Thanks. And thanks to Mooly for his idea I had never imagined was possible ( and now with the CFA versus VFA idea).
I had participated in blind tests and I had many audio friends. True audio friends which you can call maniacs. For some in different location of the world, golden ear thing is a lie but I know for sure it is not. Many times In a bt I thought that I must be the only one who can hear a difference but actually almost everyone can. Many have tested me and I passed, like putting a rubber mat on top of an amp! Things like cd player sounds better with cover removed, that's true.
I didn't know that I can hear differences with this kind of test tho. Yes, I could pick banana as my favorite very quickly and was in a hurry using $5 headphone, but I'm not sure if I can hear so clearly if not using my speaker (which is a solution to a long time mystery of what a good speaker should be designed to be).
I mentioned that I didn't think a headphone will do it. Why? Because one of the newest things I learn in audio is about believable soundstage. Soundstage cannot be bigger than it should. Cheap cdplayers will not sound like mega bucks ones, but tube effect was nice. It can make vocals float. And that is a problem because in a believable soundstage the singer must stand on the floor, and so is piano. It is i believe only a few understand well. And I just don't know how this can be possible with headphones as they have weird soundstage. In other words, the performer should be in the room, not the listener on the stage.
Thanks for that, Mooly ... I'd like to see if anything makes sense in terms of the results ...
My results were all over the place, if one were to pick an order based on theoretical merit, from published spec's, etc. Many of my thoughts mirrored that of Jay musings - but my first choice was bizarre, the TL072! 2nd was Direct, 3rd was LT1229, so that makes more sense. Overall, not a good correlation to what "should be".
Unfortunately the path is rather poor, the 2nd test with the -20dB track demonstrated that - the losses are quite severe. Since I was listening for the sonics that were most pleasing, perhaps there was something that the TL072 was doing right, at the moment when the test was done, that was countering the other issues, for me at least.
To be expected, the clocks don't match, so doing a simple diff'ng will get you nowhere - more involved analysis - perhaps Diffmaker can come through this time,
- will be needed.
Last edited:
Ahhh.... that changes everything (for me). I'll have to do some blind ABX testing against the original file, the loop and the opamps to hear if I can tell. Maybe I can't. Sure had a hard time telling the opamps apart.
Well, for me, the status of the path at the time of the recording seems to be key. I have just tried running the Master vs. Direct vs. TL072 in a better player, Nero. The order is Master, TL072, Direct (!) - Direct is definitely more 'contaminated' than the TL072 run, the latter is cleaner, closer to the Master.
Hehe if you look at my ranking based on distortion, you will see that the banana was the most polluted and peach was the cleanest. But I have suspected the cause. Kiwi was the next cleanest. Is that the tl072?
Perhaps, perhaps not ... what I would like to know is what aspect in the graphic is telling you "something" - personally, even knowing the answer, I can't see anything in it that identifies anything.
Eww it's CA741! Mooly was right that it performed much better than its "spec". But 4558 was the worst, in this blind test or in the real world.
To lay it out totally, this is what I got:
Pear - TL072
Banana - Direct
Peach - LT1229
Kiwi - 741
Raison - 4560
Apple - TLE072
Pomegranite - LM4562
Orange - OP134
Satsuma - 4558
Damson - NE5532
I went 3 rounds of grading, so that I was comfortable that a track above another was in some sense an improvement, "better" to listen to than the one below it.
Thanks for doing the poll Mooly.At the moment this one test has been enough work tbh. There's a lot of hours work in there to pull all this together
One thing we could have done differently was to have the poll hidden so no one could see the results of the voting.
It's likely I'm the only one who has an idea of the HUGE amount of time & trouble it took for you to organise this test .. cos I've done it for real.
The worst is when you carry out such an expensive series including involving your expensive true golden pinnae blind listening panel and later find out that some small mistake on your part invalidates all your results. This could be simply 0.2dB difference in level.
______________
We definitely have to hide the results from being accidentally seen by people who truly want to know if they hear differences.
Alas, it will be more difficult to get people NOT to post their results in the general thread.
______________
We have to seriously avoid people trying to guess what is being heard. When we do the next test we'll just tell everyone we are evaluating 10 different brands of 741
Of course we might actually do this for real and see how many people assign Chalk & Cheese differences between the same 741 presented 10 times.
It is interesting that some people seem to be so sensitive to level - I barely notice such things, even 3,6dB changes hardly register. What I listen for is the 'muckiness', the unpleasant edge intruding, congestion of sound when things get busy, flatness or dullness overlying the texture of the replay -- these sort of things I'm very sensitive to ...
Here's a fun thing. Since we now have the master file (not looped) I thought it would be fun to hear if it's identifiable. It is, very easily - it's about 5dB lower than the other files. Well that's no fun.
So I took Banana, Orange & Pomegranate and reduced their volume to precisely match the master file. All saved as 24 bit files. Now it's so hard to tell them apart that I gave up. Not on headphones, not on speakers, not with Solid State or with Valves. Oh, I keep thinking I can hear the difference, but Foobar ABX says 70% chance I'm guessing. Oh well.
Maybe if someone pointed out the differences I could get it - but on my own, no way.
Pretty amazing demo. Thanks Mooly!
So I took Banana, Orange & Pomegranate and reduced their volume to precisely match the master file. All saved as 24 bit files. Now it's so hard to tell them apart that I gave up. Not on headphones, not on speakers, not with Solid State or with Valves. Oh, I keep thinking I can hear the difference, but Foobar ABX says 70% chance I'm guessing. Oh well.
Maybe if someone pointed out the differences I could get it - but on my own, no way.
Pretty amazing demo. Thanks Mooly!
A good example of what my previous post was about, that aspect went straight by me, never registered. Now, by all accounts,Here's a fun thing. Since we now have the master file (not looped) I thought it would be fun to hear if it's identifiable. It is, very easily - it's about 5dB lower than the other files. Well that's no fun.
, that should have made the master sound much worse to me, heavens, 5dB softer!! Instead, it was clear that this was superior to the other two ...Wind up the volume, right, right up - listen for the congestion in the choral climaxes, the ringing of the single notes of the piano, how bitey, edgey the violin is, in the jazz piece the clarity of the drummer's cymbals in the busy stretch, and precisely the tonality of the final cymbal crash at the end.
To me this Foobar ABX thing is like playing adjacent notes on the piano, completely at random, and demanding that one can successfully say higher, higher, lower, higher, lower, etc, and get it right every time. I would completely mess it up, trying to do this. What I do is play lower, lower, lower, lower, lower, ... higher! The difference hits you straightaway, and you get a ranking ... then you go downwards, and that should confirm the ranking. Do this a couple of times, with major time intervals inbetween, to confirm the assessment and the ranking is locked in ...
Edit; where the ABX way goes down the tubes is that your mind "knows" what the slightly better version sounds like, and fills in the difference when the poorer one comes along - I've had this happen to me over and over again, it ain't the way ...
Last edited: