Yes, there were a few comments I recall reading about it being "phasey" or something. I should have had a close listen at that stage suspecting a problem but.......... 😀
IMO, slow amp should be aggressive at trebles such as violin. I'm glad to know that something is wrong with previous orange because orange cannot be 741. I guess orange is one of the fastest. I will find out today.
Now pomegranate is aggressive. It could be because of low slew rate. But 4558 is also slow but i believe the 4558 is a new production, not vintage (even CA4558 cannot be as good as CA741 i believe). Generic 741 will not sound like pomegranate either.
If Mooly could reveal the "wild card" I would try to find out which opamp is which file based on their parameter and perceived sound quality.
I'll let it run a wee while longer yet 🙂
Some interesting points and observations you make there. The "chalk and cheese" bit is interesting because I often see in other threads comments such as "the xyz" opamp just has no bass or no midrange and your left thinking "what 😱" what on earth are you hearing.
OK. I'm officially on record. I cheated and looked at the files - because I can't really tell by listening. Mooly is making this pretty good. One of the files I thought was one thing has the channels reversed which completely changed my (lack of ) opinion. I THINK I know the unmodified file but after that ??? And I may be wrong there too.
G²
It's not good practice to visually determine the difference of audio files. Wave shapes can look very different but still sound exactly the same.
I have made some examples so people can experience this for themselves.
The sound files can be uploaded here: Zippyshare.com - waves.rar
Point of topic: Never use visual clues when you want audible clues.
I have made some examples so people can experience this for themselves.
The sound files can be uploaded here: Zippyshare.com - waves.rar
Point of topic: Never use visual clues when you want audible clues.
Attachments
Indeed 🙂
Get the best headphones you can get and listen to a square vs sine wave. Once you get above a certain frequency (that is pretty low in the scheme of things) then sine vs square vs triangular all sound the same. Your age and hf hearing cut off point will determine what that frequency is.
If you sweep a square wave up in frequency, the harsh raspy tone gradually becomes smoother and smoother until you find that last trace of "audible harmonics" dissappear completely. Try it 🙂
Ok Cliff, I have not fully followed this thread, nor have I listened to the files....yet.
What do you mean by 'the elephant in the room " in this case.
Dan.
Another Blind Test For Distortion
I’m not well since Saturday evening so as expected I couldn’t do the blind test for musicality well on Wednesday (made mistake thinking that a file couldn’t be banana while in fact it was). But blind test for distortion is much easier so now I have a few results.
From listening all files I could hear one file stood out with the most distortion (Banana) and three files with the least distortion (Orange, Kiwi, Peach). So I selected 3 files for blind test (Banana as the worst, Orange as one of the best, and Pomegranate).
More than 5 blind test cycles were successful (100% accuracy) and was very easy. It was clear that the ranking based on the lowest distortion is: (1) Orange (2) Pomegranate (3) Banana.
Then I replaced Pomegranate with Pear and did around 8 cycles or more. On the first cycle I made mistake thinking that Pear was Banana but after hearing the Banana it was clear that Banana is worse so for another 4 cycles where I have to choose between Banana and Pear I made no mistake. Of course, when it was Orange I could easily know.
So it was clear that the ranking based on the lowest distortion is: (1) Orange (2) Pomegranate (3) Pear (4) Banana. (To be continued tonight)
My plan was to rank the files based on:
(1) DISTORTION of music#1. The lowest should be original, next best should be expensive opamp like TLE2072. The worst should be old opamp like 741, 4558 and 4560. Here there is issue with how the files are having different SPL etc.
(2) TREBLE/VIOLIN AGGRESSIVITY AT HIGH SPL. The aggressive ones will be linked to low slew rate opamps.
(3) DETAILS AT COMPLEX PASSAGE OF MUSIC#4. The blur ones will be associated with low slew rate and cheap opamps.
(4) BELIEVABLE SOUNDSTAGE of music#3 or music#4. I don’t know how to associate this to the opamps except that it will be favored by me.
(5) ACCURACY OF VIOLIN HARMONICS. In some files the violin doesn't sound like string instrument.
To guess which file is which opamp, it will be good if there is data about input noise, unity gain bandwidth and CMR/SVR. But I will start with slew rate:
TLE2072 (45V/us)
OPA134 (20V/us)
LM4562 (20V/us)
TL072 (13V/us)
NE5532 (9V/us)
4560 (4V/us)
4558 (1V/us)
741 (0.5V/us)
I’m not well since Saturday evening so as expected I couldn’t do the blind test for musicality well on Wednesday (made mistake thinking that a file couldn’t be banana while in fact it was). But blind test for distortion is much easier so now I have a few results.
