Head-Fi.org › Forums › Equipment Forums › Sound Science › If it graphs bad then it is bad; yes or no?
New Posts  All Forums:Forum Nav:

If it graphs bad then it is bad; yes or no? - Page 6

Poll Results: If it graphs bad

 
  • 43% (21)
    Then it is bad
  • 43% (21)
    Then it is probably, but not certainly, bad
  • 12% (6)
    We don't need no steenkin' graphs, hombre!
48 Total Votes  
post #76 of 130
Thread Starter 
Quote:
Originally Posted by jaddie View Post

Here are 3 50Hz square waves.  Tell us how they would differ in sound.  

Edit: Also, tell us how they would relate to the overall sound quality of the device that produced them.

 

You can refer to them as Top, Middle and Bottom.

 

 

 

I'd be delighted with any of those. If you'd spent a few minutes looking at sq wave graphs  - intelligently, of course - you'd realize that they'd all be pretty damn exceptional! If you gave me these three headphones then I certainly wouldn't be able to match the sets with graphs by listening. Even the worst of these make the best of the real graphs I posted look awful. I think you're confusing the fact that you haven't managed the minimal accomplishment of knowing what a reasonable real world square wave graph looks like with the idea that doing so is impossible. No.

 

Now take a look at the graph for the 668B and read what I wrote: I *would* be pretty damn sure that I could tell the Fantastic Three from the 668B. But from the HD25s with their excellent bass, end if the graph isn't as "good"? I don't know. (Especially as you have removed large chunks of information by not having a scale!)


Edited by scuttle - 2/12/13 at 3:40pm
post #77 of 130
Quote:
Originally Posted by scuttle View Post

I'd be delighted with any of those. If you'd spent a few minutes looking at sq wave graphs  - intelligently, of course - you'd realize that they'd all be pretty damn exceptional! If you gave me these three headphones then I certainly wouldn't be able to match the sets with graphs by listening. Even the worst of these make the best of the real graphs I posted look awful. I think you're confusing the fact that you haven't managed the minimal accomplishment of knowing what a reasonable real world square wave graph looks like with the idea that doing so is impossible. No.

Now take a look at the graph for the 668B and read what I wrote: I *would* be pretty damn sure that I could tell the Fantastic Three from the 668B. But from the HD25s with their excellent bass, end if the graph isn't as "good"? I don't know. (Especially as you have removed large chunks of information by not having a scale!)

I'm sure you'd be happy with these if they were headphones, but that's not what I asked.

I'm asking you to please comment on how each of these represent some aspect of sound quality. They are simple examples. If you can't do it it's simple examples, how could you tell what a really messed up square wave, like those found in headphone measurements, would indicate?
post #78 of 130
Duplicate post...sorry.
post #79 of 130
Hmmm. No reply. There's even been enough time to google for the answer.

Ok, here are a couple of hints: the top one is a theoretically "perfect" square wave. This would be the original test signal used to stimulate a device under test. The middle one has had a common modification, a change resulting from imitations found in nearly every audio device, applied to it.

What is that change, and how would it affect sound quality?

This is pretty simple square-wave test interpretation. I would think someone claiming to have mastered the art of making judgments of sound quality based on square wave graphs would get this in seconds.
post #80 of 130
post #81 of 130

Thanks for that, ultrabike.  

 

I'm posting one more square wave shot, just to make the point. 

 

 

 

I'm fairly sure that even though this wasn't produced by a set of headphones, it would be recognized as a fairly "messed up" square wave.

 

I'm still inviting scuttle to interpret these for us, since he claims to have that ability.  

 

What we have are now one undistorted 50Hz square wave, and three distortions, all of which began as the perfect square wave and had something done to them that can easily occur in audio devices, including transducers.  

 

I'll make this very very simple.  Of the three distorted waves, which one sounds the "worst", and how would a device that produces that result sound with real music?  I'm not asking for general subjective terms like "bad" or "terrible", but something a little more definitive like "less bass", "more treble", "shrill", or some other adjective.  

 

Perhaps if we get some answers we might even get to hear audio through these processes...that's if there's any interest anymore.

post #82 of 130

Which of the two bottom filtered 60 Hz square waves do you think is easier to tell apart from the top (unfiltered) one in a listening test ?

post #83 of 130
Quote:
Originally Posted by jaddie View Post

Beg to differ here.  If you're going for neutrality, the targets are well known.  If not, well, subjectivity is what marketing is all about.

 

Exactly if you are going for a certain sound signature, then there is a "perfect" frequency response - for you. You should really say "flat" rather than neutral - suggesting a flat headphone is "neutral" suggests it is presenting sound as the producer/artist heard it, or intended it, which assumes two things -

 

a. That the producers studio monitors were also completely flat and the damping/construction of their room was perfect. Uncommon today, increasingly impossible the further back in time the album you are listening to was produced.

b. That they made absolutely no compensation in their choices for the average commerically available speaker.

 

In the cases of most albums neither is true. The producer did not expect you to be listening in flat speakers and compensated for the average speaker response of whatever time the album was made, and even at the point of production was probably not using a perfectly flat room/gear themselves.

 

This is before you even take into account that what the manufacturer used to define flatness would have been based on a HRTF different to yours. 

 

This arrogance creeping into head fi that there is an objective "right" and "wrong" sound signature to headphones, and some kind of elitist hierarchy of listeners who like "neutral" or "coloured" headphones needs to stop. It's not based on anything remotely rational.


Edited by EddieE - 2/13/13 at 1:26am
post #84 of 130

Further to the above -

 

I personally do like –mostly- flat headphones. For me about +5db emphasis between 1k-2k (as the modern Stax Lambda models display) sounds perfectly natural to me. If people are looking for “neutral” they need to trust their ears. I’m used to hearing pianos, violins, guitars and many other instruments in real life, and we are all used to human voices. If headphones with this subtle tweak on “flat” sound real to me, then I am not going to second guess myself because a graph tells me that is a diversion from “flat”.

 

But ignoring this, the reason I go for –mostly- flat headphones is because most of my favourite genres feature acoustic instruments and human voices.

 

What if I listened exclusively to electronic music genres? This music is often engineered to be played in clubs, venues which tend to massively boost both bass and treble on their sound systems. The producer of that music knew that the music would not be played on a “flat” system so they compensated when they produced it. They will likely sound a little bass-shy and a little lacking in energy when played back on a flat system. If that sort of music was my main interest, I would go for headphones with a V-shaped response, and there would be nothing wrong with that.

 

There are other reasons some people might prefer a bit of emphasis in the bass on headphones. Headphones exclude your body from the listening experience and a lot of the bass response comes through the body and not the ears. On speakers you both “feel” bass in your organs and it conducts through your bones to your ears as well. That is why bass will often sound louder and fuller on speakers (most of which start a gradual rolling off around 100hz) than they will on headphones that achieve an even response right down to the threshold of hearing.

 

There are lots of reasons this idea of a "right" and "wrong" frequency response is nonsensical. While there is probably a very large cross-over area in the huge venn-diagram that comprises everyone's personal tastes, and there is such a thing as a clearly "bad" set of headphone measurements - there is no single "right" kind of sound signature. 

post #85 of 130
Thread Starter 
Quote:
Originally Posted by jaddie View Post


I'm sure you'd be happy with these if they were headphones, but that's not what I asked.
 

 

Ok: my point is that your rhetorical question is a silly one. You are trying to prove that examining 50Hz square waves is never worthwhile by showing that it is not in a particular case where all three waves are (possibly, because you left the scale off) extremely good. This is fundamentally appalling logic: a heuristic doesn't have to be useful in 100% of cases to be worthwhile. For example, not everything that is poisonous is so labelled, but you are an idiot to drink from a bottle marked "poison" if you don't know what it is.

 



I'm asking you to please comment on how each of these represent some aspect of sound quality. They are simple examples. If you can't do it it's simple examples, how could you tell what a really messed up square wave, like those found in headphone measurements, would indicate?

 

Who says these are "simple" examples to compare? They are not - these signatures would be very hard to tell apart by listening. They have simple shapes and were easy for you to draw - but that is NOT the same thing! You're making a whole raft of foolish assumptions here: you are drawing waves that will sound very alike to the ear because you don't understand how the test is used and then trying to claim that because people can't tell them apart the test isn't useful - but what is actually being measured is your own ignorance and inability to understand a simple explanation.

