Can we agree headphone amplifiers are solved?

[Fluffy]

I fully understand the spec thank you. That is all the info they published, but to take your point I can assure you they are publishing info that shows the product in its best light

You are ignoring the simple fact that the "professional" manufacturers you mention do not have access to any different or superior ADC chips over the semi professional manufacturers. The good semi professional products have maxed out the performance of these devices. There is nowhere for the "professionals" to go.

Knowledge goes beyond what's published in ASR. Just take a walk through the Texas Instruments, Analog Devices, AKM, ESS, Maxim etc ADC parts catalogues. They are full of very detailed spec sheets.

Not sure what you were trying to prove after all this. Do you support this statement:

ADC is not yet started, we gotta find a way to push it to the same performance as dac first then surpass performance of current dac offerings.

?

 

[March Audio]

ADC makers seem to play the biggest games with dynamic range. Also we know the -10 dbu with gain set to 20 db will get worse as signal level rises and if you use additional gain the noise floor will go up. They cherry picked a sweet spot for those specs.

Generally you aren't going to find ADC's with claims of better than -100 to -110 SINAD, or SNR of better than 110 db or dynamic range if properly tested of more than 115 to 120 db. Many of your most revered and expensive models won't even list actual specs. That in itself tells you something. They don't want to explain why the basic specs of the really expensive stuff is about the same as a good semi-pro device.

Oddly in the pro recording community a near consensus has been that ADCs are a solved problem for awhile now. Seems once you get most of your THD and noise specs to around -100 db they don't perceive any benefit. Other devices raise or lower signals as needed for them and if they have a 100 db wide window to work in they seem to think nothing more is needed. Now there are plenty who go for esoteric gear and don't agree with what I wrote, but a solid portion of those making their living in that field would agree with these statements. It seems to be the difference between an audiophile wanting everything below -120 db so there can never be the theoretical possibility of something being audible, and people who make their living deciding ,"if you need more than spurious signals lower than -100 db I've never heard it or run into it." Many audiophiles are chasing these numbers assuming they'll be better even though they've heard nothing wrong with what they've got. Of course pro recording people are just trying to make some cool music. They managed that when it was within the restricted confines of reel to reel tape going onto LP.

I've put up 8th generation DAC/ADC loopback copies of music. There isn't much audible degradation there, and I was using interfaces maybe a couple steps up from the lowest of them. Neither the ADC nor DAC had SOTA specs. I would have nothing against better ADCs, but I don't know if they'll gain us a whole lot audibly if anything. I think what Zoom and Sound Devices are doing with 32 bit float recording so you can never clip and never have a noise floor problem are actually more useful.

This is a very pertinent point. Nothing up front of the ADC, the mic, mic pre amp etc is going to have as good performance, so what is to be gained?

 

[March Audio]

Not sure what you were trying to prove after all this. Do you support this statement:

?

Simply trying to explain that there was no real basis for your assertion that "professional" equipment ADCs will perform any better than anything we have seen from the semi professional products tested here in ASR.

Yes I agree with John's statement in so far as ADC performance isn't as good as Dac performance. From an engineering perspective it would be nice to see it surpass 120dB SINAD (not implying that is the only important parameter) as we are with the best DACs.

However, just as with DACs, it is unlikely to provide any real world benefit.

 

[Fluffy]

Simply trying to explain that there was no real basis for your assertion that "professional" equipment ADCs will perform any better than anything we have seen from the semi professional products tested here in ASR.

As long as you are using only logic to assert that and not data, there is reason to question that assessment – however probable it might sound.

Come to think of this, arranging the measurement of a professional device by amir shouldn't be out of reach. Sound devices gear can be probably rented at any film gear renting place (not sure how it works in the US). Would be interesting to put these things to an actual test.

 

[solderdude]

I suppose a LOT more DA chips are produced than (High quality) AD chips.
DAC chips are 'scrutinized' by audiophiles as well and the quality of it can make or break sales of the devices they are put into.
LOTS more DA chips thus.
Manufacturers thus concentrate on getting this right with the lowest cost.
When designing a device that uses such chips and sells in the 10 thousands or more even a few cent is important.

AD chips are needed (sold) a lot less so less money goes into developping this. In the end its all about money.

 

[MRC01]

Lol solved? It's just started...
There is one performance metric I am still thinking has room for improvement (lower level output, like 50mV considering all the super sensitive and annoying IEM/headphone designs these days with awfully low impedance to boot). ...

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.

 

[March Audio]

As long as you are using only logic to assert that and not data, there is reason to question that assessment – however probable it might sound.

Come to think of this, arranging the measurement of a professional device by amir shouldn't be out of reach. Sound devices gear can be probably rented at any film gear renting place (not sure how it works in the US). Would be interesting to put these things to an actual test.

I'm not only using logic. I'm using real data. As I mentioned you can look for yourself at the ADC chip manufacturers data sheets to see where the performance limits are.

Yes it would be interesting to see these devices tested. My view is you won't see anything better than the good semi pro.

 

[solderdude]

It is solved in the sense that it is technically very possible to design and build this. At least to the point where it goes well below audible thresholds.
There are 'fixes' for amps that need to have less noise, with such a solution this too is solved for very high efficiency IEMs.
I would go as far and say it is even a solved issue even on a budget. At least it can be.
The fact that a lot of manufacturers build stuff that is considered below 'solved' is a practical matter and has nothing to do with manufacturers not being able to build to the set requirements. That is a financial decision not a 'technical solved one''
What definition 'solved' is depends on what thresholds must be met.

