[CraftyClown]

  krkrkrkr <- sound of me laughing
 
yes it's done with the veritas, else there is no point. a test tone could be full scale(reaching 0db) or not, as the multimeter will measure the signal, if the signal isn't maxed out you need to account for that when estimating the actual max voltage you get from the amp at a given setting. it's a silly detail, but I've messed up with that a significant number of time so I mention it ^_^.
 
I'm all for dirt cheap multimeter, I was just warning that those stuff aren't audio analyzers, the AC voltage measurements will most likely be made to work well on 50 or 60hz and 110 or 220volt for home usage. it's not designed to measure 15mv at 20khz. some can be, but they cost real money.
so what I did with my dirt cheap one was to take my fiio X1 look online for the max voltage output and measured several sine waves at different frequencies to check if at least one would give a proper result. then I checked with the tones that seemed to work into other DAPs, cellphones, dac/amps,  where I could get the max output voltage to confirm that it wasn't luck. once you know you can trust a method for your multimeter, it's all good.
 
they most likely measure the efficiency the very way we want to do it with the veritas. they have a calibrated coupler, they send a 1khz tone into it at 0.1 or 0.2V or whatever, and measure the output with the calibrated coupler. even if it's expressed in db, what is measured is the voltage.
yes I said certificate in reference to the paper with measurements that comes with the er4. but really the general specs are fine too. for example it's written on mine that the drive level is 200mV and the IEMs give respectively 104.5 and 104.6db at 1khz.
the specs for the er4sr are 98db for 0.1V.   double the voltage is 6db louder, that's how it always is for all analogous audio signal. but if in doubt you can always jump there to avoid having to calculate it yourself http://www.sengpielaudio.com/calculator-gainloss.htm 
so 98db with 0.1V means 98+6=104db into 0.2V we're 0.5db wrong and I think we can survive that magnitude of error ^_^.
 
I'm not sure I get the rest of the questions. if you have a multimeter you can check current whenever you like with any frequency. the accuracy might be given in the specs of the multimeter. but a little very serious advice for those who aren't too familiar with multimeters, the current settings are where you can blow the fuse! it's fairly hard to mess up with voltage and kind of impossible to mess up with resistance. but it's easy to mess up with current measurements, so always care about the setting and check that you're not on current next time you're using the multimeter. I always set it back to voltage and move the probe before turning it off, because I'm paranoid and I don't trust myself

and yes different test tone frequency equal different current frequency and different interaction with the IEM. but I really was talking about using lower frequencies and the multimeter just to make sure we get the right voltage calibration. once we move on to using the veritas, the frequency response isn't properly calibrated so there will be some matter of guessing anyway based on the FR of the IEM. nothing is perfect, and it's even less perfect when like me you go for cheap gear

 
Hey Castle,
 
so I have a multimeter now but electronics novice that I am, I'm still not sure where I'm taking the reading from. 
 
If it's not too much to ask, could you point me in the right direction?
 
Cheers mate

 

[csglinux]

 
 
they most likely measure the efficiency the very way we want to do it with the veritas. they have a calibrated coupler, they send a 1khz tone into it at 0.1 or 0.2V or whatever, and measure the output with the calibrated coupler. even if it's expressed in db, what is measured is the voltage.
yes I said certificate in reference to the paper with measurements that comes with the er4. but really the general specs are fine too. for example it's written on mine that the drive level is 200mV and the IEMs give respectively 104.5 and 104.6db at 1khz.
the specs for the er4sr are 98db for 0.1V.   double the voltage is 6db louder, that's how it always is for all analogous audio signal. but if in doubt you can always jump there to avoid having to calculate it yourself http://www.sengpielaudio.com/calculator-gainloss.htm 
so 98db with 0.1V means 98+6=104db into 0.2V we're 0.5db wrong and I think we can survive that magnitude of error ^_^.
 
I'm not sure I get the rest of the questions. if you have a multimeter you can check current whenever you like with any frequency. the accuracy might be given in the specs of the multimeter. but a little very serious advice for those who aren't too familiar with multimeters, the current settings are where you can blow the fuse! it's fairly hard to mess up with voltage and kind of impossible to mess up with resistance. but it's easy to mess up with current measurements, so always care about the setting and check that you're not on current next time you're using the multimeter. I always set it back to voltage and move the probe before turning it off, because I'm paranoid and I don't trust myself

and yes different test tone frequency equal different current frequency and different interaction with the IEM. but I really was talking about using lower frequencies and the multimeter just to make sure we get the right voltage calibration. once we move on to using the veritas, the frequency response isn't properly calibrated so there will be some matter of guessing anyway based on the FR of the IEM. nothing is perfect, and it's even less perfect when like me you go for cheap gear

 
Thanks again @castleofargh!  (Going slightly off at a tangent - I presume your username is Python-inspired?!?) I greatly appreciate all the info from members on this site. It's truly an amazing resource.
 
