[pinnahertz]

You can make a frequency correction chart for a budget DMM.
All it requires is a CD player and a test tone disc. Most CD player are more than accurate enough. At least from 30 Hz to 15 kHz.
Burn a WAV not a MP3.

Why not just play the tones out of your computer and not bother with the CD? Audacity generates all the tones you need.

You can make more accurate measurements with software and a sound card. Some even have calibration routines that flatten out the response of a less than perfect super-cheapo sound card.

Look at: https://trueaudio.com/rta_abt1.htm (there's a free version)
and of course REW: https://www.roomeqwizard.com/

I can't think of any reason someone going that far in wouldn't pony-up for a resistor assortment kit
and a set of clip leads. This stuff isn't expensive, and you can do so much more when you have the right tools.

 

[castleofargh]

I bought some clips for nothing that really look a lot like in your link, those suckers added about 1ohm each. I ended up soldering some speaker cable between the crocos to be able to measure stuff without too much impedance change.
can't ever tell enough good of REW, getting so much for free is amazing. but somehow many curious amateurs around here seem to get scared away from it and end up with ARTA instead.

 

[pinnahertz]

I bought some clips for nothing that really look a lot like in your link, those suckers added about 1ohm each. I ended up soldering some speaker cable between the crocos to be able to measure stuff without too much impedance change.

Odd. I keep these in my kit, use them all the time, they measure a fraction of an ohm. There are some cheap ones that use extremely thin wire, but they're so short I wouldn't think they'd be an ohm. You must have gotten some really awful ones. Perhaps look for the "heavy duty" ones, they look like they have like 20ga wire.

can't ever tell enough good of REW, getting so much for free is amazing. but somehow many curious amateurs around here seem to get scared away from it and end up with ARTA instead.

Yep, REW rocks. So much so I've never even bothered to evaluate ARTA. I use TrueRTA too (paid full version), less often, but nice for quickie measurements and the scope function is pretty good. I also have WinRTA from the THX days (it was the D2 replacement), but it was expensive. I got into about $2K for the mic (Earthworks M30 with cal charts), sound interface (originally a Sound Devices USB Pre), and software. I still use WinRTA for spacial/temporal averages at 1/12 octave because it can integrate over a specified interval, when I don't have the time for multi-point REW stuff doing commercial systems.

Fuzzmeasure is pretty cool too.

 

[castleofargh]

oh yeah the crocodile stuff I got must be different then, I still have one or 2 untouched that I kept as a freak curiosity. I can't remember where I bought those, hoped it was amazon because they keep track of everything... EVERYTHING. but it didn't come from there.
I have trueRTA too that I sometimes start 10times a day just to check the silliest things when my obsessive double check mania kicks in, and then I won't use it for 6 months. I have a few other stuff(all free versions because I'm cheap) that I pretty much never use and often forgot I had when I would actually need them. but keep in mind that I am a curious noob and nothing more so my uses are obviously at my level. but I'm losing myself again way beyond OP's matching levels and ABx, sorry about that.

 

[Speedskater]

Of course you could use a computer if you are sure that no DSP functions are enabled.
A decade ago, Benchmark & Stereophile found that Windows was resetting audio settings.

 

[RollsDownWindowsManually]

I finally got a replacement battery for my old multi-meter, it only goes to 00.1 AC volts resolution on the lowest setting, so I don't think that's low enough to give me meaningful information. I was getting the same voltage, around ~0.5 both with and without the 75 ohm resistor on various headphones playing a 100hz sine wave tone in audacity. I ordered a new multimeter with better low voltage range to see if that helps.

 

[RollsDownWindowsManually]

So I went ahead and purchased the A/B switcher and have done some sighted testing with myself trying to differentiate the sound between my cellphone's headphone output and a pretty "good" headphone amplifier/DAC, the LH labs geek out 2 pro infinity.

I'm using the periodic beryllium IEM's.

Despite my best effort to use the multimeter for level matching, I can't get reliable mili-voltage readings from it, so I'm using the Radio shack SPL meter, playing the same test files of pink noise/sine waves to level match the two sources within 1 db, which is the best I can do given the volume graduations on the devices.

