[castleofargh]

1dB is the limit of spl differences I can hear on a 1khz test tone. Probably more like 3db on music, but I haven't tested that. About .05 volts of variance seems to be where volume differences in music are apparent when I'm level matching with my multimeter. I haven't used an spl meter to try and see what that equates to in dB.

no need. but you do need a voltage value and not just the variation as the dB are calculated from a ratio. so the 0.05V variation alone doesn't tell us much.

 

[bigshot]

It's really handy to actually do listening tests and determine rule of thumb JDD's. Then you don't need to do math.

 

[RollsDownWindowsManually]

I have a Question on Level matching for A/B Headphone comparisons.

I have an amateur flat plate headphone measurement rig with a cross spectrum labs calibrated UMIK-1. Is it best to level match transducers using full band pink noise, bandwidth limited pink noise, or 1khz sine waves (or sine wave at some other frequency?)? I have half a dozen headphones that I'm trying out and am trying to discover which ones I like the best without being influenced by unequal volume levels.

 

[pinnahertz]

Band limited pink noise, usually 500Hz - 2kHz. Never sine waves, and full band pink can result in mid-band level mismatch. You want to match mid-band, then notice the differences above and below.

 

[bigshot]

It really isn't that hard to level match with mid band pink noise. Pink noise is better at that than setting EQ curves.

 

[RollsDownWindowsManually]

I finally got around to measuring the output impedance of a few devices' headphone jacks:

Pioneer Elite SC-55 Audio/Video Receiver: 34.53 ohms

Alienware 13 Laptop: 14.77 ohms

Google Pixel XL cellphone: 5.17 ohms

LG V35 ThinQ = 0.94 ohms

Line 6 Tone Port GX/GearBox = 2.79 ohms

Headroom Micro Amp: 0.88 / 0.92 / 0.98 ohms on low/med/high gain

Schiit Audio Jotunheim Single Ended Output: 0.19 / 0.17 ohms on low/high gain


I ended up just using a planar magnetic headphone with a flat 23 ohms impedance instead of a resistor to get a loaded voltage value to subtract from the unloaded voltage value. Then used the formula: Zout = (Rload * (Vnoload - Vload)) / Vload

Edit: realized I could actually measure the impedance of the headphone with my multimeter, it read 22.1 instead of the published spec of 23 so I used 22.1 ohms and it changed my Zout numbers slightly. Also I'm not sure but maybe the 3 foot headphone cable I'm using between the amp and the multimeter adds a little bit of extra impedance to my measurements.

 

[amirm]

I am building a databases of output impedances as I measure/review gear with headphone output. Here is the latest:

 

[RollsDownWindowsManually]

I am building a databases of output impedances as I measure/review gear with headphone output. Here is the latest:

Yeah, I saw that. Very nice, keep it up, I enjoy the measurements! Your measurements are much more accurate than mine I'm sure, I'm just using a cheap OWON B35T multimeter. I think my readings my be slightly high, as Schiit Specs their Jotunheim at less than .1 ohm vs the .19/.17 I measured, and I saw on XDA developers that the Pixel phone was measured at 4.8 ohms vs the 5.2 I measured on my Pixel XL (granted XL is a different model).

 

[eucariote]

Hey mate, if you still want to run an blind test (have someone else choose a source by flipping a coin, you judge source a or b or same/different), post your results and I'll run the stats for you.

 

[RollsDownWindowsManually]

I just measured the headphone output of an ASUS ROG Maximus X Hero motherboard on a new computer and the Output Impedance was 80 ohms! It still sounds OK with my planar magnetic headphones though (flat impedance and I'm not sure how much damping factor matters with them).