[vid]

Sigh... I decided not to answer you after all of that because I already know where the discussion will head.
 
All you showed is that your EQ made decay shorter at certain frequencies, but that's not all there is to it. If you look closely at the initial frequency response wall (time 0), the frequency curve is still not perfect. That's what I meant by "not overcoming physical limitations".

 
I actually showed you a published study that had roughly 99% success at overcoming physical limitations with EQ - in this case, the fact that the drivers are right up to the ears, killing spatial cues. You ignored the study and went on about how HRTF is DSP, which (1) makes no sense, and (2) has nothing to do with the study.
 
My own examples also didn't report an overcoming of all physical limitations, nor were they trying to. They showed physical limitations being overcome, which is enough in my opinion to counter the argument that EQ can't overcome physical limitations. You might say the examples only show a masking of problems, yet, as the decay plots showed, actual fidelity was being improved; problems were removed rather than masked.
 
On the other hand, the proof you gave in support of your argument was along the lines of, 'I can't give proof, because this hasn't been tested, because everyone knows it's true'. That's a fine argument for the regular Head-Fi forums, but I don't think it helps us understand audio any better when we get down to it in the sound science forum.

 

[Bill-P]

   
I actually showed you a published study that had roughly 99% success at overcoming physical limitations with EQ - in this case, the fact that the drivers are right up to the ears, killing spatial cues. You ignored the study and went on about how HRTF is DSP, which (1) makes no sense, and (2) has nothing to do with the study.
 
My own examples also didn't report an overcoming of all physical limitations, nor were they trying to. They showed physical limitations being overcome, which is enough in my opinion to counter the argument that EQ can't overcome physical limitations. You might say the examples only show a masking of problems, yet, as the decay plots showed, actual fidelity was being improved; problems were removed rather than masked.
 
On the other hand, the proof you gave in support of your argument was along the lines of, 'I can't give proof, because this hasn't been tested, because everyone knows it's true'. That's a fine argument for the regular Head-Fi forums, but I don't think it helps us understand audio any better when we get down to it in the sound science forum.

1) I have repeatedly asked you to show me where in the article they mentioned EQ or equalization. You have not shown me that. In fact, the article itself was explicitly a case study about HRTF, which, again, for the umphteenth time, is NOT EQ. It's DSP. EQ is a part of DSP, but DSP is so much more than just EQ.
 
2) Your example showed less decay, but NOT overcoming physical limitations. Again, please note that you could NOT alter the way the transducer reproduced an impulse (as evidence at, again, time 0 where the natural frequency response still shows through). All you could do was shape how the impulse itself was sent to the transducer to be reproduced. This is a completely different issue. It's like saying that if you make a target as big as 1km wide, then everyone is shooting 100% accurately, and that they have somehow overcome their physical limitations. This is simply NOT true.
 
3) You read only what you wanted to read from what I gave you. I read the ENTIRE document you gave me and found ZERO mention of any EQ or equalization or ANYTHING related to that aside from the fact that DSP was mentioned. And again, DSP is NOT EQ. EQ is a PART of DSP. I have no idea how this has escaped you for so many times, but I have tried to make it as explicit as possible.
 
I honestly have predicted that this will happen, so I will make it clear now that this will be the last time that I will try to respond to you. Challenging a claim is fine, but repeatedly stating your opinions, ignoring what others give as possible proof, reading only what you want to, and then proceeding to claim that your side is right is not the way to discuss something, and I honestly would not want to entertain any such debate. Especially when I stand to gain nothing from being right.
 
Believe what you will, I guess, but it simply won't help preventing the truth from getting further from you.

 

[krismusic]

This is meant to be humorous by the way. I intend no rudeness.

 

[mvyrmnd]

Forgive my noobishness, but from first principles I can understand that EQ cannot override the inherent inabilities of hardware.
 
If your headphones can't produce a certain frequency, then how can shoving more of that frequency down the cable help?
 
Am I on the right track here or have a missed the point of the argument?

