Does the amplifier voltage swing limits headphone dynamics?

[Fluffy]

In another forum someone has raised the issue that the dynamics of high impedance headphone can be limited by the output voltage of the amplifier. He says that because there is a peak in the bass frequency area (associated mostly with dynamic headphones as far as I know), the voltage swing of the amplifier can limit their dynamic range (or perceived dynamics; he wasn't very clear…). I want to emphasize that he says this occurs before any clipping takes place.

I want to know if this is a real phenomenon, or what he says is simply a result of compression that takes place at the end of the amplifier's ability to deliver any more wattage to the headphone, that will naturally be accompanied by distortion.

Furthermore, this is extra interesting for me because I think I experienced something similar. My Focal Clear sounds a bit more bassy and dynamic when it's connected to my Arcam rHead amp than to my Fiio X3 2nd gen (at the same volume level). though I can't be sure if I'm really hearing this or it's just a result of my expectation bias that I should be hearing a difference.

What say you, oh wise science-y folk?

 

[Frank Dernie]

Being limited by the voltage swing is clipping so the idea it could happen before clipping is bonkers.
Clipping does indeed bring about the onset of more distortion.
Clipping into high impedance 'phones will be a voltage limit and into low impedance 'phones a current limit, normally.

 

[Fluffy]

Clipping into high impedance 'phones will be a voltage limit and into low impedance 'phones a current limit, normally.

He said the same. but right after I asked about clipping, he said "I'm not talking about clipping, before that…" (I'm translating from Hebrew).

 

[Fluffy]

Ok follow up question – can clipping occur at one frequency and not at another? For example, the Focal Clear has an impedance of 70 ohm through most of its frequency range, but shoots up to 350 ohm at around 50 hz. When playing music, can the 50 hz area get compressed\clip before other frequencies do?

 

[Frank Dernie]

He said the same. but right after I asked about clipping, he said "I'm not talking about clipping, before that…" (I'm translating from Hebrew).

Sounds like a non-science person. That is absurd.

Ok follow up question – can clipping occur at one frequency and not at another? For example, the Focal Clear has an impedance of 70 ohm through most of its frequency range, but shoots up to 350 ohm at around 50 hz. When playing music, can the 50 hz area get compressed\clip before other frequencies do?

It has nothing to do with frequency just amplitude, but the amplitude at low frequencies is almost always way more than at higher frequencies in music.
If your headphones have a big variation in impedance with frequency a headphone amplifier with a higher than normal output impedance will cause changes in frequency response.

 

[Fluffy]

If your headphones have a big variation in impedance with frequency a headphone amplifier with a higher than normal output impedance will cause changes in frequency response.

That I know. But will the low frequency compress more relative to high frequencies, say if the amplifier doesn't have enough voltage swing to drive them passed a certain point? If you are familiar with multiband compression effects, I ask if this condition has a similar outcome to that effect.

 

[Frank Dernie]

That I know. But will the low frequency compress more relative to high frequencies, say if the amplifier doesn't have enough voltage swing to drive them passed a certain point? If you are familiar with multiband compression effects, I ask if this condition has a similar outcome to that effect.

I don't know what multiband compression effects are, unless you are referring to adjustments in an equaliser.
I suppose if you go loud enough to go out of the linear range of the headphone driver before the amp clips this would be a compression effect but otherwise if it is a limit in the amp it is clipping - it may be soft clipping at the onset but if the limit is the voltage swing of the amp it is simply clipping.

 

[Fluffy]

Multiband compression is a type of effect in audio mixing, that compresses different parts of the spectrum differently.

I suppose if you go loud enough to go out of the linear range of the headphone driver before the amp clips this would be a compression effect

Just to understand, "going out of the linear range" is what seen in this graph as the sharp rise in THD at a certain wattage level?

 

[lucian]

You usually don't have abrupt clipping. Some amps even have soft clipping to have better transition. You can also see this in some reviews from Amir, as he get better performance if he deals down the gain.

Maybe that's what is meant ....

https://en.wikipedia.org/wiki/Gain_compression

 

[Frank Dernie]

Multiband compression is a type of effect in audio mixing, that compresses different parts of the spectrum differently.


Just to understand, "going out of the linear range" is what seen in this graph as the sharp rise in THD at a certain wattage level?

View attachment 36632

No.
What I wrote is the driver in the headphones going out of its linear range, that means the coil is moving out of the linear part of the magnetic field or thermal effects.
Compression in the amplifier caused by the limit in voltage swing is clipping. It may not be sudden but that is the only thing.

 

[Fluffy]

Alright then. Do you think you can provide an objective explanation to my subjective experience I described?

Furthermore, this is extra interesting for me because I think I experienced something similar. My Focal Clear sounds a bit more bassy and dynamic when it's connected to my Arcam rHead amp than to my Fiio X3 2nd gen (at the same volume level). though I can't be sure if I'm really hearing this or it's just a result of my expectation bias that I should be hearing a difference.

Of course, I'm not rejecting the possibility of this being simply expectation bias, because the effect was really small. I just wonder if limited voltage swing and\or amplifier power can by itself result in a less dynamic or less bassy sound out of a headphone.

 

[lucian]

No.
What I wrote is the driver in the headphones going out of its linear range, that means the coil is moving out of the linear part of the magnetic field or thermal effects.
Compression in the amplifier caused by the limit in voltage swing is clipping. It may not be sudden but that is the only thing.

I dont want to split hair - but compression is NOT clipping - when clipping occurs here is no increase in voltage at the output anymore. With compression here is still some increase of voltage following the input.

 

[Fluffy]

I dont want to split hair - but compression is NOT clipping - when clipping occurs here is no increase in voltage at the output anymore. With compression here is still some increase of voltage following the input.

I think that is what is meant by soft clipping - basically really high ratio compression.

 

[Nango]

I took from a german EE forum that all "serious" headphones will start their impedance never below 80-100 ohm just because a lower impedance will get EVERY amp (namely its voltage converter) into self-oscillation, means distortion. He was referring to the region around 100Hz where the body of the music resides and where one regularly detects lack of punch very soon.

He also explained that folks from Burson at a certain point where the first to build opamps where they claimed to "workaround" the issue and marketed their opamps would sound even better than conventional amps. But he also added it is just "physics" and you definitely can't workaround this behaviour.

Fact is, most of the serious mfr release their flagship headphones with impedance starting around 80 ohm and much higher.

 

[Fluffy]

Are you being real now? There are plenty of high end headphones with impedances lower than 80.

 

[Nango]

Yes, being the only reason for that to play loud enough when driven by a mobile phone or similar and not for SQ reasons. What I mentioned was written in 2007 so doesn't apply to orthos Hifiman/Audeze.

 

[Fluffy]

This is a pretty bold claim, so I would like to see some proof of that. Preferably an actual explanation of why the limit is 80 ohm specifically.

 

[Nango]

No, not 80 Ohm specifically, but the lowest, the higher the better.

 

[Fluffy]

Why is 80 the lowest? What's wrong with 70? Or 50?

 

[Nango]

Fluffy, I have no idea why he wrote 80 Ohm the lowest, the story is: the lower the load, the sooner the converter gets into self-oscillation, that he said. Take 70 Ohm if you feel better with.