Here's one more general question (aimed at anybody) about measurements. If you take a look at any of the Innerfidelity data sheets on any IEM they've ever measured, it appears they've taken at least four measurements to get their FR curve plots (always shown at the top left of the sheet), because the L and R are always different, and the L and R diffuse-field compensated curves aren't consistently shifted from the L and R raw measurement data, i.e., there are additional discrepancies between the L and R measurements in both raw and df comp graphs. So it would appear to me that they measured L and R once for their raw curves and once again with the delta correction. Here's my issue with all this... If somebody wants to show a graph indicating the discrepancy between the L and R earbuds, that's fine, but what should really matter (to everybody) is the average FR of the headphone. If these discrepancies are indeed due to IEM manufacturing tolerances, neither Innerfidelity's L nor R measurement are going to exactly match mine, even if we were using the exact same coupler. (Heck, Innerfidelity's L and L measurements wouldn't agree with themselves if they re-inserted the eartip into the dummy head very slightly differently on the second measurement!)
 
Why not present an ensemble average over multiple measurements, and over both L and R? I would want to see one curve showing the average response of the headphone. Have Innerfidelity just been a bit sloppy or am I overlooking something critical?

 

[csglinux]

   
Hey Castle,
 
so I have a multimeter now but electronics novice that I am, I'm still not sure where I'm taking the reading from. 
 
If it's not too much to ask, could you point me in the right direction?
 
Cheers mate

 
You should have an AC voltage setting on your multimeter. Connect the multimeter terminals to the L or R output from your source. (If you have a 3.5mm to phono cable, for example, that would expose the contacts you need.)

 

[castleofargh]

exactly, usually when you start fooling around you end up with lots of short interconnect cables, crocodile plugs, breadboard, soldering stuff... it's like a hobby in the hobby.

in case it's not clear, if you go for a male to male cable and measure that from a device, the multimeter has very high impedance and the measurement is considered unloaded. and even when you do use a load and measure the voltage in parallel, voltage law tells you that you get the same value as the one going into the load, and because the multimeter has really high impedance, it's not changing the circuit in any significant way.
all in all, as I said before, as long as you stick to resistance and voltage, it's hard to do wrong. end even current isn't hard at all as long as you mind the settings and don't let them as is for the next time you measure voltage ^_^.
 
 

  Thanks again @castleofargh!  (Going slightly off at a tangent - I presume your username is Python-inspired?!?) I greatly appreciate all the info from members on this site. It's truly an amazing resource.
 
Here's one more general question (aimed at anybody) about measurements. If you take a look at any of the Innerfidelity data sheets on any IEM they've ever measured, it appears they've taken at least four measurements to get their FR curve plots (always shown at the top left of the sheet), because the L and R are always different, and the L and R diffuse-field compensated curves aren't consistently shifted from the L and R raw measurement data, i.e., there are additional discrepancies between the L and R measurements in both raw and df comp graphs. So it would appear to me that they measured L and R once for their raw curves and once again with the delta correction. Here's my issue with all this... If somebody wants to show a graph indicating the discrepancy between the L and R earbuds, that's fine, but what should really matter (to everybody) is the average FR of the headphone. If these discrepancies are indeed due to IEM manufacturing tolerances, neither Innerfidelity's L nor R measurement are going to exactly match mine, even if we were using the exact same coupler. (Heck, Innerfidelity's L and L measurements wouldn't agree with themselves if they re-inserted the eartip into the dummy head very slightly differently on the second measurement!)
 
Why not present an ensemble average over multiple measurements, and over both L and R? I would want to see one curve showing the average response of the headphone. Have Innerfidelity just been a bit sloppy or am I overlooking something critical?

yup the holy grail is in the castle of aaaaaaagggggghhhhhhhhhhhhhhhhhhh.................... and I'm scared of bunnies, as anybody should.
 
Tyll's measurements are done on a dummy head, so the left and right measures are done by 2 different couplers. it's just that. and why the compensated curve isn't the same on left and right, I don't really know, maybe the raw measurement is really a pure uncalibrated raw measurement, and the compensated one has the fake diffuse field kind of curve applied, plus some correction to deal with differences between left and right couplers? you'd have to ask Tyll to be sure.
but it's only 1 measurement done on IEMs(one for each ear), maybe he runs it a few times like we can do in REW to try and even things out in case of noises, current spike or whatever, but that wouldn't show on the results anyway.
for headphones, now he does a all bunch of measurements with the headphone at different positions on the head. and the compensated curve is an average of all the raw measurements, and then the compensation curve applied and whatever it is that he does. but for IEMs he seems to stick to one position and one measure(per ear).
 
for your concerns about manufacturing tolerance and reinserting the IEM into a measurement gear, welcome to the reality of what people ironically decided to call high fidelity. ^_^
there will always be some variations and our "job" is to try and get a method that is as consistent as possible over time when you come back and measure the same IEM again. if the purpose was really to just show variations between left and right for 500 of the same IEM, that would be rather easy to implement, make a mold for the specific nozzle size and be done with plug unplug variations. but a system to measure all sorts of IEMs, it's not simple.  for example, I wish the veritas was transparent to make a marking on it and see when it's good. just that would help a lot for some IEMs.
 
doing measurements makes you mad for a time, then it makes you actually more relaxed about the all audio fidelity war and how to care and understand measurements while not being obsessed by 0.2db on a graph. it most certainly has an educating effect on me.