 

[krismusic]

I know very little technically unfortunately.
What I can say is that a headphone that is tuned to give too much of a frequency, such as the midbass of the IE8, is improved immensely by reducing that frequency.
I am less happy boosting a frequency but again it seems to help.
Forcing the driver to give more of a frequency that it is reluctant to reproduce.
Not so much trying to make the driver produce something that it cannot.
More tailoring the signal that it receives in order to emphasis parts of the signal and reduce the prominence of others.
It is my belief that the driver of the IE8 is very capable but has been tuned in a way that I do not particularly like.
It is my limited understanding that this is achieved by crossovers. As I cannot alter those, EQ offers a way to alter the signal that those crossovers receive and cheat or "correct " their effect.
The "correct" signal exists in the recording. The driver is equally capable of reproducing all frequencies within its range.
It's what happens in between that is up for grabs.

 

[castleofargh]

  Forgive my noobishness, but from first principles I can understand that EQ cannot override the inherent inabilities of hardware.
 
If your headphones can't produce a certain frequency, then how can shoving more of that frequency down the cable help?
 
Am I on the right track here or have a missed the point of the argument?

EQ never was intended to extend frequencies. that will never work and only make the headphone turn into crap. but now there are many headphones with less than 10db differences over most of the 50hz to say 14khz, and the never ending debate is about knowing if you can take one headphone and make it sound a lot like the other one?
the answer to that is clearly yes. but then some say that you can, not only have about the same sound signature, but also make the decay of each frequencies closer thanks to the EQ alone. and Vid showed us that it happened and that balancing a signature could also balance the decay differences. I also think I saw the opposite idea once, so I would guess that it might depend on a few parameters. but having the same king of headphone(dynamic, open back, about the same size...) should lead to several situations where you can almost turn one headphone into another one. it just seems logical to me. just like in some instances, you couldn't do it will all the DSP's in the world(tuning a porta pro into a hd800 might be slightly problematic ^_^).
so the right answer is probably "it depends". but the audio world doesn't like "it depends" conclusions ^_^.
 
on a very subjective level,I know that I often feel a little change in soundstage between 2 EQ settings. it proves nothing but it let me believe that EQ can do more than just make the guitar louder.

 

[castleofargh]

I know very little technically unfortunately.
What I can say is that a headphone that is tuned to give too much of a frequency, such as the midbass of the IE8, is improved immensely by reducing that frequency.
I am less happy boosting a frequency but again it seems to help.
Forcing the driver to give more of a frequency that it is reluctant to reproduce.
Not so much trying to make the driver produce something that it cannot.
More tailoring the signal that it receives in order to emphasis parts of the signal and reduce the prominence of others.
It is my belief that the driver of the IE8 is very capable but has been tuned in a way that I do not particularly like.
It is my limited understanding that this is achieved by crossovers. As I cannot alter those, EQ offers a way to alter the signal that those crossovers receive and cheat or "correct " their effect.
The "correct" signal exists in the recording. The driver is equally capable of reproducing all frequencies within its range.
It's what happens in between that is up for grabs.

the difficulty is that the "tuning" of an headphone can be electrical or physical. electrical can be done with EQ by just sending more or less volts at a given time, or inside the headphone with some resistors+capacitors (I doubt that's how it's done on a single dynamic driver like the IE8, and seeing as the impedance response is dead flat, it would suggest no such thing as caps or resistors to tune the FR.
but then there are all the physical conditions(type of driver, kind of membrane, overall impedance, shape of the inside casing, physical filters(even if it's only the metal grid at the end of the nozzle) and any game with pressure or reflection.
all being able to do what the other does in some ways, the main problem being not to have one counter the other too much. like the airflow making the sub bass to roll off on some designs because that's pure physics. and an EQ would be hard pressed to counter that correctly.

 

[Bill-P]

  so the right answer is probably "it depends". but the audio world doesn't like "it depends" conclusions ^_^.
 
on a very subjective level,I know that I often feel a little change in soundstage between 2 EQ settings. it proves nothing but it let me believe that EQ can do more than just make the guitar louder.

 
Well, it's not that the audio world doesn't like "it depends". I think Vid is the only one who is not liking that.
 
You'll find that many audiophiles bank on this "it depends" concept, like... when an amp is paired with a DAC to "synergize". Though "synergy" actually goes above and beyond EQ as you may find that most modern amps and DAC measure almost ruler flat in frequency response and also into inaudible range of distortions, but they still somehow sound different. I suspect this is due to subtle time domain behavior of certain electrical components, or more precisely, how these components deal with impulses sent to them.
 
