Skullcandy Mix Master Mike versus Audeze LCD-3

[Chromako]

A lot of this discussion is a bit over my head (and I'm worried about hostility... be nice, everyone!)
 
But actually, I've been asking myself very similar questions. Why do headphones sound so different when so many of the numbers are so similar (that is, some cheap ones seem remarkably similar in numerical tests to uber expensive ones). 
 
My conclusion: the particular numbers we are looking at don't tell the whole story (duh). Perhaps our analysis, to get a fuller picture and answer the question, needs to go further. Maybe more square or triangular waves, or harmonic tests, of more frequencies. The interesting thing is that Headphone.com tests occur with the microphones in a synthetic "head" with ears and such that should match human noggins pretty well. 
 
I'm not entirely buying the "larger diaphragm" or "differing angle" argument too much, as first, the test uses a synthetic realistic head/ears, and also, changing driver angles should be pretty easy anyway. Not something that justifies kilobuck investments.
 
Also: The "golden ears" tidbit- rubbish. Good instrumentation can capture far more precise/more information than humans could ever capture. 
 
 
What I can say is that the headphones do sound different, but have similar numbers in our tests. Interesting as the human nervous system is surprisingly "digital." Conclusion #2: This needs further study. 

 

[kiteki]

Originally Posted by xnor /img/forum/go_quote.gif
 
Right, but what many don't seem to understand is that CSD is just a visualization technique. A visualization of the impulse response to show both frequency and time domain data on a single graph. That's all there is to it.
If you see ringing in the impulse response, you will see in the CSD too, of course.
 

Moral of the story: gotta use the right tools. FR is only one of them. FR + CSD is better. FR + CSD + few scattered quality and meaningful non-linear distortion measurements gives us a much better picture.

I'd say impulse response first (--> FR, CSD ...). Distortion measurements second (THD, IMD).

 
Xnor, so you mean impulse response is the most important measurement?  I'm just wondering why don't more headphones use anechoic materials to improve IR?  In speakers it makes sense, since the echo is in the room.  In IEM's, well some CIEM's use special less echoic tubes.

 

[kiteki]

 
Also: The "golden ears" tidbit- rubbish. Good instrumentation can capture far more precise/more information than humans could ever capture. 

 
Nothing to do with golden ears, talking about someone completely deaf in their left ear, that can hear the differences in the Apple earbud and a high-end STAX within half a second.  It's due to the driver size, geometry, angle, distance etc. which the human ear can locate, and which we don't have in any of our (current) data.
 
However, people assert over and over, that instruments are more precise than any human, in every single instance.
 
Another example is voicing, do you know how hard it is to program software to identify someones voice?  Not talking about accents, pronounciation, just a voice.  Do we apply that software to FR or something?  No.
 
Asserting that data is total is one thing - asserting "this is so complex, deep, yet factually intact, that anyone contesting it simply has no idea"... is another.
 
(not aiming that at you, and the rest of your post was OK)

 

[xnor]

Quote:

 
I'm just wondering why don't more headphones use anechoic materials to improve IR?  In speakers it makes sense, since the echo is in the room.  In IEM's, well some CIEM's use special less echoic tubes.

 
Ignorance or simply other goals than high sound quality, such as, saving costs where possible or making headphones look nice (acoustics as an afterthought). Grado for examples designs their headphones by ear (that's what they claim anyway), which is why they don't measure well and sound harsh to my (and a lot of other guys') ears. Now Skullcandy, on the other hand, has changed their mind with some of their latest headphones and actually hired someone with a degree in acoustics and EE and the results speak for themselves.
 
Similarly to the crappy headphones, there are many computer speakers, 2/5/7.1 sound systems that just sound awful but look nice, are small, or simply are inexpensive.
 
That doesn't mean that it's impossible to make great sounding headphones for little money, but I doubt many manufacturers are even seriously trying. Some just take OEM models and put their brand name on it.

 

[kiteki]

Interesting as the human nervous system is surprisingly "digital."

 
Everything analog will look digital as soon as you zoom in and slow it down so we can process it.
 

 

[kiteki]

 
Ignorance or simply other goals than high sound quality, such as, saving costs where possible or making headphones look nice (acoustics as an afterthought). Grado for examples designs their headphones by ear (that's what they claim anyway), which is why they don't measure well and sound harsh to my ears.
 