 

[csglinux]

Perhaps only one measurement per side is an indication of Innerfidelity's supreme confidence in their measurements?! I think I'll continue to take averages

I understand your comment about wanting a transparent (see-through) coupler. If I were 3D printing my own coupler, I think I'd do away with the conical taper of the Vibro Veritas device. All my silicone tips would just slowly pop out of this thing, and trying to hold the IEM in with putty makes repeatability pretty hard. I really want to measure everything with SpinFit tips, because I almost always use them for listening. What I did was to glue a small rubber washer to the end of the coupler (one of my other fun hobbies is plumbing), so that the inner diameter of the washer is just a bit smaller that that of the coupler. Now, when I push in a SpinFit tip, the washer holds it at the exact same insertion depth each time. I can't gauge how close that distance would be compared to the actual distance from the tip to my eardrum when I'm listening, but at least it's consistent from measurement to measurement.

 

[CraftyClown]

   
You should have an AC voltage setting on your multimeter. Connect the multimeter terminals to the L or R output from your source. (If you have a 3.5mm to phono cable, for example, that would expose the contacts you need.)

 
Sorry but I'm still rather confused. I attached a 3.5mm to phono cable to my sound card whilst playing a looped 1khz tone and then with the multimeter set to measure ac voltage I'm touching one prong on the meter to the left output of the phono cable and the other prong on the meter to the right output.
 
Is this correct? What kind of result should I be expecting?

 

[csglinux]

   
Sorry but I'm still rather confused. I attached a 3.5mm to phono cable to my sound card whilst playing a looped 1khz tone and then with the multimeter set to measure ac voltage I'm touching one prong on the meter to the left output of the phono cable and the other prong on the meter to the right output.
 
Is this correct? What kind of result should I be expecting?

 
You can use either 1) the two connections to the left earpiece or 2) the two for the right.

 

[CraftyClown]

   
You can use either 1) the two connections to the left earpiece or 2) the two for the right.

 
So both prongs touch either the left phono or the right phono?
 
When both are touching I am getting 0.01v

 

[csglinux]

   
So both prongs touch either the left phono or the right phono?
 
When both are touching I am getting 0.01v

 
Yes, there are two wires going to each earpiece. You can take the two connections from either the right or the left, but don't mix connections from left and right. The two connections to the left earpiece OR the two connections to the right earpiece. I had to crank my USB DAC/amp up pretty high to hit the 0.2V I needed to match my Etymotic spec. Depending on the soundcard you're using, you will likely need to raise that volume a lot or use something else as your output. I have a StarTech soundcard, and it's not an ideal output source, IMHO, because it has quite a high output impedance, which can mess up the FR from lower-impedance phones.

 

[CraftyClown]

   
Yes, there are two wires going to each earpiece. You can take the two connections from either the right or the left, but don't mix connections from left and right. The two connections to the left earpiece OR the two connections to the right earpiece. I had to crank my USB DAC/amp up pretty high to hit the 0.2V I needed to match my Etymotic spec. Depending on the soundcard you're using, you will likely need to raise that volume a lot or use something else as your output. I have a StarTech soundcard, and it's not an ideal output source, IMHO, because it has quite a high output impedance, which can mess up the FR from lower-impedance phones.

 
Ok I'm clearly misunderstanding something here.
 
I'm using a Geekout 100 to output and I have a 3.5mm to left and right phono cable connected to it. When I touch both the red and black poles of the multimeter to either the left phono pole or the right one, I only get 0.01v
 
What have I done wrong here?

 

[csglinux]

   
Ok I'm clearly misunderstanding something here.
 
I'm using a Geekout 100 to output and I have a 3.5mm to left and right phono cable connected to it. When I touch both the red and black poles of the multimeter to either the left phono pole or the right one, I only get 0.01v
 
What have I done wrong here?

 
If you have the appropriate cable, you should have two connections per side. If you're connecting your two multimeter leads to the same piece of metal, it will read zero You need to connect the multimeter contacts to the two separate terminals on one side (left or right channel).  Just be careful not to short the outputs with your multimeter leads.

 

[CraftyClown]

 
 
If you have the appropriate cable, you should have two connections per side. If you're connecting your two multimeter leads to the same piece of metal, it will read zero You need to connect the multimeter contacts to the two separate terminals on one side (left or right channel).  Just be careful not to short the outputs with your multimeter leads.

 
Excellent, got it!
 
So the next question is do I just tweak my volume output until I can get almost exactly 200mv?

 

[csglinux]

   
Excellent, got it!
 
So the next question is do I just tweak my volume output until I can get almost exactly 200mv?

 
That's what I did. Do you have an Etymotic IEM handy? Or something else who's output you know at 200 mV?

 

[CraftyClown]

I do yes. I have my ER4-XR.

 

[csglinux]

  I do yes. I have my ER4-XR.

Same one I have According to the spec sheet, the ER4XR is supposed to output 104 dB at 1 kHz, when supplied with 200 mV. But as castle says, you may find the little certificate that came in your ER4XR box has a slightly more accurate read for your particular set.