[autoexec]

So, are there any from older batches that has problems? From what serial number is considered new batches. And what is the earliest serial that has problems? I've not noticed any imbalances or mid-bass hump. From what I hear from mine I think mine is from the older batches (I hope).
 
This is serious issue if the test is correct. I'm dying to hear a defective pair. :/

 

[slwiser]

Purrin, I have a suggestion for future headphone tests...first do not listen to them and get any bias set in your mind.  I can see it happening that you get the tests results that a listening bias would suggests.  I know you try to be better than this but this can happen to the best of us.  Test at least three times with a complete reset of the phones onto your setup first to see if you can get a consistent measurement and when you can then listen to see if what you measured sounds as you would have suspected.  I just think that listening first would set anyone up for the potential for measurement bias'.

 

[anetode]

It's worth pointing out that the 3 measured sets are essentially random selection and that the disparities between them go far beyond what should be acceptable in a high-end product by a serious brand with decades of experience.

It's also worth pointing out that even though I own what's arguably the weidest measuring pair, I still think they sound good. I'll have to spend some type listening to bass-heavy tracks to figure out why the mid-bass hump doesn't sound so bad and I already know that sins of omission in the early-to-mid treble are easier on the ear. Since the 1350s are otherwise well-behaved, it will be interesting to find out how they sound after EQ. It might be that, when EQ-ed to an even response I might not think as highly of them. Of course then I'd have to figure out whether to blame my plebian taste, near-deafness or a microphone

 

[skyline315]

Quote:

It's worth pointing out that the 3 measured sets are essentially random selection...

 
No, they aren't.  Volunteers are sending him their headphones, which, by definition, is not a random sample.  I understand what you're saying, but there's a reason that random samples are one of the foundations of most valid statistical studies.
 
Again, this isn't a criticism against purrin.  It's not as if beyer is going to supply him with a true random sample to test.

 

[anetode]

No, they aren't.  Volunteers are sending him their headphones, which, by definition, is not a random sample.  I understand what you're saying, but there's a reason that random samples are one of the foundations of most valid statistical studies.

Well, you really can't say that they're not random either, statistically, unless you have a large enough sample of volunteer donations to determine their non-randomness. I'm not saying that the sample size is sufficient from the point of view of the manufacturer, indeed I hope that beyer measures and keeps track of all of their stock. Still the results from an arbitrarily random sampling like the aforementioned three are either very unfortunate or very troubling.

 

[skyline315]

Quote:

Well, you really can't say that they're not random either, statistically...

 
Yes.  You can.  The results of this non-random sample may end up being the same as a true random sample, but we have no way of knowing that.  As purrin said himself, people who hear funny things in their headphones may be more inclined to send them in to him for testing.  Of course, the opposite can be true, in that people with good headphones may be eager to know that they sound good and send them in for that reason.  There's just no telling.  This is why volunteer-based sampling is typically frowned upon (too many lurking variables that can't be accounted for) unless you're required to use volunteers for ethical reasons (medical testing, etc.).  
 
Sorry to argue the point and be technical about it, but like I said earlier...stat teacher here 

 

[anetode]

Yes, I understand the point of contention. My only intent was to point out that there are statistical tests to determine the randomness of a sample which would be useless when talking about a sample size of 3. So in the same way that we don't know whether the results of this non-random sampling would coincide with a truly random sampling, there's also a paucity of evidence to indicate that self-selection bias is tainting these results towards any specific end. We have a few results because they were convenient to collect, I don't think that anyone is disputing that. Rather it's that the data's already troubling and since it is being collected by forum volunteers who lack the resources of, say, Consumer Reports, we're lucky there's any data to begin with.

 

[ecosse_011172]

How comfortable are these headphones for extended use?
I've given up on my Etymotic ERP-4 IEMs due to lack of comfort and also my Sennheiser HD25-SP II as I find them too unconfortable after a while due to having a large head.
I'd love to be to replace them both for something that will have great isolation, low leakage great sound and enough sensitivity to be loud enough with an iPod when travelling on airplanes, trains etc.
 
I love my AKG K550 headphones for office use but they aren't much cop as a portable headphone.
 
 
Help!