In science, "it depends" means... "it's a part of a puzzle". And that's what I have been trying to say. EQ alone is just not enough. You can only make a headphone have "better frequency extension" by directly interfering with its physical structure... (change the membrane? Or change the filter in front of the membrane? Etc...). And on another note, some have opted to modify the HD800 acoustically to reduce its resonance. EQ can probably achieve the same effect, but... I have read from some who own the HD800 that when they EQ it that way, they can still hear some "harshness" in its high frequency response, whereas modding it acoustically pretty much gives it a super smooth characteristic.

And yeah, EQ can do more than just make the guitar louder or make everything but the guitar less loud. I don't deny that. It's a great tool for tone-shaping, especially when one has achieved a certain level of technicality with their entire headphone rig, and then all that is left is tone.
 
I think what you are hearing subjectively as a little change in soundstage is an effect of properly (or "improperly"?) reproducing the differences in levels between different frequencies that are representing... different physical sources of sound. In that case, the difference is what will cause your brain to perceive the soundstage being a certain way. For instance, an instrument playing softer will tend to sound more "distant", or "deeper" into the soundstage, whereas one that sounds louder will tend to sound like it's closer to you, or more "forward". So I'm not too surprised you'll hear a different soundstage with EQ. But I don't think EQ can do either "widening" or "focusing" of soundstage, since EQ deals mostly in the frequency domain. It won't really be able to fix left-right separation, which I think is the most important aspect of achieving a "wide" or "focused" soundstage. But then that's where the rest of DSP comes in.

 

[castleofargh]

  Well, it's not that the audio world doesn't like "it depends". I think Vid is the only one who is not liking that.
 
You'll find that many audiophiles bank on this "it depends" concept, like... when an amp is paired with a DAC to "synergize". Though "synergy" actually goes above and beyond EQ as you may find that most modern amps and DAC measure almost ruler flat in frequency response and also into inaudible range of distortions, but they still somehow sound different. I suspect this is due to subtle time domain behavior of certain electrical components, or more precisely, how these components deal with impulses sent to them.
 
In science, "it depends" means... "it's a part of a puzzle". And that's what I have been trying to say. EQ alone is just not enough. You can only make a headphone have "better frequency extension" by directly interfering with its physical structure... (change the membrane? Or change the filter in front of the membrane? Etc...). And on another note, some have opted to modify the HD800 acoustically to reduce its resonance. EQ can probably achieve the same effect, but... I have read from some who own the HD800 that when they EQ it that way, they can still hear some "harshness" in its high frequency response, whereas modding it acoustically pretty much gives it a super smooth characteristic.

And yeah, EQ can do more than just make the guitar louder or make everything but the guitar less loud. I don't deny that. It's a great tool for tone-shaping, especially when one has achieved a certain level of technicality with their entire headphone rig, and then all that is left is tone.
 
I think what you are hearing subjectively as a little change in soundstage is an effect of properly (or "improperly"?) reproducing the differences in levels between different frequencies that are representing... different physical sources of sound. In that case, the difference is what will cause your brain to perceive the soundstage being a certain way. For instance, an instrument playing softer will tend to sound more "distant", or "deeper" into the soundstage, whereas one that sounds louder will tend to sound like it's closer to you, or more "forward". So I'm not too surprised you'll hear a different soundstage with EQ. But I don't think EQ can do either "widening" or "focusing" of soundstage, since EQ deals mostly in the frequency domain. It won't really be able to fix left-right separation, which I think is the most important aspect of achieving a "wide" or "focused" soundstage. But then that's where the rest of DSP comes in.

well I'm not sure I agree with the amp part. to me apart from the amp being ok and having the proper specs to drive the headphone, they are very unlikely to actually sound different in a matched levels blind test. but there are all kinds of amps with all kind of specs so I agree that we in practice reach different sounds more often than not. but we shouldn't ^_^.
 