 

Now, again, all of the three graphs you gave are good. Here's what a bad graph looks like:

 

 

700

 

 

See how different they are? Look at the pulse aligned with the q in "Square Wave": the response does NOT stay in the positive side of the graph! So you get a weak initial signal followed by an echo: you won't hear these separately, but you will hear the bass as blurred - this a wave you can look at identify as showing poor bass. All three of the graphs you drew showed very adequate bass because they lacked this signature of falling off to zero and then having a secondary peak. This isn't rocket science: it's a very crude test that I had thought was idiot proof - you look for the pulse to be entirely on the correct side of the x-axis and to lack after peaks on either side of the axis. Something like a flat top followed by a vertical decline is presumably a bonus, but I'm not sure that I'd hear the difference - let alone the difference between relatively subtle variations!

 

Really, you should read stuff before replying, and then reply to what is actually written rather than your fantasy version...

 

- Actual claim: 50hz square wave is a useful tool because it will objectively reveal SOME instances of poor bass

 

- Fantasy claim: 50Hz square wave reveals all, including differences between headphones that all have superb bass!

 

...No. Different!

 

A sort of Dummies Guide in a single example from Head Room:

 

..And indeed I have a pair of the PortaPros handy and the bass, like the 668Bs I have, is undefined/boomy/blurry.


Edited by scuttle - 2/13/13 at 3:24am
post #86 of 130
Thread Starter 
Quote:
Originally Posted by EddieE View Post

 

Exactly if you are going for a certain sound signature, then there is a "perfect" frequency response - for you. You should really say "flat" rather than neutral - suggesting a flat headphone is "neutral" suggests it is presenting sound as the producer/artist heard it, or intended it, which assumes two things -

 

a. That the producers studio monitors were also completely flat and the damping/construction of their room was perfect. Uncommon today, increasingly impossible the further back in time the album you are listening to was produced.

b. That they made absolutely no compensation in their choices for the average commerically available speaker.

 

In the cases of most albums neither is true. The producer did not expect you to be listening in flat speakers and compensated for the average speaker response of whatever time the album was made, and even at the point of production was probably not using a perfectly flat room/gear themselves.

 

This is before you even take into account that what the manufacturer used to define flatness would have been based on a HRTF different to yours. 

 

This arrogance creeping into head fi that there is an objective "right" and "wrong" sound signature to headphones, and some kind of elitist hierarchy of listeners who like "neutral" or "coloured" headphones needs to stop. It's not based on anything remotely rational.

 

You are confusing two separate ideas:

 

- That there is worthwhile information to be gained from an fr

 

And

 

- That the fr can reduced to an exact line rather than a statement like "It should be reasonably flat, with a bit of bulge for bass, and no troughs in range X deeper than Y decibels."

post #87 of 130
Thread Starter 
Quote:
Originally Posted by stv014 View Post

Which of the two bottom filtered 60 Hz square waves do you think is easier to tell apart from the top (unfiltered) one in a listening test ?

 

Can I point out that even the middle one is better than the graph from the PortaPros and 668s? At least if you think of the ear and associated software as being optimized for detecting peaks, which is a working hypothesis.


Edited by scuttle - 2/13/13 at 3:26am
post #88 of 130
Quote:
Originally Posted by scuttle View Post

 

You are confusing two separate ideas:

 

- That there is worthwhile information to be gained from an fr

 

And

 

- That the fr can reduced to an exact line rather than a statement like "It should be reasonably flat, with a bit of bulge for bass, and no troughs in range X deeper than Y decibels."

 

I'm not confusing anything. I already conceded that there can be such a thing as a "bad" set of measurements. I said there was no such thing as a "right" or a "wrong" set of measurements, which was contested.

post #89 of 130
Quote:
Originally Posted by scuttle View Post

Can I point out that even the middle one is better than the graph from the PortaPros and 668s?

