[Steve999]

This audiophile-type (?) DAC seems pretty bad. And it’s very expensive. Whether any of the problems are audible might be arguable, might be grist for the mill, I dunno.

https://www.audiosciencereview.com/...c200-signature-review-measurements-dac.44288/

For sure this seems like an extreme outlier, but maybe it’s why it might be good policy to never say never.

Personally, I just use the <$10 DAC chip in my phone, my receiver, my computer, etc. I generally don’t really get why people buy stand-alone DACs, period. I don’t really get why some “objectivist” hobbyists get so juiced about DACs.

 

[bigshot]

It isn’t great, but nothing there looks audible to me. I bet it would be difficult if not impossible to pick out in a proper ABX.

Numbers don’t tell the whole story with audibility. Ears do. But when you see two digits behind the decimal point on distortion and noise below -40dB, you aren’t going to hear anything out of line. It’s certainly not a clear difference by any stretch of the imagination. That DAC has everything it needs for your transducers and ears.

 

[gregorio]

to be fair, a lot of the time, your money buys you hardware that is in some way or another "faulty" and causes distortions, jitter, and other sometimes audible defects.

That’s tricky because some level of distortion, noise and jitter is inevitable and therefore a device which exhibits these artefacts is not necessarily faulty. For example, the minimum amount of noise created by a circuit is defined by Johnson/Nyquist (Thermal) Noise. As consumer DAC/amps are run at room temperature (rather than near zero Kelvin), the only way they could avoid producing thermal noise and being “faulty” is if they somehow broke the laws of physics, which is obviously impossible.

However, this thermal noise is well below the the threshold of audibility. The same is true of jitter and broadly true of distortion. Even cheap devices have managed to keep internal noise, jitter and distortion below or well below audibility for several decades. We could therefore argue that a modern (even modestly priced) device exhibiting these artefacts at audible levels is therefore faulty. However, do you have any examples of such devices?

the sole purpose should be to "reproduce the recording as it was made".
(and that opens up another can of worms, because people's ears are different, taste supposedly matters, and "would you want it to sound like it did to whoever mastered the recording or the way it sounded in the studio before it even entered the chain of microphones, amplifiers, recorders, etc.?

Indeed that is somewhat a can of worms, although in the vast majority of cases it’s an easy question to answer. Are there really any consumers who would want to hear a bunch of unedited, in-processed, in-mixed takes of individual instruments spread over serval days/weeks?

I feel like there's some similarity to hobbyists cyclists with 10kg overweight shelling out thousands of dollars to cut the weight of the frame by 500g.

I’m not sure, is it possible to sense a difference of 500g? A more accurate analogy would be a frame weight difference a hundred or so times below whatever the threshold is.

if you go over to "audiosciencereview", you'll find a disturbingly high number of DACs that actually produce audible distortions, noise, and jitter.

Really, where? I’ve seen almost none! While I applaud ASR for producing objective measurements, the conclusions are typically somewhat misleading. It’s rarely mentioned how those measurements relate to audibility and even when it is, the claimed audibility commonly has little relation to the real world. For example, it’s entirely possible to hear certain artefacts if we choose a very quiet passage of music and whack up the amplifier. However, that’s far from any real world scenario because if we actually tried to listen to the whole song/piece at that level we’d blow our eardrums and probably even our drivers. In many cases a simple bit of logic and knowledge suffices, REGARDLESS of individual hearing threshold differences. For example, what peak level do you typically listen with your HPs? Probably 90dB SPL or lower. Therefore (for instance) the level of jitter artefacts, say at -120dB, would be at -30dB SPL, which is about half the SPL required for a sound to even exist, so any question of audibility is irrelevant. ASR would list a DAC with jitter artefacts at say -140dB higher/better than one with jitter artefacts at -120dB and maybe rightly so but only “on paper”, it makes no difference to the actual sound that is output.

See my next response:

G

 

[gregorio]

This audiophile-type (?) DAC seems pretty bad. And it’s very expensive.