 
now I totally agree with me hearing a change in soudstage(arghhhhhh!!!! headstage) for the reasons you explain. it's just that I was on the other side of the debate a few years back, and my main reason was that from one IEM to another I had a very different headstage. so I thought that changing only the FR could do nothing about it and thus couldn't be enough to make 2 headphones to sound almost the same. 
but now that I realized that my sense of headstage could be impacted strongly by FR changes and stereo separation(crosstalk or crossfeed), I accept the possibility that with the right EQ and the right stereo separation, I'm likely to get more similarities even between 2 IEMs.
to a point of course. I tried several times to get my er4 to sound like my IE80 and my IE80 to sound like my ER4 for fun, and each time I ended up like this:

 
anyway I would have sold most of my IEMs if I couldn't EQ them a little. so EQ is a very significant part of my audio life. and I would probably be a very unhappy and forever changing audiophile if I didn't have the EQ margin of maneuver.

 

[Bill-P]

A matched level blind test inherently have some issues that I don't think have been mentioned too often by its advocates. Here are some that I see can potentially happen:
 
1) Blind test relies on a person's subjective impressions, and... well, subjective impressions are not always reliable.
 
2) People have different sensitivity, at different frequencies, and also at different amplitudes. Fletcher-Munson gives a rough idea (not the whole picture) of what this is like on average. In that case, one level matched for one person may not be someone else's "comfort zone", and you'd have to level-match for each participant in that case. Also due to this, even the headphone used for the test would matter because of people's tonal preferences. Throwing an HD800 at someone who is super sensitive to treble will simply have the person preferring whichever amp has a high frequency roll-off, which is not ideal.
 
3) Expectation bias also comes into play here. Without actually seeing a switch taking place, someone may think in their head that no switch actually occurred, and then they'd be inclined to think there should not be any difference. This is the same problem as with a non-blind test where if someone sees a switch taking place, then they would expect a difference. You really can't have it both ways here. 

 
4) The gears involved need to be able to show the difference. For instance, if the DAC or music source sucks... and has problems with its power supply (computers typically have problems with their power supplies, at least as far as noise and ripple goes), then this will feed into the amp and gets amplified. If the music involved is not something the participant is familiar with, nor is it something to their taste, they would also tend to be indifferent about it (same problem as #3). I mean... if I don't like rap, then all rap would sound the same to me, right? Ignoring the DAC and music choice altogether, the headphone needs to also be able to reveal the differences. HD800 comes into play here, but... I think you'll see #2 comes into play as well. You really can't please everyone here.
 
#3 is the biggest obstacle that has not been overcome IMO. Even for someone from here, if their expectation is for there to be no difference, then they would hear no difference no matter how the test is conducted.
 
Now... I think it makes sense from an electrical standpoint that there should be a difference. Different amps have different topologies, and they also measure... well, different on some levels. Look up the op amp measurements done by none other than the famous O2's creator. He noted the differences, and yet proceeded to claim that they are "inaudible", but... again, there are differences.
 
It is possible to create two amps that measure identically using different circuitries, and I'd say... if they indeed measure completely identical, with completely identical specifications, then one may claim the amplifiers are very unlikely to sound different, and I'd be inclined to agree with that. But that's not really how things look.
 
Is the O2 a giant killer then? Well, not really. Measurements are one thing. Subjective impressions... are the other. Note #2 above. People have different tastes, so a headphone when played at its normal and natural frequency response may not be exactly to someone's taste. Though... if it's close enough, then just a "small nudge" will probably make it perfect for them. That's where EQ comes in, but then if you can have an amp that takes care of EQ, then you would not have to be so reserved about using no EQ at all anymore.
 
So are all high-end amps doing EQ? Well, again, not really. I think what they are doing is similar to a very complex DSP that will sum up to give no apparent EQ effect, but yet when checking square wave and impulse responses may show some interesting "shaping" of the signal that they are given. This may "act" like EQ from a subjective standpoint, but it's much more complicated than just EQ.

 

[552388]

EQ can not solve all problem, it will also bring some phase distortion which is really bad to the sound quality

 

[vid]

1) I have repeatedly asked you to show me where in the article they mentioned EQ or equalization. You have not shown me that. In fact, the article itself was explicitly a case study about HRTF, which, again, for the umphteenth time, is NOT EQ. It's DSP. EQ is a part of DSP, but DSP is so much more than just EQ.
 
2) Your example showed less decay, but NOT overcoming physical limitations. Again, please note that you could NOT alter the way the transducer reproduced an impulse (as evidence at, again, time 0 where the natural frequency response still shows through). All you could do was shape how the impulse itself was sent to the transducer to be reproduced. This is a completely different issue. It's like saying that if you make a target as big as 1km wide, then everyone is shooting 100% accurately, and that they have somehow overcome their physical limitations. This is simply NOT true.
 
3) You read only what you wanted to read from what I gave you. I read the ENTIRE document you gave me and found ZERO mention of any EQ or equalization or ANYTHING related to that aside from the fact that DSP was mentioned. And again, DSP is NOT EQ. EQ is a PART of DSP. I have no idea how this has escaped you for so many times, but I have tried to make it as explicit as possible.
 
I honestly have predicted that this will happen, so I will make it clear now that this will be the last time that I will try to respond to you. Challenging a claim is fine, but repeatedly stating your opinions, ignoring what others give as possible proof, reading only what you want to, and then proceeding to claim that your side is right is not the way to discuss something, and I honestly would not want to entertain any such debate. Especially when I stand to gain nothing from being right.
 
Believe what you will, I guess, but it simply won't help preventing the truth from getting further from you.

 
1. I'm not living in a digital world, so my HRTF isn't DSP and neither is yours. Let it go, or explain why you keep saying this, because your rejection of the proof against your argument apparently relies on "HRTF = DSP" being true. As for where they mention equalization in the study, it's in the methods section - but you already knew this, because you'd also said you don't consider equalization via digital filters to be equalization.
 
2. I posted several examples - which one are you addressing? Also, you're trying to switch this around to be about what the driver is doing, but since we don't have direct access to the sound produced by the driver, this isn't the way to go. What matters is the point at which we do have access to the sound - the point at our eardrum; and at that point, we're looking at contributions to the sound from multiple sources, not just the driver. Thus, if we want to improve the perceptible sound of the headphones, we don't care about whether the driver is moving this way or that way per se, we care about what the sound is like once it reaches our eardrum, and that's what my examples were concerned with.
 
2.5. I think at this point it would help if you defined physical limitation from your point of view, and what it would take to prove that such a limitation has been overcome. Right now, it seems - to take your analogy - you're giving a moving target.
 
3. You gave me a link to a forum where some were saying EQ was bad and some were saying EQ was good. I read from this that if someone makes an argument for either of those cases, they'll need proof more solid than "everyone knows it's true". Again, digital filters were mentioned in the study, as you know, but your moving target of an argument dodged this by inventing the rule that digital filters don't count as EQ.
 
As it stands, you've literally given no proof for your argument, and the grounds on which you've dismissed the counterargument are sketchy at best. While dropping out of the conversation is fine and all in some cases, I don't think anyone should bring an argument to a science forum in the first place if they're not willing to defend it.

 

[Head Injury]

Quote:

  A matched level blind test inherently have some issues that I don't think have been mentioned too often by its advocates. Here are some that I see can potentially happen:
 
1) Blind test relies on a person's subjective impressions, and... well, subjective impressions are not always reliable.

 
Blind tests are much less subjective than sighted tests, so I don't see how you can really object to them on this basis. You can always do ABX tests instead, which take subjective impressions out of the picture even more and asks the listener to only compare X to A and B.
 
Quote:

 2) People have different sensitivity, at different frequencies, and also at different amplitudes. Fletcher-Munson gives a rough idea (not the whole picture) of what this is like on average. In that case, one level matched for one person may not be someone else's "comfort zone", and you'd have to level-match for each participant in that case. Also due to this, even the headphone used for the test would matter because of people's tonal preferences. Throwing an HD800 at someone who is super sensitive to treble will simply have the person preferring whichever amp has a high frequency roll-off, which is not ideal.

A blind test doesn't require every listener to be in the room at the same time, so volume can be adjusted on a per-user basis. Overall volume can be adjusted digitally to a listener's comfort level once level matched. As for the headphone issue, we have to establish that there's an audible roll-off first and we can do this with an ABX test which again asks the listener to identify rather than describe the amps. Why are you so hung up on finding subjective preferences? We're interested in finding audible differences.
 

3) Expectation bias also comes into play here. Without actually seeing a switch taking place, someone may think in their head that no switch actually occurred, and then they'd be inclined to think there should not be any difference. This is the same problem as with a non-blind test where if someone sees a switch taking place, then they would expect a difference. You really can't have it both ways here. 

You can in an ABX test. Because, once again, in the world of amps when most everything actually sounds the same, we have to establish that there are audible differences at all before we worry about which one people like more.
 

4) The gears involved need to be able to show the difference. For instance, if the DAC or music source sucks... and has problems with its power supply (computers typically have problems with their power supplies, at least as far as noise and ripple goes), then this will feed into the amp and gets amplified. If the music involved is not something the participant is familiar with, nor is it something to their taste, they would also tend to be indifferent about it (same problem as #3). I mean... if I don't like rap, then all rap would sound the same to me, right? Ignoring the DAC and music choice altogether, the headphone needs to also be able to reveal the differences. HD800 comes into play here, but... I think you'll see #2 comes into play as well. You really can't please everyone here.

Most differences between DACs are smaller than the ones between amps, even audibly transparent amps, and their distortion isn't affected by the load itself like an amp's. If there are differences to reveal, virtually any DAC will reveal them. The headphone is important, but the HD800 isn't the one to use. It has high impedance, making it an easy load for the amp, and it's open back which will allow ambient noise to bury the tiny noise and distortion differences in amps. Probably a better choice would be a high isolation low impedance IEM, which will present a more difficult load and increase the amp's distortion. A multi-BA IEM can be used to test for differences caused by output impedance as well if desired.
 
On a personal level, because amp performance is so dependent on the load, the headphone to use is one you will be listening with on a daily basis. It doesn't matter how an amp performs with a 300 ohm load when your headphones are 32 ohms, for example.
 

 #3 is the biggest obstacle that has not been overcome IMO. Even for someone from here, if their expectation is for there to be no difference, then they would hear no difference no matter how the test is conducted.  
Now... I think it makes sense from an electrical standpoint that there should be a difference. Different amps have different topologies, and they also measure... well, different on some levels. Look up the op amp measurements done by none other than the famous O2's creator. He noted the differences, and yet proceeded to claim that they are "inaudible", but... again, there are differences.
 
It is possible to create two amps that measure identically using different circuitries, and I'd say... if they indeed measure completely identical, with completely identical specifications, then one may claim the amplifiers are very unlikely to sound different, and I'd be inclined to agree with that. But that's not really how things look.
 
Is the O2 a giant killer then? Well, not really. Measurements are one thing. Subjective impressions... are the other. Note #2 above. People have different tastes, so a headphone when played at its normal and natural frequency response may not be exactly to someone's taste. Though... if it's close enough, then just a "small nudge" will probably make it perfect for them. That's where EQ comes in, but then if you can have an amp that takes care of EQ, then you would not have to be so reserved about using no EQ at all anymore.
 
So are all high-end amps doing EQ? Well, again, not really. I think what they are doing is similar to a very complex DSP that will sum up to give no apparent EQ effect, but yet when checking square wave and impulse responses may show some interesting "shaping" of the signal that they are given. This may "act" like EQ from a subjective standpoint, but it's much more complicated than just EQ.

Expectation bias is why we ask the ones who do hear a differences to test themselves, because there's no point testing Sound Science regulars if they don't think there will be a difference. Again, ABX can help with that slightly and reveal differences we might not have expected, but the results will still be more conclusive coming from someone who's actively trying to prove the differences exist. But anyway, I think the test you're looking for is an ABX test, not a regular AB blind test. Or we can do it the easy way and record the output of two amps and compare the results, but this bypasses the headphone load which will greatly affect the amp's output.
 
It makes no sense from an electrical standpoint that there should be audible differences. Exactly how revealing do you think our ears are?
 
The O2 isn't a giant killer. It's so far past audible noise and distortion with most loads that it's pointless. Probably one of the $80 Fiio amp/DACs is the giant killer, they don't measure as well but will be audibly transparent most of the time anyway. Probably more linear than most kilobuck amps.
 