 

[rasmushorn]

Quote:

How comfortable are these headphones for extended use?
I've given up on my Etymotic ERP-4 IEMs due to lack of comfort and also my Sennheiser HD25-SP II as I find them too unconfortable after a while due to having a large head.
I'd love to be to replace them both for something that will have great isolation, low leakage great sound and enough sensitivity to be loud enough with an iPod when travelling on airplanes, trains etc.
 
I love my AKG K550 headphones for office use but they aren't much cop as a portable headphone.
 
 
Help!

I love the sound - but more than two-three hours my ears hurt and I need a break. Often I do not need to listen that long but I have used it while playing Diablo III for longer sessions and comfort is a problem for me with the DT-1350. 

 

[Teuneu]

If I use my glasses, my ears hurt as h*ll after 2,5 hours. If I use my contact lenses, I can use them for almost 4 without too much pain. I mostly use the dt-1350's for traintrips (3 hour trips) and while studying. If I take a coffee break every 3 hours, it's no problem to use them the entire day. My former headphone was the alessando ms1i, which was wayyyyy more comfortable, but I couldn't use them in the library for studying. Comfort is a issue with these, but I adjusted from glasses to contact lenses and it's way better now...

 

[rhythmdevils]

Quote:

 
What's shocking is all this skepticism of Beyerdynamic (and the decades of experience of their acoustic engineers have) contrasted with blind faith in Purrin*.
 
*No offence to Purrin, this post isn't directed at him.

 
I would describe it more like trust with Purrin.  He's earned my trust after a year of scrutinizing his measurements and the way they compare to my own impressions not to mention getting to know him personally.  Beyer and pretty much every other headphone manufacturer has earned my distrust after 7 years being frustrated by the mediocre to crappy headphones they churn out.
 
I don't trust decades of existence.  I trust results.  I find Purrin's results to be outstanding (though of course, not perfect) and I find Beyer's results to be average at best.

 

[purrin]

Quote:

Purrin, I have a suggestion for future headphone tests...first do not listen to them and get any bias set in your mind.  I can see it happening that you get the tests results that a listening bias would suggests.  I know you try to be better than this but this can happen to the best of us.  Test at least three times with a complete reset of the phones onto your setup first to see if you can get a consistent measurement and when you can then listen to see if what you measured sounds as you would have suspected.  I just think that listening first would set anyone up for the potential for measurement bias'.

 
Agreed. That's sort of the plan. I would probably take the same measurement three times in three days, removing the headphone from the rig each time. I would also do some positional testing to see how changing the location of the headphone would affect the measurement. I definitely think this extra effort is warranted to get the bottom of this.
 
I'll be sure to build a faraday cage around my garage too.

 

[ecosse_011172]

Sounds like they wouldn't be any better than my HD25-1 II .
What about the V'MODA Crossfade M80?
What else should I look at?

 

[Magick Man]

Sounds like they wouldn't be any better than my HD25-1 II .
What about the V'MODA Crossfade M80?
What else should I look at?

I like the M-80s, they sound good to my ears and make for excellent hard rock cans. The best thing about them, though, is their durability. I've not been gentle with mine and they still look almost new.

 

[skyline315]

Quote:

Purrin, I have a suggestion for future headphone tests...first do not listen to them and get any bias set in your mind.  I can see it happening that you get the tests results that a listening bias would suggests.  I know you try to be better than this but this can happen to the best of us.  Test at least three times with a complete reset of the phones onto your setup first to see if you can get a consistent measurement and when you can then listen to see if what you measured sounds as you would have suspected.  I just think that listening first would set anyone up for the potential for measurement bias'.

 
Quote:

 
Agreed. That's sort of the plan. I would probably take the same measurement three times in three days, removing the headphone from the rig each time. I would also do some positional testing to see how changing the location of the headphone would affect the measurement. I definitely think this extra effort is warranted to get the bottom of this.
 
I'll be sure to build a faraday cage around my garage too.

 
Funny, because I would recommend exactly the opposite.  Listen first and jot down subjective notes, and then see if the measurements confirm or contradict what you heard.  I think if you see measurements that indicate a mid-bass hump, let's say, then this is exactly what you'll hear now that you know the measurements say it exists.  In reality, however, it may not be noticeable even if it shows up in the graph.  
 
I agree with the multiple tests and repositionings.  Doesn't Tyll do it 5 separate times and then average them for his final frequency response graph?