That’s a good example (of being terrible). However, it’s only terrible relative to the measurements of other DACs, how “terrible” is it in reality?

If we take THD for example, we’re used to seeing figures of 0.01% or lower but the worst case for this DAC (unbalanced RCA output) is 0.11% which is around 10-100 times worse than many other far cheaper DACs and can justifiably be described as “Very Poor” but this is still around 10 times below the threshold of audibility.

Or: The worst case dynamic range is 82.6dB, that puts it on a par with moderately priced CD players from the around the mid 1980’s, so truly terrible compared with modern DACs but that’s still a greater dynamic range than pretty much any commercial recording ever made.

The IMD really is troubling, worse case of -48.8dB is horrifically bad but even though this is within the threshold of audibility, it would still not be audible in many/most cases. It would only be audible with certain recordings, at high playback levels with especially accurate HPs/speakers and to those with good hearing and listening abilities.

As the reviewer stated: “This is the worst DAC I ever measured, regardless of the price and class. Period.” but even in this worst case, it’s unlikely most audiophiles would be able to ABX it under normal/reasonable listening conditions with their usual equipment and recordings.

G

 

[Easa]

On a different note, I would like a scientific or at least objective opinion on the topic of driver degradation. Specifically, the Fostex Biodyna driver. As it´s made from organic material, I imagine that it has a limited lifetime. Being used in the studio conditions, zero travelling, in quite stable temperature room with no humidity extremes, used daily for lets say 2 hours, how long will it take till the driver degradation / membrane damage will render them acoustically worse or straight unlistenable? I know that this is impossible to answer exactly, I just want to know if its in the time frame of years or decades.

I know that cars from 90s were almost always equipped with paper membrane speakers or paper + foam suspension, and they aged quite bad, especially the suspension will rot and decay over time, but those really are used in the harsh outdoor environment, so thats no surprise to me.

Anybody here uses TH900 as a daily driver since their release? Do they still sound the same?

 

[gregorio]

I know that this is impossible to answer exactly, I just want to know if it’s in the time frame of years or decades.

That’s impossible to answer even vaguely. There are just too many variables, what does “organic” mean, technically any compound containing carbon is organic, so diamonds and plastic are both organic. What are the organic parts coated with, how thick, how evenly applied, how resistant to humidity, etc.

One would expect an expensive product to be well made and last for many years or decades if well cared for, but in the audiophile world there can often be a disconnect between expectation and reality!

G

 

[71 dB]

Ideally you want the HP input impedance to be about 8 (or more) times higher than the output impedance of your amp.

As a rule of thumb rather than ideally, as for each headphone model what is or isn't ideal can be quite different and depends on its electro-acoustic properties. There are headphone models that can be driven from "any" output impedance without problems with sonic accuracy and there are headphone models suffering from some inaccuracies even when the 1/8 rule is being applied.

The output impedance of an amp driving headphones affects the frequency response, damping and distortion. Of these the frequency response criteria seems to be the most demanding: If you make it good enough, the other things are taken care of also. So, headphone manufacturers could calculate it as a spec data point for the customers using this equation:

Rout ≤ (0.06 * Zmax * Zmin) / (Zmax - 1.06 * Zmin),
​

were Zmax and Zmin are the maximum and minimum impedances respectively measured on a "head" to mimick the real life acoustic impedance seen by the headphones (headphone impedance changes a little bit depending on whether it is on head of not). This equation limits the frequency response error to 0.5 dB or less. So, if you measure Zmax = 200 Ω and Zmin = 50 Ω you get

Rout ≤ (0.06 * 200 * 50) / (200 - 1.06 * 50) = 600 / 147 = 4 Ω
​

If Zmax is just a little more than 6 % bigger than Zmin, the allowed amp output impedance becomes very large. If Zmax is exactly 6 % bigger than Zmin, this equation gives infinite amp output impedance and if Zmax is less than 6 % bigger than Zmin, the "allowed" output impedance is negative. These are of course non-sensical values and in these cases the limiting factor for the amp output impedance is the voltage division between the amp output impedance and the headphone: In order to limit the voltage division attenuation to 3-4 dB at most, we want at least 2/3 of the amp voltage over the headphone and at most 1/3 over the amp output impedance:

20 * log (2/3) ≈ -3.5 dB
​

This limits the amp output impedance to half of the headphone impedance at most. So, if you measure Zmax = 111 Ω and Zmin = 98 Ω (Zmax is about 13 % bigger than Zmin), you get

Rout ≤ (0.06 * 111 * 98) / (111 - 1.06 * 98) = 91.7 Ω

​

Since this violates the voltage division rule, the largest allowed amp output impedance is limited to Zmin/2 = 49 Ω. It is easy to show that the voltage division rule steps in whenever Zmax is at most 20 % bigger than Zmin. We can put all of this together by modifying the original equation into a form that avoids issues such as negative/infinite impedances and division by zero:

Rout ≤ (0.06 * Zmax * Zmin) / MAX( 0.14 * Zmin, Zmax - 1.06 * Zmin ),​

where MAX( a, b ) simply means picking the bigger value of a and b.

 

[JackarlTwo]

Anybody here uses TH900 as a daily driver since their release? Do they still sound the same?

SN: 35xx here, bought them back in early 2015 directly from Japan, so I guess they were made a couple of months before that. They have been used heavily for at least 5+ years, less so in the past couple of years. But I like to revisit them once in a while and they still sound the same as I remember. FWIW I have opened the cups a few times because of a cable issue, and the material inside still looks fresh.

 

[Ghoostknight]

Can higher Amp impedance have an effect on SQ (or objective measurements) with higher impedance headphones? ( 10ohm amp and 250ohm headphone )
beside the fact that lower amp impedance is generally better for low impedance headphones

 

[71 dB]

Can higher Amp impedance have an effect on SQ (or objective measurements) with higher impedance headphones? ( 10ohm amp and 250ohm headphone )
beside the fact that lower amp impedance is generally better for low impedance headphones

There aren't 250 Ω headphones so crazy 10 Ω output impedance can really affect the sound quality. If the output impedance is 20 Ω, that's when something mild might happen... ...but the probability is very low.

 

[WoodyLuvr]

There aren't 250 Ω headphones so crazy 10 Ω output impedance can really affect the sound quality. If the output impedance is 20 Ω, that's when something mild might happen... ...but the probability is very low.

Totally concur about the point you are making. The 1/8th rule has been heavily preached as the end all, be all for many years now but I have rarely experienced the touted drastic changes in transducer performance when the output to headphone impedance ratio is well under the mythical 1/8 mark. Yes, on a few occasions I could hear some loss of resolution and a noticeable decrease in impact in the lower frequency range but again very, very minor as you have stated. It may also really depend on the quality of the circuit design (and build) of the amp you are using as well.

A bit confused about your saying there are no 250Ω impedance headphones (I might be missing a joke or another conversation... if so, my apologies) as Beyerdynamic has a number of 250Ω impedance versions available for a number of their models (e.g. DT 990, DT 880, DT 1990, etc.), as well as the Sony MDRZX110NC which has an impedance of ~220-250Ω when the noise-cancelling circuit is activated (turned on).

 

[71 dB]

Totally concur about the point you are making. The 1/8th rule has been heavily preached as the end all, be all for many years now but I have rarely experienced the touted drastic changes in transducer performance when the output to headphone impedance ratio is well under the mythical 1/8 mark. Yes, on a few occasions I could hear some loss of resolution and a noticeable decrease in impact in the lower frequency range but again very, very minor as you have stated. It may also really depend on the quality of the circuit design (and build) of the amp you are using as well.

A bit confused about your saying there are no 250Ω impedance headphones (I might be missing a joke or another conversation... if so, my apologies) as Beyerdynamic has a number of 250Ω impedance versions available for a number of their models (e.g. DT 990, DT 880, DT 1990, etc.), as well as the Sony MDRZX110NC which has an impedance of ~220-250Ω when the noise-cancelling circuit is activated (turned on).

