Which is why, when setting up speakers and maybe doing some minor tweaking on a a system, I always tell my clients to wait until the Placebo/Confirmation effect wears off in a week or two before deciding their next move.
I wonder if Mark's 12 in a row lash up was socketed... Could he sub in different opamps and see how the results compare? That would be interesting.
No. I had a simple 2-mike ORTF setup into a NAGRA. At the other end of the scale, someone had 5 mikes, one for each instrument plus a another pair.
Recorders varied between portable digital ones to restored large Grundigs.
Jan
One moment of insight for me was years ago when I read about Floyd Toole's test in his Harman speaker lab. He invited his speaker design personell to a rating test for 5 or 6 speakers, their own and some competition. He told them it would be double blind, and his people joked that they wouldn't be fooled, they were his speaker designers! They knew speakers, they all had their proverbial 10,000 hours listening under their belt.
Toole started with a viewed test, and specifically asked them to rate the bass performance.
Then he switched to blind (look up the 'Harman speaker changer') and again asked them to rate the bass performance.
When he looked at the results, he found that in the viewed test the best rated bass was for the largest speaker. In the blind test, that was no longer the case and speaker ratings were more in line with measured bass performance.
In the viewed test, one of Harman's own came out best, in the blind test, it was one of the competition.
Remarked Floyd dryly, 'so much for company loyalty'.
Jan
Toole started with a viewed test, and specifically asked them to rate the bass performance.
Then he switched to blind (look up the 'Harman speaker changer') and again asked them to rate the bass performance.
When he looked at the results, he found that in the viewed test the best rated bass was for the largest speaker. In the blind test, that was no longer the case and speaker ratings were more in line with measured bass performance.
In the viewed test, one of Harman's own came out best, in the blind test, it was one of the competition.
Remarked Floyd dryly, 'so much for company loyalty'.
Jan
Yep, sure is... but still we're only asking it for about 1 or 2 volts every 50us, so still well within it's unity gain bandwidth.
Attachments
Last edited:
Yes the mics make a difference, but this is all about preference, and that can be manipulated. Ask any successful marketing executive.
My mikes were $ 50 a pair from the Far East against some multi-kilo-$ Neumann's.
So if actual performance was important, I'd be kicked out right away.
I like to think I didn't cheat but 'prepared the battle field'. Change the odds in your favor. What any good soldier would do.
Jan
Last edited:
Well, you can have ALL the 741s you can eat. I don't think they sound good, and these days, at $1 apiece for an NE5532, I'll take all of those I can eat, and be happy about it.
I guess you missed the part where I said I don't use the 741s anymore.
But it's simple logic that if they were good enough for audio 40 years ago, they're still going to be good enough now. Audio isn't hard... it's the lowest segment of the EM spectrum and only a very tiny slice at that. Just because something sounds high pitched, doesn't mean it's high frequency... play with some RF circuits for a while, you'll get the idea.
Yes the 5532 is better, no argument.
Last edited:
Did. If you want the kind of headroom most +4dbu mixers, tape machines etc have its not enough.
A cascade of unity gain stages is an obvious test-case, but not much use in audio. The longer cascades in mixing consoles give overall gain, and also internal gain in "unity gain" blocks like EQ.
On your existing breadboard, add 220r resistors sum-points to ground so all work at gain of 10.
As you point out, common-mode distortion is a thing and non-inverting is more common in audio. Means a major re-build. Use 20dB pad (1800r+220r), then gain-up again, twelve times.
The old TASCAM -10dBv standard started when '741-like opamps were all we could get (cheaply enough). There was a '5532 at the end to get the +4dBm master output.
On your existing breadboard, add 220r resistors sum-points to ground so all work at gain of 10.
As you point out, common-mode distortion is a thing and non-inverting is more common in audio. Means a major re-build. Use 20dB pad (1800r+220r), then gain-up again, twelve times.
The old TASCAM -10dBv standard started when '741-like opamps were all we could get (cheaply enough). There was a '5532 at the end to get the +4dBm master output.
Last edited:
Be very interesting to see the measurements for all three - its perfectly possible there's a problem somewhere that's both audible and measurable, perhaps missing decoupling, or a bad relay contacts, filter time-constants poorly matched, or some such. And a difference in gain is another confounding factor - louder always sounds betterHi
I have now a setup that can switch between 3 opamps installed as an output filter of the same DAC (AK4497). The DAC+/DAC- go to the 3 identical low pass filters and the output of the filters (using different opamps) is selected by relays as in a preamp selector using a simple knob.
Test results:
1) my wife detects the differences after about 2 seconds on acoustic music. And she does not know which one is selected (she has no idea what i am switching anyway...).
2) my fiend took about 5 seconds to identify the difference between opamp #1 and #2 but not sure about the 3rd one. It was rock music used this time.
3) of course I can hear the differences too but not on all music and not easily as my wife.
The 3 opamps are VFB types and indicated for audio applications from the datasheet .
I am not claiming anything but just reporting a free A/B test result.
Fab
And you do know that single-blind testing isn't good enough to remove expectation bias. Simple clues like the sound of the knob turning may leak information about the setting, as will a fixed sequence - this has to be rigorously excluded to remove expectation effects, the choice must be random and unknowable.
Alas no, the cheapest NE5532's are SOIC8 surface-mount, and I was buying 100 off, possibly for an opamp-array amp at some point.
There are some tap-points along the chain brought out too. BTW I was running off two 9V batteries.
And I wasn't going to buy 6 Burson's to test them! Though it would be an interesting comparison...
With the NE5532's fed with fast-edged square waves there are very obvious problems with slew-rate limiting that get worse down the chain, I might rig that up for this thread too, as I've recently upgraded to 4-channel 'scope. Moral is audio opamps don't work at logic speeds!
Last edited:
As I said, 3 x same output stage DAC filters (3 x same assembled pcb DAC output section) with exact same gain. Opamp decoupled with same caps and same power supply board as source. Same (new) relays for switching. I myself asked if there was something else than the opamp itself...
I am not saying the difference is night and day just saying it is more obvious in specific passages (soundstage, sibilance , etc) . For my wife testing I have used a music piece she knows very well and I am not sure if she noticed when I switched the knob (at least she did not know which one of the 3 ). Quickly for that piece of music she was able to notice a difference and even was able to determine repeatedly which one was which
With this setup I am able to select which one I prefer upon specific style or piece of music. Maybe my wife, my friend and I are all imagining all this ...
I am interested in your work and also would be interested if you can test other opamps than the NE5532 because TI (Burr-Brown too) and AD have made a lot of so called audio opamps since then with very good specifications and there must be a reason...
If you have the time you can try OPA1612, AD4898-2, AD8620 and OPA1642 all dual opamps.
Fab
I am not saying the difference is night and day just saying it is more obvious in specific passages (soundstage, sibilance , etc) . For my wife testing I have used a music piece she knows very well and I am not sure if she noticed when I switched the knob (at least she did not know which one of the 3 ). Quickly for that piece of music she was able to notice a difference and even was able to determine repeatedly which one was which
With this setup I am able to select which one I prefer upon specific style or piece of music. Maybe my wife, my friend and I are all imagining all this ...
I am interested in your work and also would be interested if you can test other opamps than the NE5532 because TI (Burr-Brown too) and AD have made a lot of so called audio opamps since then with very good specifications and there must be a reason...
If you have the time you can try OPA1612, AD4898-2, AD8620 and OPA1642 all dual opamps.
Fab
Last edited: