HD. Some SS designs have much lower THD but it is maynly third harmonic and thus perceived as dry and harsh sound.
There is quite some psychoacustics involved here.

 

[mbriant]

THD levels also come into play when discerning the power output of an amplifier. Power output figures should be accompanied by a THD number as well. Some boomboxes and car amps fudge their power output figures by quoting them at 10% THD instead of the standard <1% THD. A small car amp for instance claiming a power output rating of 50 watts or 100 watts, is very likely quoting at the 10 % THD level. This means that the amp is really only approximately only 5 or 10 watts at less than 1% THD.

 

[AndreYew]

Oddness and evenness of distortion really doesn't predict audibility. 6th order distortion sounds pretty darn bad, compared to 3rd order. What's important is comparing the distortion spectrum against human masking thresholds, as that will directly tell you if something is audible or not. 2nd order is hard to hear not because of its inherent evenness, but because it's close enough to the main signal, in frequency, to be masked by the main frequency.

--Andre

 

[blr]

AndreYew.
You're right that higher order harmonics are easier to hear (as i also said) but I don't think 2nd order is inaudiable due to band overlap. perhaps at very low frequency but at higher freq. the2 nd harmonics are far appart. i guess it's a psychoacustics issue.

 

[System1Engineer]

I spent years in electronic classes learning A,B,C (yes C class amps used in microwave) and D class came in later. Class A amps have very very low THD as the output power transistors are biased on 100% of the time. A clean output "Tube" or Transistor provided incredible low total harmonic distortion. THD.
Class B amps are a dual push / pull design with an NPN / PNP Transistors coupling the output typically through huge transformers to help reduce distortion caused at the crossover / switching point (some manufactures biased each transistors into the opposing side to get out of the nonlinear region of the power transistors on - off switching point. "This is where THD was generated on class B amps.

Now overdriving an amp will put any transistors into "saturation" / also non linear region - distortion will start to rise there also - but THD typically not measured at 110% output.

In closing.
Macintosh Amps = 0.002% THD. So clear the sound is boarding. I own one... Over kill clean. For $$$$$$$$
Old school Rockford Fosgate (and comparable Orion, pioneer amps) < 0.05% super clean. The standard.
Newer MOSFET technology varies widely.
Amps you want to listen to music????
<0.5% is detectable if you listen to everything.
If the amp is rated at 0.5% or up, it's not clean by old school and audiophile standards. These amps with that level of total harmonic distortion will cook your ears at loud volumes, verses I have listened to clear music at ratings of <0.05% THD in cars and <0.002 THD in home and I'm 55 years old and still hear every bend of a chord Gilmore pills on his Fender playing Pink Floyd and Every note Stevie Ray ever played.

THD no matter how you package it / call it was taught to me to be the "Tell" of the quality of an amplifier.
It's 100% true. What good is an amp the distorts when it reproduces music.

And if it distorts 0.5% at a low volume then it's going to do at least the same thing at high volumes.

Hope this helped educate as well as inform. If you want to listen to it loud and clear get the lowest THD rating possible. If you want to cook your ears and let them ring 24-7 buy any amp you want.

My bottom line for BOTH Car sound and my big bass 8000w Orion (car 0.05% or less with sub at <0.5%).

And YES you can hear 1% distortion in the Bd class amps - they START at 1% advertised. You want bass guaranteed 0.5% or less across the board (THD to rated power).

Thanks for reading.

Dr. E

 

[GRUMPYOLDGUY]

Harmonic distortion is just one measurement describing what the spectrum looks like. It is not the complete picture...

If I tell you a car has four wheels, can you tell me what color it is? Of course not.

In terms of frequency domain analysis, I'm looking for spurs, noise floor, harmonics, accuracy (if I put a 1KHz tone in, do I get a 1000Hz tone out, 1001? 1050?), gain or attenuation, etc.

There other interesting measurements too... Phase noise is a big one.


Sent from my E5803 using a highly trained, special forces carrier pigeon