Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: How do perform a controlled listening test at home (Read 13002 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

How do perform a controlled listening test at home

Reply #25 –
Quote from: Rich B on
That sounds like a lot of work and a lot of uncertainty involved. How would I know whether my ADC was good enough not to mask differences?


You can try an ABX between the D/A-A/D loop only (no amplifier), and the original sample first. This way, you can test the transparency of your recording equipment. Or extract the difference signal (with something like Audio DiffMaker) instead of a listening test. You can also perform frequency response, noise, THD, etc. tests; RMAA is popular for this purpose, and is easy to use. Although it is somewhat limited in its functionality and it has some bugs, it can reveal obvious problems like a frequency response that is not flat enough.

Another approach is to record the output of a single device, and then ABX it against the original sample to test its transparency. I have some examples of this here. This way, the possibility of a non-transparent ADC makes the results pessimistic, rather than masking the differences (which in practice is unlikely in my opinion, by the way).

If you are worried about differences getting masked, then the most likely source of such problems is your headphones/speakers and listening environment. That is, transducers having a rather non-flat frequency response, high distortion in the low frequency range, possibly high ambient noise, and - with loudspeakers - room acoustics. Compared to those factors, even a fairly cheap ADC can be very transparent.

You should be careful, however, that your testing equipment does not interfere with the test. In loopback recording setups, ground loop issues are not uncommon. The ADC also needs to be able to handle the output of the amplifier without clipping, but also have a low enough noise floor to be able to record a low level signal (for example, a headphone output driving a sensitive IEM) without adding audible noise. Another potential source of problems is when the DAC and ADC do not share the same clock, and thus the recording slowly "drifts" over time (this makes difference extraction difficult, for example). And, of course, you cannot record the output of a balanced/bridged amplifier with a typical unbalanced sound card line input. Finally, you need to make sure that software related problems are avoided (bad mixer settings, buggy drivers, unwanted sample rate conversion by Windows drivers, buffer over/underruns, etc.).

How do perform a controlled listening test at home

Reply #26 –
It seems like such a task to do, and there are lots of things that can go wrong.

I would love to test a few amplifiers. But that would involve an ABX box. I don't think any are made today, but it sure would come in handy! I wish there was an easy way to test the differences between amplifiers that did not involve 101 steps. The ABX box method would probably be the best.

How do perform a controlled listening test at home

Reply #27 –
Quote from: Rich B on
I have friends who have bought expensive gear who have challenged me to prove that they can't hear differences between equipment.

This should have been addressed earlier.

First, ABX demonstrates whether someone can hear a difference with statically significance.  A statistically significant result does not prove they can't.

Second, and more importantly, the burden of proof is on your friends to demonstrate that they can.

How do perform a controlled listening test at home

Reply #29 –
Quote from: Rich B on
Quote from: greynol on
Quote from: Rich B on
I have friends who have bought expensive gear who have challenged me to prove that they can't hear differences between equipment.

This should have been addressed earlier.

A statistically significant result does not prove they can't


What does it prove then?


Nothing at all.

An ABX test can only prove that you hear a difference.  All other results are inconclusive.  Proving negatives is actually quite hard.

How do perform a controlled listening test at home

Reply #31 –
"Proof" is a too strong word imo. You can still guess your way to a statistically significant result.

I'd say a positive ABX result is evidence that you could very likely hear a difference.


You could disprove the claim "I cannot hear a difference between A and B" with a solid ABX result, but proving "nobody can hear a difference between A and B" is quite a bit harder..
"I hear it when I see it."