 

That is not what the question was, though. What audible differences do you expect from those square wave graphs ?

post #90 of 130
Quote:
Originally Posted by scuttle View Post

 

Ok: my point is that your rhetorical question is a silly one. You are trying to prove that examining 50Hz square waves is never worthwhile by showing that it is not in a particular case where all three waves are (possibly, because you left the scale off) extremely good. This is fundamentally appalling logic: a heuristic doesn't have to be useful in 100% of cases to be worthwhile. For example, not everything that is poisonous is so labelled, but you are an idiot to drink from a bottle marked "poison" if you don't know what it is.

You claimed, "The measurements that really concern me on headphones are the square wave graphs and maybe the CSD."  I agreed on the CSD/Waterfall, I simply questioned why you were so concerned with square waves.  Then, in the same post, you contradicted yourself with, "Otoh, a good 50Hz sq wave doesn't seem to guarantee punchy bass."

 

Quote:
Originally Posted by scuttle View Post

Who says these are "simple" examples to compare? They are not - these signatures would be very hard to tell apart by listening. They have simple shapes and were easy for you to draw - but that is NOT the same thing! You're making a whole raft of foolish assumptions here: you are drawing waves that will sound very alike to the ear because you don't understand how the test is used and then trying to claim that because people can't tell them apart the test isn't useful - but what is actually being measured is your own ignorance and inability to understand a simple explanation.

I say they are simple to compare.  I've used square wave testing of many different devices for decades.  This isn't hard stuff, but you do need to understand the principles, which would then actually lead you to a valid conclusion.  You are correct when you say they would be very hard to tell apart by listening.  And, that's my point.  They look radically different, but don't sound radically different.  So then how could you look at another graph and correlate it to a particular quality of sound?  I've asked that too.

 

Oh, and there's no need for name calling in this forum.  Just sayin'. 

Quote:
Originally Posted by scuttle View Post
Now, again, all of the three graphs you gave are good. Here's what a bad graph looks like:

 

 

700

 

 

See how different they are? Look at the pulse aligned with the q in "Square Wave": the response does NOT stay in the positive side of the graph! So you get a weak initial signal followed by an echo: you won't hear these separately, but you will hear the bass as blurred - this a wave you can look at identify as showing poor bass. All three of the graphs you drew showed very adequate bass because they lacked this signature of falling off to zero and then having a secondary peak. This isn't rocket science: it's a very crude test that I had thought was idiot proof - you look for the pulse to be entirely on the correct side of the x-axis and to lack after peaks on either side of the axis. Something like a flat top followed by a vertical decline is presumably a bonus, but I'm not sure that I'd hear the difference - let alone the difference between relatively subtle variations!

 

Really, you should read stuff before replying, and then reply to what is actually written rather than your fantasy version...

 

- Actual claim: 50hz square wave is a useful tool because it will objectively reveal SOME instances of poor bass

 

- Fantasy claim: 50Hz square wave reveals all, including differences between headphones that all have superb bass!

 

...No. Different!

 

A sort of Dummies Guide in a single example from Head Room:

 

..And indeed I have a pair of the PortaPros handy and the bass, like the 668Bs I have, is undefined/boomy/blurry.

I'd hardly call that graph a Dummy's Guide.  If the caption wasn't there, would you have reached the same conclusions for the Shure and Denon? Couldn't you have seen this from the FR plot, which assigns a magnitude to the FR variation?

 

Did you look at the fourth graph I posted?  How'd that look to you?  How would that sound to you?  

 

As to your statement of claims...who's were they?  Not yours, you didn't say that at all.  I've quoted you exactly above.  I do agree with the "actual claim", though.  However, I'm at a loss to why anyone would value a square wave over an actual response plot for evaluating bass, or any other response problem.  There's no scale, only ambiguous shape.  

 

Again, forgive me if I repeat...I'm not asking for "good" or "bad" here.  You could easily see that on a FR plot.  

 

I'm asking:

 

Why put such weight on a square wave test?

 

What, specifically does it show you (my examples are far simpler to analyze than any headphone, but if you can't do that...you get the idea)

 

How would the examples translate to what is heard?  You've said the first three are all good and hard to tell apart.  What about the fourth?  It looks pretty different to me.  

 

Last question: would it be possible for you to just answer the questions without being demeaning, disrespectful or name calling?   

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › If it graphs bad then it is bad; yes or no?