I'm going to have my wife randomly assign A and B to be X and/or Y then administer a blind test foobar ABX style, playing the same tracks from foobar, one on my PC using the LH amp through USB and one on my smartphone using its on-board headphone out behind a barrier so that I cannot see.

I have a schiit Jotunheim and an older Headroom Micro stack I'd like to try it with as well if she has the patience. I'll say "Play A, Play B, Play Y, Play B, Play X, Play A" etc etc and then make a guess that A is X or Y and B is X or Y like I would in foobar's ABX. With the IEM's in I really can't hear the clicks of the switcher, so there is no way for me to cheat by counting clicks even if I wanted to. I know that it has been said that the tester needs to not know the correct answers for it to be valid, but I really don't have any else I can ask over to torture this way.

In my initial sighted tests, it's difficult to tell the difference between the cellphone and the amp when level matched.

I know it's not scientific because there are a lot of variables other than just the amps I couldn't eliminate (computer vs cellphone, clicks from the switcher, level match precision), but at least the source files are the same.

 

[bigshot]

In my initial sighted tests, it's difficult to tell the difference between the cellphone and the amp when level matched.

Odds are the differences are going to be even harder to detect with a blind test then. When doing comparisons like this, you have to decide what level of testing is good enough to come up with an answer that suits your purposes. If you're trying to determine if one is better than the other, you've probably already gotten your answer... unless you enjoy the testing process. It seems some people really do. I'd rather listen to music myself.

 

[RollsDownWindowsManually]

Odds are the differences are going to be even harder to detect with a blind test then. When doing comparisons like this, you have to decide what level of testing is good enough to come up with an answer that suits your purposes. If you're trying to determine if one is better than the other, you've probably already gotten your answer... unless you enjoy the testing process. It seems some people really do. I'd rather listen to music myself.

I did a best effort abx test with my wife administering, and only got 18% correct between the cellphone headphone jack and LH labs amp/dac using the periodic beryllium (Be) IEM. I did 11 trials and only got 2 correct. Was going to do 30 but knew I was probably doing very poorly so ended it. Granted I had to try and use a whole song on repeat playing on both devices, didn't have the luxury of playing a 5 second section over and over. Tried to listen for differences like punchier bass, lower noise with the amp but it the difference wasn't noticeable in a repeatable way. At least I know I have no reason to be lugging around a portable amp with these headphones.

 

[castleofargh]

So I went ahead and purchased the A/B switcher and have done some sighted testing with myself trying to differentiate the sound between my cellphone's headphone output and a pretty "good" headphone amplifier/DAC, the LH labs geek out 2 pro infinity.

I'm using the periodic beryllium IEM's.

Despite my best effort to use the multimeter for level matching, I can't get reliable mili-voltage readings from it, so I'm using the Radio shack SPL meter, playing the same test files of pink noise/sine waves to level match the two sources within 1 db, which is the best I can do given the volume graduations on the devices.

I'm going to have my wife randomly assign A and B to be X and/or Y then administer a blind test foobar ABX style, playing the same tracks from foobar, one on my PC using the LH amp through USB and one on my smartphone using its on-board headphone out behind a barrier so that I cannot see.

I have a schiit Jotunheim and an older Headroom Micro stack I'd like to try it with as well if she has the patience. I'll say "Play A, Play B, Play Y, Play B, Play X, Play A" etc etc and then make a guess that A is X or Y and B is X or Y like I would in foobar's ABX. With the IEM's in I really can't hear the clicks of the switcher, so there is no way for me to cheat by counting clicks even if I wanted to. I know that it has been said that the tester needs to not know the correct answers for it to be valid, but I really don't have any else I can ask over to torture this way.

In my initial sighted tests, it's difficult to tell the difference between the cellphone and the amp when level matched.