The 1/8 rule is from the time when portable listening wasn't such a thing it is today. Headphones didn't need to be that sensitive. People plugged them on their stereo amps (which typically have very high output impedance!). The headphones of the past tend to have more mechanical damping meaning they didn't need electric damping (provided by low output impedance).

In the era of portable audio and dedicated headphone amps sensitivity is important and it leads to headphones often having less mechanical damping. That's why in my opinion 1/16 or 1/20 is the new 1/8 rule.

Of course there are 250 Ω headphones, but not "crazy ones" needing lower output impedance than 10 Ω.

 

[WoodyLuvr]

The 1/8 rule is from the time when portable listening wasn't such a thing it is today. Headphones didn't need to be that sensitive. People plugged them on their stereo amps (which typically have very high output impedance!). The headphones of the past tend to have more mechanical damping meaning they didn't need electric damping (provided by low output impedance).

In the era of portable audio and dedicated headphone amps sensitivity is important and it leads to headphones often having less mechanical damping. That's why in my opinion 1/16 or 1/20 is the new 1/8 rule.

Of course there are 250 Ω headphones, but not "crazy ones" needing lower output impedance than 10 Ω.

Ah, ok gotcha sorry for the misunderstanding good man.

 

[MakeshiftApe]

I'm going to be honest, I really don't understand the science very much. Initially just from peoples' amp reviews I got the impression that every amp would drastically colour the sound, then as I read up more, I came to the conclusion that if an amp provided enough power that was all that mattered.

...Then more recently I discovered material about how the amp's output impedance can affect distortion. If I recall correctly, they were saying that actually the lower resistance the headphones themselves are, and the higher the resistance the amplifier, the more likely there is to be distortion? Though don't quote me on that.

Anyway, my question is - I'm currently running my K702, so 62ohm with 105db/V sensitivity, off my Behringer UMC1820 audio interface. I can't find definite specs on it but I did find one post on their forums where someone says the built in headphone amp has 16ohm output impedance. Is this a good match or not? My K702s get plenty loud and sound fantastic on it, but I'm wondering, am I missing anything and could an amp upgrade get even more out of these?

 

[71 dB]

I'm going to be honest, I really don't understand the science very much.

Nobody can know and understand everything. The fact that you are aware of this and you are willing to listen to those who do understand things is a good thing itself.

Initially just from peoples' amp reviews I got the impression that every amp would drastically colour the sound, then as I read up more, I came to the conclusion that if an amp provided enough power that was all that mattered.

Well, amps definitely should not drastically colour the sound! In many (but not in all) cases having enough power is indeed enough.

...Then more recently I discovered material about how the amp's output impedance can affect distortion. If I recall correctly, they were saying that actually the lower resistance the headphones themselves are, and the higher the resistance the amplifier, the more likely there is to be distortion? Though don't quote me on that.

Distortion happens, when the amp doesn't have enough control over the headphones. The smaller the output impedance of the amp is compared to the headphone impedance, the better control the amp has and the less distortion there is. However, this is not the only factor. Some headphone models "need" more control than others. That's why some headphones work nicely even when the output impedance of the amp is almost as big as the headphone impedance.

Anyway, my question is - I'm currently running my K702, so 62ohm with 105db/V sensitivity, off my Behringer UMC1820 audio interface. I can't find definite specs on it but I did find one post on their forums where someone says the built in headphone amp has 16ohm output impedance. Is this a good match or not? My K702s get plenty loud and sound fantastic on it, but I'm wondering, am I missing anything and could an amp upgrade get even more out of these?

62 Ω is minimum at 0 Hz, about 100 Ω is maximum at 20 kHz.

Rout < 100*62*0.06/(100-1.06*62) = 11 Ω.

This means your 16 Ω amp has a bit too large impedance for K702, but the sound may still be good. There shouldn't be issues with distortion. There might be too much treble. Don't worry about it if the sound is great in your opinion. Lower output impedance will change the sound very marginally.