Speaker distortion vs Preamp-Amp-DAC-distortion

[RayDunzl]

So yeah, as a genre, ESL's are among the lowest distortion speakers... (if not THE lowest distortion type...)

I've seen some measures of cones and domes do as well.

Magico comes to mind.

I heard a big pair with the big amps and the big subs in a big room at the big show here in Tampa. They made me take notice.

Somehow, wherever we were in the room, the sound was just "there".

 

[dlaloum]

The good manufacturers were happy to publish their comprehensive distortion data back in the day. Here's an example, Yamaha NS-1000X from 1991/2:

Woofer:
View attachment 225083

Midrange:
View attachment 225084

Mid directivity:
View attachment 225085

Tweeter:
View attachment 225086

Tweeter directivity:
View attachment 225088

The whole system:
View attachment 225090

Power and system directivity:
View attachment 225091

Most of these seem to make my point...

The woofer graph shows 34db signal/noise @ 40Hz = 2% THD

The Midrange shows 40db SN @ 2500Hz = 1% THD

Tweeter is getting there with 45db @2500Hz = 0.56% (it would depend on where it was crossed over, as it does better higher up... 0.25% @ 9kHz

Still... an ESL at its worst gets up to 0.5%... and a dynamic struggles to get there - as a genre the ESL is low distortion whereas magnetic drive cones/domes etc basically aren't. (which isn't to say there aren't exceptions in the categories... but as a generalisation by category, that seems to fit!)

 

[MaxwellsEq]

Distortion is not coherent, so one distortion cannot hide behind another. There are some references in the thread to "masking", which is lifted from psycho-acoustics, but is out of context when discussing distortion.
Imagine a sound event (not a sine wave, real music) arriving perfectly at a power amp which distorts slightly: the sound event will be modified. The modified sound goes to the speaker. Imagine that this speaker has low distortion for this specific sound event. In which case the modification introduced by the amp passes exactly to the listener.
Distortion is NOT like noise, it isn't a floor that is hard to hear through.

 

[dlaloum]

Distortion is not coherent, so one distortion cannot hide behind another. There are some references in the thread to "masking", which is lifted from psycho-acoustics, but is out of context when discussing distortion.
Imagine a sound event (not a sine wave, real music) arriving perfectly at a power amp which distorts slightly: the sound event will be modified. The modified sound goes to the speaker. Imagine that this speaker has low distortion for this specific sound event. In which case the modification introduced by the amp passes exactly to the listener.
Distortion is NOT like noise, it isn't a floor that is hard to hear through.

If it is sufficiently low level, and it is harmonically related to the signal - it is frequently very difficult for a listener to perceive - music is made up of harmonics, timbre and tone are made up of harmonics - you add a sub layer of additional harmonics... and it can be very hard to perceive.

Even more so if the recording being listened to is not "acoustic" but "electronic" in nature, and therefore there is no baseline - no measure in acoustic tangible non recorded reality that one can refer to. - When the audio is itself a synthesised collection of harmonics, how does one differentiate between the intentional and that which is added by the replay signal chain? - so Yes it can be masked.

Known recordings with a real life benchmark will often show up distortion like no recording will - the voice of family members - the sound of an acoustic instrument you are intimately familiar with in real life, playing... - do they sound real .... typically, only you can tell, because only you have heard them "acoustically" - it may not seem "audiophile" - but the sound of your wife's voice, recorded live - can be the best test of all!!

 

[Tomino]

To what extent are the cumulative distortions in the electronics chain masked by speaker distortions? My WAG is that they are completely masked, presuming the electronics are even halfway decent, but that's just a guess. And even if they aren't, their audible contribution to signal reproduction must certainly be minimal.

I've had the same thought and have built my system around the assumption this is true. I wish I knew.

I'm surprised that more attention isn't paid to this question, because it so plainly would inform how a person ought to spend their audio budget.

My speakers are far nicer than my other components, and I often wonder whether I would hear a difference if I upgraded my amp. My guess is maybe -- quite possibly not -- especially as I don't listen to a lot of power-hungry bass music.

FWIW, I run Revel F208s off of a relatively cheap AV amp (Onkyo TX-NR686, purportedly with 100 watts per channel (2 channels) at 8 ohms; 0.08% THD).
Signal source is Roon, with DSP filters set up using REW, using a $140 DAC (SMSL Sanskrit Mk2), a Raspberry Pi, and plain speaker wire. No sub engaged when listening to music. No room treatments other than furniture placement. The speakers are the highest-end component by a very wide margin.

So what do people think? Should the speakers be more than ten times the cost of the amp, assuming basic power needs are met?

 

[fpitas]

So what do people think? Should the speakers be more than ten times the cost of the amp, assuming basic power needs are met?

I'd say yes. Basic competent electronics is getting cheap enough.

 

[RayDunzl]

If it is sufficiently low level, and it is harmonically related to the signal - it is frequently very difficult for a listener to perceive - music is made up of harmonics, timbre and tone are made up of harmonics - you add a sub layer of additional harmonics... and it can be very hard to perceive.

A single note of "music", the electrical output of a bass guitar, just to show the level of harmonics in the timbre of the instrument as plucked.

It was a "soft" pluck, less higher frequencies, as I wanted the fundamental and not the second harmonic to register at the highest level.

 

[dlaloum]

I've had the same thought and have built my system around the assumption this is true. I wish I knew.

I'm surprised that more attention isn't paid to this question, because it so plainly would inform how a person ought to spend their audio budget.

My speakers are far nicer than my other components, and I often wonder whether I would hear a difference if I upgraded my amp. My guess is maybe -- quite possibly not -- especially as I don't listen to a lot of power-hungry bass music.

FWIW, I run Revel F208s off of a relatively cheap AV amp (Onkyo TX-NR686, purportedly with 100 watts per channel (2 channels) at 8 ohms; 0.08% THD).
Signal source is Roon, with DSP filters set up using REW, using a $140 DAC (SMSL Sanskrit Mk2), a Raspberry Pi, and plain speaker wire. No sub engaged when listening to music. No room treatments other than furniture placement. The speakers are the highest-end component by a very wide margin.

So what do people think? Should the speakers be more than ten times the cost of the amp, assuming basic power needs are met?

Yes! (but the assumption of basic power needs being met, can be more complex than some people think... depending on the speakers)