Most amps are perfectly flat from 20 Hz to 20 kHz, and most competent solid state amps have distortion specs too low to hear at any frequency. With a low output impedance, there's no EQ effect at all. There's exceptions of course, but bringing it up in a discussion of EQ is more of a distraction than anything.

 

[Rajikaru]

   
3) Expectation bias also comes into play here. Without actually seeing a switch taking place, someone may think in their head that no switch actually occurred, and then they'd be inclined to think there should not be any difference. This is the same problem as with a non-blind test where if someone sees a switch taking place, then they would expect a difference. You really can't have it both ways here. 

 
#3 is the biggest obstacle that has not been overcome IMO. Even for someone from here, if their expectation is for there to be no difference, then they would hear no difference no matter how the test is conducted.
 
 

 
If a listener can't tell the difference either way, then for all practical purposes, it's inaudible. It sounds like you're just trying to compensate for something so insignificantly small by making it out to be much bigger than it actually is. 

 

[Bill-P]

  Blind tests are much less subjective than sighted tests, so I don't see how you can really object to them on this basis. You can always do ABX tests instead, which take subjective impressions out of the picture even more and asks the listener to only compare X to A and B.
 
Blind tests are subjected to the same expectation bias as sighted tests as I mentioned somewhere in that post. When someone does not see the actual switching, the expectation is that... there was no switching that actually occurred, and they may tend to think it should be the same as before. I'm not necessarily objecting to it as much as raising some potential issues that I think would affect the test result and causes the test to be inconclusive, as it has for the past while. ABX test also has its own issues, because the subject is under pressure rather than relaxing and just giving what they think. They have to strain to compare now, which is not ideal.
 
A blind test doesn't require every listener to be in the room at the same time, so volume can be adjusted on a per-user basis. Overall volume can be adjusted digitally to a listener's comfort level once level matched. As for the headphone issue, we have to establish that there's an audible roll-off first and we can do this with an ABX test which again asks the listener to identify rather than describe the amps. Why are you so hung up on finding subjective preferences? We're interested in finding audible differences.
 
Digital volume should not be adjusted because you may cause dynamic range to be compressed or some other oddities depending on the software setup. It is best for this to be done in the analog domain... But that's not the real issue I wanted to raise. The issue is that people simply hear things differently. Whether a roll-off is audible or not can be determined using an objective measurement of the frequency response, and some amps on the higher-end do intentionally EQ high frequencies down. If the amp does not do this then some DACs do. Case in point: look up measurements for the iQube V3 and AMB Gamma2.
 
I'm not hung up on finding subjective preferences, but I'm trying to explain how personal preference can affect the test results of a comparison test. Again, this is purely subjective, and is not along the same line as audibility. It's not like we're asking listeners to determine whether or not they can hear something playing in a completely silent room.
 
You can in an ABX test. Because, once again, in the world of amps when most everything actually sounds the same, we have to establish that there are audible differences at all before we worry about which one people like more.
 
In reality, though, amplifiers should not sound the same. They have measurable differences. Whether the differences are audible or not are tested using studies that rely on subjective impressions, which... as I have mentioned already, is not reliable. Objectively, there should be a difference. Unless you're saying measurable differences somehow don't count.
 
Most differences between DACs are smaller than the ones between amps, even audibly transparent amps, and their distortion isn't affected by the load itself like an amp's. If there are differences to reveal, virtually any DAC will reveal them. The headphone is important, but the HD800 isn't the one to use. It has high impedance, making it an easy load for the amp, and it's open back which will allow ambient noise to bury the tiny noise and distortion differences in amps. Probably a better choice would be a high isolation low impedance IEM, which will present a more difficult load and increase the amp's distortion. A multi-BA IEM can be used to test for differences caused by output impedance as well if desired.
 
People have claimed to hear the differences in DACs, though, so... I don't think it's as clear cut as saying the differences are small. Also some DACs intentionally shape the impulse a certain way. Please look up the AMB Gamma2 and "Minimum Phase" filter.
 
