Can we agree headphone amplifiers are solved?

[garbulky]

What would you gain by more power? What headphone do you have the necessitates more power than what is currently available?

I use HD700 plan to get hd800s. Also want to try hifiman units. Your same question could be directed towards the inaudible distoretion that most amps already have

 

[Fluffy]

I use HD700 plan to get hd800s. Also want to try hifiman units. Your same question could be directed towards the inaudible distoretion that most amps already have

Those don't require more power than what most amplifiers can provide.

And yeah, everything having inaudible distortions is my point exactly.

 

[solderdude]

I don't think music is 'special' - it's just much more complex than standard test tones. As I said previously, the component sine tones of music are far greater in number, and far more varied in amplitude, and unequally-spaced in frequency. None of the standard test tones are representative of this complexity. The use of simple test tone distortion to judge performance in real-world usage contains the implicit assumption that this distortion has a positive monotonic correlation with the distortion DUTs would produce with much more complex superpositions of many different sine tones. Is there any hard scientific evidence that this will universally hold for all DUTs (or at least just amps)? If there isn't, then the possibility remains this correlation may not be monotonic. And yes, Serge's findings that Df for music (real or simulated) signals does not have a monotonic correlation with Df for sine signals is, while maybe not robust evidence, at least suggestive that the previously mentioned assumption may not be correct. It may turn out none of this applies to amps - I was (perhaps erroneously) inferring that if it's a possibility for DAPs/DACs, it could be a possibility for amps. Nevertheless, as @waynel pointed out here, although an ideal amplifier is memoryless, in practice this may not be the case, which could result in their transfer function (and so distortion) being time- and previous state-dependent. In turn maybe these states could be signal-dependent (varying with complexity for example), which could result in an increase in distortion with more complex signals such as music. Who knows. And I think that's the point - how do we know 100% that the mechanisms of nonlinearity we observe with simple test tones are the only ones, and there aren't in fact any others that may manifest themselves only with more complex signals?

There are a lot of mays, assumptions, possibilities, maybes, perhapses, coulds, who knowses, thinking in your epos aren't there.
Is there a possibility that perhaps, maybe some of your assumptions and possibilities could be erroneous thinking ... who knows.

Can I ask you if you have ever done serious electronics testing, did actual well performed blind tests, did some actual nulling using electronic nulling equipment, do you have years of experience in designing, fault finding analog equipment ? Do you have a degree of sorts in electronics ?
And if so... what made you doubt that so much that you use 'mays, assumptions, possibilities, maybes, perhapses, coulds, who knowses'

 

[solderdude]

if one kind of IMD test can miss 'something you wouldn't expect' yet another IMD test reveals this something, how can we be sure there are not other 'things we wouldn't expect' that might only reveal themselves when playing more complex signals such as music?

When something is not revealed by 1 test but it shows in another test than you did not test for the correct things in the first place. This has nothing to do with complexity and music but everything with not running enough or useful tests.
Your whole thesis of maybes and coulds all depends on the belief that more complex signals (music) could reveal something that would have stayed 'hidden' otherwise but have no evidence of this.

 

[solderdude]

I use HD700 plan to get hd800s. Also want to try hifiman units. Your same question could be directed towards the inaudible distoretion that most amps already have

The HD700 is 100dB/V, the HD800S = 105dB/V so you need substantially less power for the HD800S. Also the impedance of the HD700 is lower so draws more current (and current is responsible for distortion)

When you plan to get the older HE-6 or a Susvara then indeed you will need an amp that can supply a lot more voltage/power.
Best to connect these directly to a speaker amp directly. No need for a headphone amp at all.
In that case the dual balanced monoblocks are readily available. You can lower the idle current of these class-A immensely as you won't need much current. The amps will stay very cool that way.

 

[garbulky]

Those don't require more power than what most amplifiers can provide.

I’ve tried several amplifiers and have to disagree. I have driven multiple amps to audible clipping. Also the quality of the sound even during normal volume suffers with wimpier amps. Is it power or that they don’t hold up The power output over the entire frequency band/ varying impedance I don’t know. but the best sounding amp I heard was also vastly more powerful than the others.
Also the hifiman stuff and some fostex are very power hungry

 

[solderdude]

When you have a varying impedance the output power varies per frequency (lower for higher impedance so even more easy to drive)
Voltage efficiency does not change.
An amplifier is a voltage source. The amount of 'power' depends on the drawn (not supplied) current. The dissipated power (mW in case of the Senns) is extremely low. No amp should have any problems with that. Even a simple op-amp can supply more than enough current. It is one of the easiest loads around for any amp. You do need to be able to reach enough voltage though. This too is not a technical challenge.

 

[restorer-john]

Those don't require more power than what most amplifiers can provide.

These people believing they want more power in headphone amplifiers are in dreamland. The output powers of many of the offerings are already well into the ridiculous territory.

 

[restorer-john]

Also the hifiman stuff and some fostex are very power hungry

Well just connect them up to the output of say, a Benchmark AHB-2 and see how you go. It's low noise, low THD and can swing 30V+ RMS over your cans. That should be enough to warm up those voice coils.

 

[March Audio]

I think we have seen some implementations (wasn't RMEs ADI one of them?) that pushed very slightly past the datasheet specs even, but that difference was very small if I remember correctly. I generally support your statements regarding this "professional" vs "semi-professional" gear debate, if you aren't going to do some FPGA magic you are stuck with the same ICs everyone else uses. And I know too little about FPGAs to even say wether it would be feasible to consider that route towards higher A/D performance.