I know it's not scientific because there are a lot of variables other than just the amps I couldn't eliminate (computer vs cellphone, clicks from the switcher, level match precision), but at least the source files are the same.

the general idea is indeed that once you've controlled most variables, the differences are likely to be less than what you first thought subjectively while doing sighted evaluation. and that's what makes so many people falsely conclude that controlled tests are flawed and sighted listening works better. because they "notice" bigger differences when the brain can run free and make up stuff as it keeps going uncontrolled and self confirming previous preconceptions thanks to the constant support of visual cues about the gear used.
so we say do controlled testing because it's the only way to get confidence in what you're hearing. and they say don't do those, they're worst than just listening. ^_^ the main difference in judgement is awareness of all the potential biases, or lack of.

about the voltmeter, remember that double the voltage is only +6dB, you probably get just as good if not better accuracy from the voltmeter than what you get from the SPL meter. it's just that in one case you have a log scale. if you're as lazy as I am check this http://www.sengpielaudio.com/calculator-gainloss.htm for conversions.

 

[bigshot]

At least I know I have no reason to be lugging around a portable amp with these headphones.

I did a similar test with my Oppo PM-1s with the Oppo HA-1 as opposed to just plugging straight into my iPod. I came to the same conclusion.

 

[RollsDownWindowsManually]

Finally sat down and read the manual for my new multimeter to get it working.

I can set the level on both sources to within 8 millivolts (80 mV on the amp vs 72 mV on the phone) playing the same 99hz sine wave test track (because of the way the volume steps work). This works out to within 1db on the SPL meter, playing a 99hz sine wave at 70db (mic right against the ear nozzle). Maybe I will do another ABX of an amp against the cellphone tonight with a different song, just for fun now that I can match levels with the MM (not that it gets any closer than before due to the volume steps on the devices, but it's a nice crosscheck).
-------------------------
I have tried to calculate the output impedance of my cellphone headphone jack with it - when I put a 75 ohm resistor in line with my 32 ohm IEM, the voltage increases by 8 millivolts when playing the 99hz test tone (from 72 to 80 millivolts). I don't understand this. I thought the voltage should decrease when I increased the impedance with a resistor.

 

[RollsDownWindowsManually]

Maybe the problem with my "resistor" was that I was using an Etymotic ER-4P to 4S adapter cable, which I had read was just a 75 ohm resistor. Maybe it's not.

 

[RollsDownWindowsManually]

Question for experienced measurement types:

Is it correct for the measured voltage of an amp's output to be the same regardless of frequency being played? I measured my Sennheiser hd800 with a 60hz, 200hz, and 1khz sine wave at the same amplifier volume setttings and the readings were all pretty much identical at .992 volts. I was trying to level match two different headphone amps (a tube hybrid and a solid state) to A/B compare them (sighted). I was just using a Y splitter with one side connected to the headphones, and using my multimeter to take readings off the other side. I couldn't hear much difference, if any when I level matched the amplifiers this way, granted it was a sighted and unscientific test.

 

[pinnahertz]

Question for experienced measurement types:

Is it correct for the measured voltage of an amp's output to be the same regardless of frequency being played? I measured my Sennheiser hd800 with a 60hz, 200hz, and 1khz sine wave at the same amplifier volume setttings and the readings were all pretty much identical at .992 volts. I was trying to level match two different headphone amps (a tube hybrid and a solid state) to A/B compare them (sighted). I was just using a Y splitter with one side connected to the headphones, and using my multimeter to take readings off the other side. I couldn't hear much difference, if any when I level matched the amplifiers this way, granted it was a sighted and unscientific test.

If an amplifier has a very low output impedance and/or the load has little impedance fluctuation with frequency, then yes that's normal. Even though the HD800 has major impedance fluctuations, its overall impedance is quite high, so the output of an amp won't be loaded much at all, resulting in a flat frequency response. And that would be true for the tube amp as well. Such is not always the case, though.

The rule of thumb is headphones should be 8X to 10X the output impedance of the amp, or inversely, the amp output impedance should be 1/8 to 1/10 of the headphone impedance. Under those conditions you won't get frequency response variations much at all.