how do you "listening test"?

[admc007]

I have 2 monoblock gainclones. When I want to evaluate a tweak, I build it on one gainclone only. I then play say 10 seconds on a CD on repeat (A - B loop) through one channel only, then repeat on the other channel.

So if I do a mod on the left channel, I listen to the left channel only for a few repeats. I then take the left channel interconnect and plug it into the right channel gainclone and listen to a few more repeats. This way it is easier for me to evaluate the difference. Then I have a good think about whether it is better, worse or just different!

I think when the difference is quite small, you have to listen to just a few seconds only on repeat. Then you have to quickly repeat the same test without the tweak. If you leave too long between the listening tests, your memory starts to play tricks on you. By tweaking just the one channel and then swapping the interconnect into the untweaked channel, you can properly tell what the difference in sound is.

The speakers can also be checked out this way.

The A - B loop is not a great way to listen to music but it is useful for a listening test!

I read so many endless technical arguments on this site that go on and on, and I am not qualified to pass judgement on them. However, I sometimes wonder if someone would not just build something in order to test it by actually listening to it. If it's a simple thing to build in just 30 minutes, then just try it and see if it sounds better?!?

I could spend years studying books and reading endless opinions (and get myself more confused!), but sometimes I think it would be easier to just build it in 30 minutes and then listen to it!

 

[jjrenman]

I often make a change to one channel only and then listen in stereo. Over the course of many listening sessions I will swap which side the modification is on. After a while I lose track of which side it's on but I keep listening till I can easily recognize the difference in the left/right channel.

I do differ from you in that I avoid listening to just a few seconds of a song and then switch. Reason being that IMO although a few seconds can let you hear a difference it is not enough time to tell if it is a "false positive". IOW some changes will sound good for a short period of time but become fatiguing with longer listening.

 

[Cornelis Spronk]

What is lovely question for a audio DIY forum.

When listening to speakers I listen for considerable time, and identify what is annoying. Whatever it is, the annoyance gets worse to my ear, the longer I listen to it.

Often, on first impression, a new speaker sounds interesting, which may not mean good over the long run.

I agree that short listening identifies differences, but does not immediately identify which is better.

The differences also seem to go into long term memory. I even tend to form opinions of what I hear, compared to what I heard many years ago. Difficult to verify this with a double blind test.

Most important, is to make comparisons of what I hear, to what I hear in a real live acoustical performance.

 

[admc007]

I can understand that a certain design will be good from an engineering point of view. I understand that there are certain guidelines or rules of thumb, but only by doing a listening test can I ascertain if it's actually "a big deal" in terms of the sound!

Engineering is all about compromise. If I have an idea of the relative significance to the actual sound, then I can make a better informed choice about which compromises to make.

I came across a thread about one bridge rectifier versus 2 bridge rectifiers. There were arguments for and against and I didn't read the whole thing cause there was just too much of it! I can't help wondering, surely someone should just build a channel with one bridge rectifier and a channel with 2 bridge rectifiers, then just listen to which sounds best?!

I don't think it is good for the diy audio to have arguments that go on and on without any resolution. Imagine you were working in R&D and there was an argument that just went on and on without resolution.... I just don't think high tech companies would tolerate that sort of thing!

I think we should take the principles of experimental design. In experimental design you let the data do the talking. In our case, the "data" is the sound.

An engineering company did some kind of experimental design on breakdown voltage (or something like that) and they measured some other parameter in order to infer the voltage breakdown. But surely they should have just measured the breakdown voltage directly? It is important in experimental design to measure the parameter that is most pertinent, most relevant, and in our case that parameter is sound quality, even though it is objective!

 

[jjrenman]

You state "then just listen to which sounds best".... that statement is where the rub comes in. I try to stay away from any general statement as to which sounds best. There are a few things that we all agree upon but past that is gets to be personal as to what we are optimizing for.

So to avoid arguments that go on and on we have to stay away from defending what we personally think sounds better and instead share the differences we are hearing. It then leaves it up to the individual as to what is best for them.

IMO in this hobby to many individuals will defend their preferences instead of helping each other figure out what our preferences might be. Can you imagine a wine tasting club where everyone staunchly defended their preferred wine? Would it not be better to help each other out to become educated to all the different types of wines so that one can not only recognize an excellent wine in any style but also figure out what they prefer?