[bigshot]

I'm curious, wouldn't it be better to use a mono EQ signal and balance the whole room at once? It seems to me two separate wouldn't behave the same as two together.

 

[Joe Bloggs]

I'm curious, wouldn't it be better to use a mono EQ signal and balance the whole room at once? It seems to me two separate wouldn't behave the same as two together.

I measured the room response on a per-speaker basis as well (paired with the subwoofer, though). Several speakers playing at once interfere and cancel each other creating comb filtering artifacts in the measurement, it won't sound good if you EQ according to such a measurement.

 

[bigshot]

I've always found that a first pass EQ in really good mono is the best starting point. I have tweaked channels a little different since though, but that is because my sub is on the right. Those auto adjustment things have never worked for me at all.

 

[Joe Bloggs]

How do you determine what EQ setting to use? What kind of measurement or listening test?

 

[bigshot]

I started out with the auto EQ on my Yamaha, and then I worked to balance relative levels and peel back the masking layer by layer by ear over the period of several months using directly recorded CDs of orchestral and chamber music. Keyboard runs played with even weight is very handy too. Very small adjustments at a time. I probably still have narrow spikes and dips, but the overall curve is there.

 

[jaddie]

If you measure you have to do each channel separately.  Also, each channel has to be measured using a spacial/temporal average.  Single point measurements are meaningless.

 

[higbvuyb]

Why do some custom IEMs have two of each type of driver in each ear? e.g. Two identical tweeters + two identical midrange + two identical lowrange.
 
What benefit would this provide to make up for the (often astronomical) cost increase versus a 3 driver model? Better matching? Lower impedance? slightly higher loudness?

 

[limpidglitch]

My assumption has always been that as most things a balanced armature pushed to its limits will start to distort rather badly.
So if you have two armatures you can have double the sound pressure of a single armature, with the same amount of distortion.
 
But exactly how significant this improvement is I have my doubts about. It's not exactly as if people have been complaining about excessive distortion with the ER4.

 

[Cristello]

Quote:

Why do some custom IEMs have two of each type of driver in each ear? e.g. Two identical tweeters + two identical midrange + two identical lowrange.

Some custom IEMs may not have identical pairs, but instead divide the midrange (or highs/lows) between two more narrowly tuned BAs, similar to how speakers can have multiple cones of similar size cover different frequency bands.
 
Quote:

My assumption has always been that as most things a balanced armature pushed to its limits will start to distort rather badly.
So if you have two armatures you can have double the sound pressure of a single armature, with the same amount of distortion.

 
If the BAs are identically implemented, then this makes the most sense.
In the example of the ER4, THD plots always indicate a slight rise in distortion as the frequency lowers. Adding another driver would probably help limit the over excursion of any one BA.
 
I would suppose that the rise in large BA-array IEMs is likely just good marketing: 'I have six drivers!!!... "oh yeah? well we have EIGHT. suck it!"...'
Reminds me of the common confusion amongst consumers with CPU cores (It haz more cores so it must be bettaz, derp)

 

[higbvuyb]

Quote:

Quote:

Why do some custom IEMs have two of each type of driver in each ear? e.g. Two identical tweeters + two identical midrange + two identical lowrange.

Some custom IEMs may not have identical pairs, but instead divide the midrange (or highs/lows) between two more narrowly tuned BAs, similar to how speakers can have multiple cones of similar size cover different frequency bands.

If we look at some of the UM custom IEMs, one model has two TWFKs (total of 4 drivers) and another has two TWFKs and a pair of another BA. If they're the same, all it does is increase the SPL by 3 dB for the same level of distortion. If they covered different ranges, I'm sure they'd advertise them as the 4 or 6 drivers covering different ranges instead of pairing them off as lows, mids, highs.

 

[limpidglitch]

Quote:

 
In the example of the ER4, THD plots always indicate a slight rise in distortion as the frequency lowers. Adding another driver would probably help limit the over excursion of any one BA.
 

 
I've been looking at the THD plots over on InnerFidelity, and there's no very obvious correlation between number of armatures and level of distortion.
Except maybe if you push it to 100dB, but that sort of sound pressure level will damage your hearing very quickly.
Even the cheaper dynamic IEMs, which I always had been suspecting to fare worse on this front, do very well.
 
This is just a random selection of the models I have some knowledge of.
The top three are single armature, the rest have at least dual bass drivers.
 
 
 

 

[mikeaj]

Quote:

If we look at some of the UM custom IEMs, one model has two TWFKs (total of 4 drivers) and another has two TWFKs and a pair of another BA. If they're the same, all it does is increase the SPL by 3 dB for the same level of distortion. If they covered different ranges, I'm sure they'd advertise them as the 4 or 6 drivers covering different ranges instead of pairing them off as lows, mids, highs.

 
Somebody please tell me if I'm horribly wrong, but isn't it 6 dB? They're (hopefully) outputting the same signal in phase at the eardrum.
 
Now, as for the distortion generated by the actual products as a whole...

 

[Cristello]

Quote:

 
I've been looking at the THD plots over on InnerFidelity, and there's no very obvious correlation between number of armatures and level of distortion....

 
indeed, those THD plots also factor in the housing and (in the case of the muilt-BA units) internal crossovers. The best way to test might be measuring an on-axis anechoic near-field response of a single BA and two side-by-side, then compare the two at the same calibrated dB level.
 
I am curious to see if having the extra BAs is really worth the trouble.
(Someone needs to start trying to measure large-driver-count custom monitors... Maybe get molds done for a "Pinna & HRTF Simulator"???)
 
EDIT: fixed in accordance with current protectionist guidelines.

 

[bigshot]

You can't say "dummy head" on Head Fi. It's against the rules.

 

[Cristello]

Oh. I'll fix that right away.
Thank you! it won't happen again.