My question is, whether as pure buffer they will retain their characteristics or their subjective/comparative ranking/preference in this 5 buffers situation where one opamp may suffer "more" than the other.
The noise with clip A is very high such that (from a distance) the vocal intelligibility is not clear at all. But overall I found it the most musical. I'm not sure whether I will still favor it if it is a normal buffer, because in this 5-opamp buffer situation I don't think JFET input opamp will survive, especially with recording that require dynamics/transient (but Frank seems to favor JFET and put the most dynamics at the bottom of the list??)
Jay, the way I do these sort of exercises is to quickly grade into better, worse, and dunno - gives me a rough initial sort, then I pick two adjacent ones from that grading, say listen to the "lesser", then go to the "greater" - is there definitely at least some notch upwards in quality? Or am I disappointed, feel that I've gone a bit downhill? If so, I swap those two, and then pick another adjacent two, not necessarily overlapping with the previous pair, and repeat the exercise. When I get to the point of getting confused with listening to adjacent ones, finding some aspects better, some worse comparing one to the other of the pair, then I'm essentially done. Which doesn't mean I've lost my way, 
- if I pick one from towards the top, compare it directly with one towards the bottom, then the difference, and which I prefer screams at me ...
- if I pick one from towards the top, compare it directly with one towards the bottom, then the difference, and which I prefer screams at me ...
The choice was very scientific. I looked in the drawer and saw that I had a big pile of them.
Sounds perfectly reasonable to me.
Now listen up folks ! I've had another listen to these and compared just one of these files to the master using Foobar and I only scored 4/8 showing I was guessing. I'm going to try it again later but it seems that if there are any differences between these opamps, then as used in this configuration and at these levels the differences must be very small.
Are you all finding this, that it is difficult to separate them just by listening ?
The list of devices means little to me, I would not attempt in any way to associate what I'm hearing with the particular items,
Interesting, because from many listening tests I've always found that your preferences have always been correlated with numbers, or very objective. As if, you know exactly which file is which item. For example, from your ranking, it seems it is highly correlated with distortion. So, based on your ranking, if I associate it with distortion then I'll get more or less (this is your ranking, which I believe will not be too far from the "which is which"):
1. D=LM4562 (0.00003%)
2. E=LM833P (0.002%)
3. F=TL072 (0.003%)
4. B=TLE2072 (0.013%)
5. C=NE5532 (Unknown but must be bad)
6. A=RC4560 (0.05%)
For me, there are a lot of subjectivities involved, such as "musicality" which might be a function of several variables but distortion. My ranking is very subjective (especially with D&C), I believe we will not be able to see any pattern here:
1. A
2. B
3. E
4. F
5. D
6. C
But from the opamp list, in real world implementation, I have always avoided NE5532, and I have always like TL071 (and OPA134). Not familiar with the rest.
And what interest me is the possibility to find out the correlation between what we hear with numbers. In this test, tho, I guess it would be difficult because I don't understand the effect of the 5-buffer circuit to each of the opamp.
IME with critical listening, most if not all implementation have been in gain stages, so I often asked myself, what criteria is important for a buffer (instead of unity gain stability of course), so I can use the right opamp at the right place (in a complex multi-opamp circuit such as crossovers). May be JFET-input for buffer and BJT-input for gain stages, I don't know.
Last edited:
You have to familiarize yourself first with the sound, then recognize the difference, only then you can do an ABX (That's why I easily got 8/8 at first trials). For example, and the question is, what have you recognized as the difference between the 2 files before going with ABX?
I guess it would be easiest to distinguish original versus A or some other opamps that run out of steam when producing that fast transient drum.
I could hear statistically significant differences between D on the one hand, and A and F on the other. I could not distinguish between D vs. B, C and E, or I didn't try hard enough and earphones get hot quickly on a warm summers day.
I thought A and F sounded a bit wooly compared to the others.
This being said, none of them introduced even a flint of unbearableness in a song I used to like before yesterday.
I thought A and F sounded a bit wooly compared to the others.
This being said, none of them introduced even a flint of unbearableness in a song I used to like before yesterday.
Oh dear, has the magic worn off with repeated play. It will come back. Your results on the whole
are great actually.
Got the files.
And thank you for doing this!My question was more succinct: I was wondering if there were 7 (or more) files where one or more of them were Micromega Stage 2 player > ADC (eliminating the entire opamp chain), as a control (or multiples to confound the testee and see his/her consistency). No worries there isn't--it would have been a very useful end point (more so than a levels matched straight rip of the song as at least it contains any effects of the DAC/ADC) to determine audibility of any/all of these opamp chains and nothing.
Again, I do want to reiterate that I'm very grateful for the test, and that my point is but a minor nit, although perhaps one useful for adding in future experiments.
Yep, there are always "tricks" like this to help distinguish ... but as you say, 'which is the better'
? The answer, for me, is to forget completely such notions - only concentrate on whether the sound strikes one as 'right' ... which is equivalent to there being nothing in what I hear that keeps reminding me that I'm listening to audio playback, as compared to experiencing a musical event ... The good ol' reliables: piano, and voice. Just focus on the texture of those sounds - does it just sound like a radio playback of those 'instruments', or does it have a hint, or better, of the real thing ...
I don't do special pleading of some idea about the real thing. Feel free to ascribe your results however you wish, but I chose not to.
I'm not sure we have the same interpretation of "wooliness", but here is my list based on increasing wooliness:
1.C
2.A
3.B
4.D
5.E
6.F
After that I did an ABX between A&C, B&D, E&F.
Yes, both of them were good. But I felt fatigue with D, so I put D down on the preference list (#5). B is objectively good, and subjectively I had no problem with it, so I put it in #2.
Objectively A has many fidelity problems, at first impression it was the worst, but then I found the musicality and tonality is the closest to what I have used to with the recording, except the bass was "hard". I can follow the instruments and vocal. Based on this I thought it was NE5532. But in the end I have always found NE5532 boring/"fatigue"? So listened to it longer than the others and I surprised that I had no fatigue, but the opposite of boring, due to its musicality. So it is #1.
Comparing A with C, it was easy to know that C was much clearer and well defined. But I couldn't follow the music, sounded wrong. On longer background listening (half an hour) I felt the fatigue and wanted to turn it off. I thought this was LM4562 (similar signature with LM4562 in previous test). So it is #6.
E and F, they have similar wooliness characteristics (F the most woolly). I guessed they must have similarity in specs, either both JFET, or both cheap. Listen to them briefly and I found E is enjoyable, so E #3, F #4.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.