On that note, high impedance does not necessarily mean "easy load". It only means "low current" needed, and not as much interplay with output impedance. Amps that do not have the voltage swing necessary for a certain loudness level may still audibly clip or distort. Solid-state amplifiers relying on op-amps will also run into this if power supply quality is not taken into consideration. When all else are equal, though, the differences may still exist due to noise level, as well as the op-amp's inability to cope with capacitive load, which most headphones would present due to their structure (cable has capacitance, and also the diaphragm construction has capacitance). You'd be surprised to know that most op amps don't really play well with capacitance in the output, even when said capacitance is extremely small.
 
Back on the discussion, I don't think IEM is the way to go, because not everyone can fit IEM. Also because not all IEM can stand up to the technicality of something like the HD800, so they may not be able to resolve as much as the HD800 can. Stax SR-009 would be my pick, but then it doesn't work with many amplifiers.
 
On a personal level, because amp performance is so dependent on the load, the headphone to use is one you will be listening with on a daily basis. It doesn't matter how an amp performs with a 300 ohm load when your headphones are 32 ohms, for example.
 
Agreed. The test should be conducted with a headphone the user is familiar with. However, you wouldn't want to use a tube amp with 120-Ohm output impedance on a 32-Ohm headphone. I have tried. The distortion goes way beyond "more bass and warm sound".
 
Expectation bias is why we ask the ones who do hear a differences to test themselves, because there's no point testing Sound Science regulars if they don't think there will be a difference. Again, ABX can help with that slightly and reveal differences we might not have expected, but the results will still be more conclusive coming from someone who's actively trying to prove the differences exist. But anyway, I think the test you're looking for is an ABX test, not a regular AB blind test. Or we can do it the easy way and record the output of two amps and compare the results, but this bypasses the headphone load which will greatly affect the amp's output.
 
But Sound Science regulars are the ones who think there is no difference.
 
At least that is the vibe I'm getting from here.
 
When you say something like "all amps sound the same or the differences are not audible", and someone say "I can hear amp X is warmer than amp Y" and then the actual objective measurements tend to show minute differences, as a scientist, who am I inclined to believe more, right?
 
It makes no sense from an electrical standpoint that there should be audible differences. Exactly how revealing do you think our ears are?
 
It makes sense from an electrical standpoint. In fact, allow me to show you an article on measurements of different types of capacitors...
 
I think our ears are as revealing as they can be. Tests that ask people to replay their subjective impressions should be taken as a yard stick rather than as facts because... again, people are really unreliable.
 
When I can measure a difference, and my ears cannot pick those up, that means the problem lies with me (or perhaps my headphone/speaker), and not with the amp or DAC itself.
 
The O2 isn't a giant killer. It's so far past audible noise and distortion with most loads that it's pointless. Probably one of the $80 Fiio amp/DACs is the giant killer, they don't measure as well but will be audibly transparent most of the time anyway. Probably more linear than most kilobuck amps.
 
You would think it's past audible noise and distortion, but I can show you otherwise. It would still interact with IEMs... which should be easy loads, and even when its output impedance is at 0.5 Ohm. That should tell you 
 
Most amps are perfectly flat from 20 Hz to 20 kHz, and most competent solid state amps have distortion specs too low to hear at any frequency. With a low output impedance, there's no EQ effect at all. There's exceptions of course, but bringing it up in a discussion of EQ is more of a distraction than anything.
 
Most amps are actually only "relatively" flat from 20Hz to 20KHz. Zoom in on the measurements and you'll see bumps. Notice they usually present frequency response graphs in massive scales. This is actually more due to limitations of the measuring equipment.
 
But anyway, even when frequency response is not taken into account, distortions and noise are still different, so the amps should, theoretically, be different. Saying they don't sound different is just a subjective impression, and not necessarily science IMO.
 
Unless I am mistaken, science is the pursuit to ultimately understand and comprehend a phenomenon, and not a pursuit to dismiss something based on subjective perception.
 
On that note, I only brought it up to further expand because some may tend to think high-end amps do EQ, but that's not the whole truth.

 
Sorry, I couldn't split things out, so I responded in bold.
 
Also link to capacitor measurement as mentioned above:
http://diyaudioprojects.com/mirror/members.aol.com/sbench102/caps.html