From listening all files I could hear one file stood out with the most distortion (Banana) and three files with the least distortion (Orange, Kiwi, Peach). So I selected 3 files for blind test (Banana as the worst, Orange as one of the best, and Pomegranate).
More than 5 blind test cycles were successful (100% accuracy) and was very easy. It was clear that the ranking based on the lowest distortion is: (1) Orange (2) Pomegranate (3) Banana.
Then I replaced Pomegranate with Pear and did around 8 cycles or more. On the first cycle I made mistake thinking that Pear was Banana but after hearing the Banana it was clear that Banana is worse so for another 4 cycles where I have to choose between Banana and Pear I made no mistake. Of course, when it was Orange I could easily know.
So it was clear that the ranking based on the lowest distortion is: (1) Orange (2) Pomegranate (3) Pear (4) Banana. (To be continued tonight)
My plan was to rank the files based on:
(1) DISTORTION of music#1. The lowest should be original, next best should be expensive opamp like TLE2072. The worst should be old opamp like 741, 4558 and 4560. Here there is issue with how the files are having different SPL etc.
(2) TREBLE/VIOLIN AGGRESSIVITY AT HIGH SPL. The aggressive ones will be linked to low slew rate opamps.
(3) DETAILS AT COMPLEX PASSAGE OF MUSIC#4. The blur ones will be associated with low slew rate and cheap opamps.
(4) BELIEVABLE SOUNDSTAGE of music#3 or music#4. I don’t know how to associate this to the opamps except that it will be favored by me.
(5) ACCURACY OF VIOLIN HARMONICS. In some files the violin doesn't sound like string instrument.
To guess which file is which opamp, it will be good if there is data about input noise, unity gain bandwidth and CMR/SVR. But I will start with slew rate:
TLE2072 (45V/us)
OPA134 (20V/us)
LM4562 (20V/us)
TL072 (13V/us)
NE5532 (9V/us)
4560 (4V/us)
4558 (1V/us)
741 (0.5V/us)
Last edited:
A lot of interesting discussion. My process was as follows:
1. listen to the musical selections
2. pick a favorite
3. vote
4. keep my mouth shut
1. listen to the musical selections
2. pick a favorite
3. vote
4. keep my mouth shut
Hey that was the same process I used. I do believe step 4 was one of the most important steps to not poison the process.
Last edited:
I was thinking of doing the big unveiling later today as interest seems to be waning now.
I think you'll find some surpises 😀
I think you'll find some surpises 😀
What process 😀
I have mentioned that whatever the poll result, it doesn't mean much. From validity point of view, this poll is less valid than any of the many blind tests I have participated in. For most of you taking part in this poll, it is just for fun, isn't it?
Well, I have different view, interest and intention than you do. That's why I posted my result. And only Mooly has the right to decide whose intention should be prioritized.
There has been many debate between subjectivists and objectivists regarding how ears can really hear differences. The conclusion so far is: the golden ears are only fooling themselves and the subjectivists are right with saying that every opamps sound the same (or similar claims).
This debate will continue until next decades because nobody have proof, especially the subjectivists.
That's why I want to try to pin point each file to each opamp so if it is successful it can be a proof that it is really possible to hear differences and hopefully those experts who has terrible ears will understand their limitation.
For me, it is clear that I can hear differences. Otherwise, how can I have such a high accuracy in the blind test? Well, some of you will think that I cheat or something, that's why I want to do it down to detail so no such excuse is possible.
Of course I can post later at the end of the day, but I'm not sure I will have time, so that's why I did it while I have time.
Let me tell you something. Up until this moment, the subjectivists are not joining the poll. And they consider those who join are audio fools. And I'm not happy with that. But that's alright because at least I didn't vote either 😀
The orange not a discrete opamp?
By telling your own results, you invalidate the test for people who haven't yet done it. I was going to do the test but I've now read too many results for me to be completely unbiased .. so I shan't bother loading the huge wav files on my slow line.
ANY previous results that are traceable to specific presentations WILL skew the results for people who are yet to do the test.
In my DBLTs, the listeners are always tested separately and are not allowed to discuss it with each other until the final results are out.
And you NEVER tell the listener what he is supposed to be listening too so he doesn't try to guess what is what.
Mooly, would it be possible to do another test after you release the results of this one? You could just change the names and order. Can the poll be one where you only see the results after you've done it and can we post all results on another thread. Discussions on the original poll thread strictly limited to stuff that ISN'T specific to identifiable presentations.
This would get better valid data out of the test. Please feel free to slip in some repeat presentations to weed out the deaf Golden Pinnae 🙂