Similarly, there are many computer speakers, 2/5/7.1 sound systems that just sound awful but look nice, are small, or simply are inexpensive.
 
That doesn't mean that it's impossible to make great sounding headphones for little money, but I doubt many manufacturers are even seriously trying.

 
Thanks, time to buy some velvet. )
 
That Koss DJ100 thing could be fun, considering it's IR.

 

[liamstrain]

Quote:

 
If an open-air STAX, open-air dynamic and open-air isodynamic planar all had the exact same FR and SWR, they'd still sound completely different, identifiable by humans.

 
This is an assertion, and does not mesh with my user experience. In reality, it was very difficult to distinguish them aside from FR trends. When they were EQ'd to be the closer in response, I could not tell them apart aside from fit/weight.

 

[liamstrain]

Quote:

 
Another example is voicing, do you know how hard it is to program software to identify someones voice?  Not talking about accents, pronounciation, just a voice.  Do we apply that software to FR or something?  No.
 

 
We've been over this. Software can immediately tell that there is a difference between two voices - and can enumerate those differences very specifically and accurately. What it cannot quickly do, is match that voice to a particular person - that's pattern recognition and a whole host of other human processes. 

 

[kiteki]

 
This is an assertion, and does not mesh with my user experience. In reality, it was very difficult to distinguish them aside from FR trends. When they were EQ'd to be the closer in response, I could not tell them apart aside from fit/weight.

 
Far too vague, which STAX?
 
 

 
We've been over this. Software can immediately tell that there is a difference between two voices - and can enumerate those differences very specifically and accurately. What it cannot quickly do, is match that voice to a particular person - that's pattern recognition and a whole host of other human processes.

 
Humans can identify differences in voices faster and more accurately than any software can, asides from identifying them to a specific person.
 
Likewise if you have an ebony versus maple fretboard guitar, they will have different voicings, it's a similar concept.  Now, where is the code of ebony / maple in the FR, CSD, SWR, IR, THD+N, IMD?  That's the important part, which a lot of people are dismissing - why?

 

[liamstrain]

Quote:

Far too vague, which STAX?
 
Stax 009, 007 mk2, and 507,  which I tested against the Hifiman HE-500, and Audeze LCD2 along with Sennheiser HD800, HD600 and Beyer 880. On their own they were easy to distinguish - and part of that is fit/weight. But they do have characteristic frequency responses. But once EQ'd the distinctions were much harder to make with any certainty. If there was a way to have them all fit/feel the same without affecting the sound, I expect it would have been nearly impossible to tell. Especially between those with flat transducers (Ortho and Electrostatics) - I know many people claim different transient response, but honestly, I heard very very very little difference once EQ'd (and level matched). Consequently, I will not be shelling out for the Stax myself. 
 
 
Humans can identify differences in voices faster and more accurately than any software can, asides from identifying them to a specific person.
 
Evidence? Software is pretty good at doing quick comparisons in raw data and showing differences - crunching numbers. All we're looking for is X is different than Y. Is that how you mean the word "identify."
 
I don't doubt that humans do a good bit of synthesis and comparative analysis unconsciously... but I'd like to see evidence for your statement. 
 
 
Likewise if you have an ebony versus maple fretboard guitar, they will have different voicings, it's a similar concept.  Now, where is the code of ebony / maple in the FR, CSD, SWR, IR, THD+N, IMD?  That's the important part, which everyone is dismissing, such as "we've been over this".
 
I do not know. But I suspect you would see it show up in SWR and waterfall's showing differences in decay and resonance. 

 

[Joe Bloggs]

Quote:

 
Xnor, so you mean impulse response is the most important measurement?

I think he was just saying that impulse response is more important than distortion.  Most of us have been saying that frequency response stands way above the other measurements in importance, xnor included AFAICS.

 

[firev1]

 

[rhythmdevils]

Quote:

 
Xnor, so you mean impulse response is the most important measurement?  I'm just wondering why don't more headphones use anechoic materials to improve IR?  In speakers it makes sense, since the echo is in the room.  In IEM's, well some CIEM's use special less echoic tubes.

 
In speakers it's not just the room.  Resonance can come from the driver itself, or from the enclosure of the speaker.  In the speaker world, people pay attention to resonances in all of these places including the room.  With headphones, for some reason no one cares.  Xnor was right on in his earlier post. 