On original topic:

I think they are a solved problem. From my point of view it's all about the currents. Lower currents flowing means worrying less about non-linearities of components, ground layout and so on. Not that you don't have to deal with that kind of stuff, you obviously have to, but you will usually be able to do so by applying common engineering "best-practices" or sometimes maybe by throwing more copper at the problem.

There are some tricks you can do such as paralleling channels which will reduce noise (but not distortion), but the fundamental performance of the ADC is the limiting factor.

As mentioned in an earlier post it may be possible to implement some esoteric solution that isn't based on these ADC chips but that won't be seen in the sorts of "professional" product we have talked about above.

 

[garbulky]

When you have a varying impedance the output power varies per frequency (lower for higher impedance so even more easy to drive)
Voltage efficiency does not change.
An amplifier is a voltage source. The amount of 'power' depends on the drawn (not supplied) current. The dissipated power (mW in case of the Senns) is extremely low. No amp should have any problems with that. Even a simple op-amp can supply more than enough current. It is one of the easiest loads around for any amp. You do need to be able to reach enough voltage though. This too is not a technical challenge.

What you say makes sense. There's no reason for tiny drivers to need huge power. But it doesn't reflect when I listen. Like I said I've hit audible clipping. I can also hear - not sure the correct term -saturation. When things become compressed, highs become distortion and strongly sibilant. While my much more powerful amp doesn't exhibit that. It's a speaker amp Bas-x A-100 in direct drive mode getting the full power of the headphones. Though when I do add a dropping resistor to drop its power output (albeit drastically increase its output impedance to 220 ohms), I can also easily hit clipping/distortion.

What was the class A monoblock option you were talking about? I only know of one class A headphone amp and that's the unreleased Emotiva headphone amp and it is not a monoblock.

 

[garbulky]

@solderdude I would be interested in hearing real world measurements of different headphones. Like measuring power levels and then measure real headphone output for the given power levels and see where distortion and compression sets in.

 

[Tks]

I agree. Headphone amps are not "solved" until we see higher SNR at low output levels, and perfect channel balance all the way down. The few headphone amps that do well in this area are rare enough to be the exceptions that prove that this is not "solved" in general.

Yeah channel matching is something more practical I think needs solving, especially at lower levels. But as solderdude mentions, this is actually already solved, the only issue is, manufacturers aren't doing it. Like for instance if someone wants to say you can get perfect balancing with stepped attenuates or something, well as a consumer I'm not really concerned why manufacturers aren't implementing this, the fact so few are, is what leads me to deduce it's not solved, because the notion of something being solved, would practically entail very little reason to abstain from implementing such solution.

It's like saying "IMD Hump on ESS chips is solved". Not really, and not without drawbacks, or not without someone privy to how it's done, or knowledgeable enough on how to get it done without drawbacks, or ESS hiding information concerning this issue and being silent.

Going by the strict definition of solved, technically almost everything is solved given enough impractical investment or approach to the issue. It'd also be like saying "oh you want to know how low the noise and distortion on an Toppind L30 is? That's simple! Just wait until Audio Precision releases a better performing and costlier measurement device in the next generation". Solved in my context is always with concern to practicality and attainability for normal people.

 

[JohnYang1997]

There are some tricks you can do such as paralleling channels which will reduce noise (but not distortion), but the fundamental performance of the ADC is the limiting factor.

As mentioned in an earlier post it may be possible to implement some esoteric solution that isn't based on these ADC chips but that won't be seen in the sorts of "professional" product we have talked about above.

Theoretically, if budget permits you can use ADC at low input level with extremely low distortion. Then paralleling them to gain back the noise performance lost then increase the noise performance further.

 

[March Audio]

Theoretically, if budget permits you can use ADC at low input level with extremely low distortion. Then paralleling them to gain back the noise performance lost then increase the noise performance further.

Very true, you might need a lot a ADC channels though, less than 3dB per doubling

 

[JohnYang1997]

Very true, you might need a lot a ADC channels though, less than 3dB per doubling

4x5578 per channel is definitely doable.

 

[March Audio]

4x5578 per channel is definitely doable.

32 to 1 ADC channels. Cool. About 14dB improvement. How is it's thd performance at -14dBfs?

 

[JohnYang1997]

32 to 1 ADC channels. Cool. About 14dB improvement. How is it's thd performance at -14dBfs?

I would say it really depends on implementation. -6dB to -9dB already can bring so much distortion improvement. Where around -15dB it should be under -130dB. This is my prediction.

 

[Blumlein 88]

Here is a spec sheet for Lavry's best ADC. LIsts THD+N at various levels.

I've checked some of the ADC's I have and at -20 dbFS the highest harmonic is -120 db from the signal at about -140 dbFS. If notching out the tone and looking at noise the SINAD would be 87 db referenced to the - 20 dbFS tone which is pretty much all noise. I don't have anything which is using close to a SOTA ADC chip. Lavry is only claiming -108 Sinad for - 20 dbFS signals as he has a much lower noise floor than my ADCs.

 

[March Audio]

Here is a spec sheet for Lavry's best ADC. LIsts THD+N at various levels.

I've checked some of the ADC's I have and at -20 dbFS the highest harmonic is -120 db from the signal at about -140 dbFS. If notching out the tone and looking at noise the SINAD would be 87 db referenced to the - 20 dbFS tone which is pretty much all noise. I don't have anything which is using close to a SOTA ADC chip. Lavry is only claiming -108 Sinad for - 20 dbFS signals as he has a much lower noise floor than my ADCs. View attachment 77997

Done well to get the noise floor to - 130dB, Rme Adi2 pro manages - 124dB.

However distortion let's it down. - 102dB at - 6dBfs. Rme is managing - 112dB at - 1dBfs. The Lavry would be doing worse again at - 1dBfs.

Edit: just found its 0.0012% for full scale so - 98dB