One needs to differentiate with what is technically possible (solved) and what manufacturers actually sell the public. This contains trade-offs.

 

[garbulky]

No, I don't think they are solved. Headphone amplifiers can improve by providing more power.
Other things they can do...how about some class A headphone amps?
My dream headphone amps are a set of monoblock class A balanced headphone amps.

 

[solderdude]

These can be build with todays components. What is not solved about this ?
Why would it be solved when a headphone amp can provide 10W and you need headphone cables with 2 connectors into 2 amps that have a topology that adds absolutely nothing in the case of headphones.
Your wishlist is not the 'solved' thing. Its a wish from you that is not available in the shops. A practical and financial thing.

How can amps be improved by providing more power when most headphones around are rated between 50mW and 200mW with a few exceptions ?

 

[Fluffy]

No, I don't think they are solved. Headphone amplifiers can improve by providing more power.
Other things they can do...how about some class A headphone amps?
My dream headphone amps are a set of monoblock class A balanced headphone amps.

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

 

[sergeauckland]

Simply trying to explain that there was no real basis for your assertion that "professional" equipment ADCs will perform any better than anything we have seen from the semi professional products tested here in ASR.

Yes I agree with John's statement in so far as ADC performance isn't as good as Dac performance. From an engineering perspective it would be nice to see it surpass 120dB SINAD (not implying that is the only important parameter) as we are with the best DACs.

However, just as with DACs, it is unlikely to provide any real world benefit.

The 'real world' benefit to me would be to be able to use an inexpensive interface as a measuring device. I completely agree that for the purpose most users put these interfaces to, i.e. recording audio, it will make no difference to any extent, but what annoys me is that a cheap interface can give a very low SINAD on the DAC section, but that's not matched on the ADC section, and I can't see any (technical) reason why this should be so.

S

 

[March Audio]

The 'real world' benefit to me would be to be able to use an inexpensive interface as a measuring device. I completely agree that for the purpose most users put these interfaces to, i.e. recording audio, it will make no difference to any extent, but what annoys me is that a cheap interface can give a very low SINAD on the DAC section, but that's not matched on the ADC section, and I can't see any (technical) reason why this should be so.

S

Absolutely, we are looking for the same thing

 

[kn0ppers]

As I mentioned you can look for yourself at the ADC chip manufacturers data sheets to see where the performance limits are.

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.

 

[bobbooo]

I think if the other version of IMD testing (using twin tones close together and equal amplitude) were used you would not see such an anomaly.

If true then this suggests that more complex test tones (unequal instead of equal amplitude in this case) can indeed reveal distortion that others show no signs of (at least in DACs), and so supports testing using music signals or a simulated spectral analogue thereof in the form of the standardized BS EN 50332-1 Program Simulation Noise.

 

[Blumlein 88]

If true then this suggests that more complex test tones (unequal instead of equal amplitude in this case) can indeed reveal distortion that others show no signs of (at least in DACs), and so supports testing using music signals or a simulated spectral analogue thereof in the form of the standardized BS EN 50332-1 Program Simulation Noise.

You can find various small anomalies in many places that this rather than that test uncovers. But even in the case being discussed here it wasn't a huge discrepency. It is simply the nature of how circuitry works, the transfer functions and the signal involved that we really don't gain much from the kind of testing you have in mind. The result you brought up here hints that maybe that device has something asymmetrical in the output stage that you wouldn't expect.

 

[bobbooo]

This is a DAC. We are talking about amplifiers. Have you found similar plots for amplifiers ?
besides also in DACs this non linearity is not missed using test tones. It is clear to see in the plots.
Why would there be distortions using music that would be missed using multitones for instance. What is so special about music ?
Music, just like test tones is just a voltage changing level over time. The difference with music is that it isn't as easy to analyze as there is no pattern.
Nulling can do this but as explained a couple of time there are disadvantages to this method.


It's fine to suspect or think or reason that music is special in some way. It would be better if you presented some actual evidence of this.
Evidence not being Df or sighted listening tests as these have more issues than measurements using artificial and defined test signals.

You could start with proving this by building an actual analog null tester which has different but not the same challenges as a digital null tester.
So you know I already did such about 30 years ago and till this day still believe that measurements using artificial signals has advantages over nulling and disadvantages. Measurements are enough. What is lacking is the knowledge of the general public in how to interpret all of those measurements and perhaps a basic grasp of audibility levels.

Nulling is a great method but you need to listen to it, amplified and not amplified in order to assess whether or not the null is detrimental to the sound.

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?

 

[bobbooo]

You can find various small anomalies in many places that this rather than that test uncovers. But even in the case being discussed here it wasn't a huge discrepency. It is simply the nature of how circuitry works, the transfer functions and the signal involved that we really don't gain much from the kind of testing you have in mind. The result you brought up here hints that maybe that device has something asymmetrical in the output stage that you wouldn't expect.

And that's the whole point - 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?

 

[sergeauckland]

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?

Because of the Superposition Principle. As far as I know it's never been found to be at fault.

S

 

[bobbooo]

Because of the Superposition Principle. As far as I know it's never been found to be at fault.

S

I've mentioned a couple of empirical examples where this doesn't seem to hold in practice for all DACs/DAPs. Even if these cases don't apply to amps, you need more than the superposition principle though - you additionally need distortion to be time and state-independent for a distortion measurement at time t to be the same as at t+1, and for a distortion measurement in state S1 to be the same as in state S2 (where S2 may be different to S1 possibly due to a signal consisting of a more complex superposition of sine tones such as music). As I understand this is only true of an ideal amp, not generally of all amps in practice, which can be subject to 'memory' effects.