 

[kiteki]

Originally Posted by liamstrain /img/forum/go_quote.gif
 
Stax 009, 007 mk2, and 507, which I tested against the Hifiman HE-500, and Audeze LCD2 along with Sennheiser HD800, HD600 and Beyer 880. On their own they were easy to distinguish - and part of that is fit/weight. But they do have characteristic frequency responses. But once EQ'd the distinctions were much harder to make with any certainty. If there was a way to have them all fit/feel the same without affecting the sound, I expect it would have been nearly impossible to tell. Especially between those with flat transducers (Ortho and Electrostatics) - I know many people claim different transient response, but honestly, I heard very very very little difference once EQ'd (and level matched). Consequently, I will not be shelling out for the Stax myself.

 
Your story there, if you think it's nearly impossible to tell apart the SR-009, 007, 507, HE-500, LCD-2, HD800, HD600 and DT880, once they're all equalized to have the same FR, apart from fit and weight, why haven't you advocated for the MMM to be tailored into the LCD-3 yet?
 
 

Originally Posted by liamstrain /img/forum/go_quote.gif
 
Evidence? Software is pretty good at doing quick comparisons in raw data and showing differences - crunching numbers. All we're looking for is X is different than Y. Is that how you mean the word "identify."
 
I don't doubt that humans do a good bit of synthesis and comparative analysis unconsciously... but I'd like to see evidence for your statement.

 
I don't have evidence for everything, it's just common sense, you know.  I don't only look at what's (currently) on paper, since then I'd believe the world was flat if satellite imagery wasn't invented yet, clearly that's not right, is it?
 
Anyway here...
 

 Text-Dependent, Text-Independent and Text-Prompted Methods Speaker recognition methods can also be divided into text-dependent (fixed passwords) and text-independent (no specified passwords) methods. The former require the speaker to provide utterances of key words or sentences, the same text being used for both training and recognition, whereas the latter do not rely on a specific text being spoken. The text-dependent methods are usually based on template/model-sequence-matching techniques in which the time axes of an input speech sample and reference templates or reference models of the registered speakers are aligned, and the similarities between them are accumulated from the beginning to the end of the utterance. Since this method can directly exploit voice individuality associated with each phoneme or syllable, it generally achieves higher recognition performance than the text-independent method. There are several applications, such as forensics and surveillance applications, in which predetermined key words cannot be used. Moreover, human beings can recognize speakers irrespective of the content of the utterance. Therefore, text-independent methods have attracted more [url=http://www.scholarpedia.org/article/Attention][color=#0066cc]attention[/color][/url]. Another advantage of text-independent recognition is that it can be done sequentially, until a desired significance level is reached, without the annoyance of the speaker having to repeat key words again and again. Both text-dependent and independent methods have a serious weakness. That is, these security systems can easily be circumvented, because someone can play back the recorded voice of a registered speaker uttering key words or sentences into the microphone and be accepted as the registered speaker. Another problem is that people often do not like text-dependent systems because they do not like to utter their identification number, such as their social security number, within the hearing of other people. To cope with these problems, some methods use a small set of words, such as digits as key words, and each user is prompted to utter a given sequence of key words which is randomly chosen every time the system is used. Yet even this method is not reliable enough, since it can be circumvented with advanced electronic recording equipment that can reproduce key words in a requested order. Therefore, a text-prompted speaker recognition method has been proposed in which password sentences are completely changed every time.

 

[kiteki]

 
In speakers it's not just the room.  Resonance can come from the driver itself, or from the enclosure of the speaker.  In the speaker world, people pay attention to resonances in all of these places including the room.  With headphones, for some reason no one cares.  Xnor was right on in his earlier post. 

 
...and much like your previous post, what does this have to do with anything being discussed?  I assume if I answer, you won't continue the dialogue either way, and just come back later to leave more negative comments.
 
Even the Etymotic ER-4's were evaluated with listening and voting as well.  http://gilmore2.chem.northwestern.edu/ubb/showpost.php?fnum=1&tid=1718&pid=13330&fpage=3
 
The reason fewer care with headphones, is because if you're aiming for an apex sonata open-window to the recording with no fluctuation in colour or tones anywhere, you buy speakers.
 
imho.