Tube Rolling: Does it Make a Difference?

[majingotan]

That's just it.

Why are you presuming that sonic differences that can be identified under blind tests are not heard during normal listening?

(For instance, I hear the differences between my tube and SS preamp during normal listening - the same characteristics I heard under blind listening).

It's actually quite the opposite for most. They hear the magical differences during sighted and it's gone when tested blind. I could agree with that statement regarding lossless and hi-res or even 256 AAC against ALAC 16-bit, but tubes against SS isn't some magical placebo that disappears upon double blind level matched testing to my experience. Completely different amp topology, completely different behavior with headphones / speakers. A 30-50 dB SINAD standard tube amp is pretty easy to distinguish to 90+ dB SINAD SS

 

[beefkabob]

(I presume you think you could reliably identify the voices of two people you know very well, under blind conditions, minutes apart).

If your tube rolling makes Dolly Parton sound like Leonard Cohen, we need to talk about your hifi goals.

 

[MarkS]

Why are you presuming that sonic differences that can be identified under blind tests are not heard during normal listening?

I think it's possible that a rapid-switch test could reveal (for example) small differences in frequency response, but that these would be too small to be noticeable in single-end listening. AFAIK, this is standard audio lore, which is why beefkabob wrote that "Blind tests for sounds that are minutes apart don't work."

I'm agnostic on this. I just know that, for me, if I can't tell which amp is in my system when I listen to it for a while, then, for me, those two amps are identical.

 

[MattHooper]

If your tube rolling makes Dolly Parton sound like Leonard Cohen, we need to talk about your hifi goals.

What if it's the reverse?

 

[MattHooper]

I think it's possible that a rapid-switch test could reveal (for example) small differences in frequency response, but that these would be too small to be noticeable in single-end listening. AFAIK, this is standard audio lore, which is why beefkabob wrote that "Blind tests for sounds that are minutes apart don't work."

I'm agnostic on this. I just know that, for me, if I can't tell which amp is in my system when I listen to it for a while, then, for me, those two amps are identical.

I was mostly confused because you'd made what seemed to be a broad dismissal of the results of blind tests (even "minutes apart"), as if they only identify sonic differences that don't make a difference in normal listening. But of course they can. However, if you mean to say "I'm just talking about differences so minute I don't care about them in normal listening" then that's fine, except it doesn't say anything about blind testing in general.

(After all, Harman Kardon blind tests speakers because the differences heard under those conditions are perceived in regular listening).

And of course it will depend on how significant a difference is to the individual. Maybe you'd detect the differences I did in my blind test. Perhaps not.
And if you did, you might well say "that's not a big enough difference for me to care about." But for me it certainly is, for my everyday listening.

 

[SIY]

I think it's possible that a rapid-switch test could reveal (for example) small differences in frequency response

Also small differences in level.

 

[beefkabob]

What if it's the reverse?

Hallelujah!

 

[MarkS]

Also small differences in level.

There is an easy way to avoid the need to level match, which is to have the volume start at zero at the beginning of each ABX trial, and have the listener adjust it to preference. (This requires that the volume control not give visual or tactile feedback to the listener as to what the level is, but I claim that is not too hard to arrange.)

 

[MarkS]

I was mostly confused because you'd made what seemed to be a broad dismissal of the results of blind tests (even "minutes apart"), as if they only identify sonic differences that don't make a difference in normal listening.

No, that was not at all my intention.

But of course they can. However, if you mean to say "I'm just talking about differences so minute I don't care about them in normal listening" then that's fine, except it doesn't say anything about blind testing in general.

I'm a big fan and proponent of blind testing. It's saved me a great deal of money!

The average at-home audiophile can pretty easily do the kind of blind testing that matters, which is: can I tell which component is intstalled, the new one or the old one? That almost no one ever does this is pretty shocking to me.

 

[beefkabob]

The problems with home blind testing include but are not limited to the lack of a double blind, the lack of a control, lack of level matching, poor test design, and a lack of training. Psychology's power over sound is so absolute that we cannot be sure of what we're hearing. There's plenty of literature on this.

Hence, the need for scientific controls, to seek out the truth. Hence the value of measurements.

Poor level matching alone can make somebody choose crap equipment over something state of the art.

 

[SIY]

There is an easy way to avoid the need to level match, which is to have the volume start at zero at the beginning of each ABX trial, and have the listener adjust it to preference. (This requires that the volume control not give visual or tactile feedback to the listener as to what the level is, but I claim that is not too hard to arrange.)

I'm not sure how you can get validity that way since there's no control over relative levels and the volume knobs would be accessible to the test subject. A master volume control would be much more suitable.

Level-matching is usually very easy, suitable voltmeters are cheap ($10 at Harbor Freight). No idea why someone trying to do testing wouldn't take proper care of that obvious variable.

 

[roderickvd]

What if it's the reverse?

Wrong RPM

 

[MattHooper]

The problems with home blind testing include but are not limited to the lack of a double blind, the lack of a control, lack of level matching, poor test design, and a lack of training. Psychology's power over sound is so absolute that we cannot be sure of what we're hearing. There's plenty of literature on this.

Hence, the need for scientific controls, to seek out the truth. Hence the value of measurements.

Those are broad propositions, but you aren't directly addressing my particular test as far as I can tell (for instance, which of those issues plausibly impact my blind test results).

Further, it's worth pointing out that your level of skepticism of home blind tests seems to be at odds with Amir, who produced a video explaining the worth of doing home blind tests, and that they don't have to be double blind to be informative.

Also note in Amir's video when it comes to matching levels, he suggests using a voltmeter, HOWEVER he also says even AFTER matching levels with a voltmeter "play some music and do the A/B test and make sure they (levels) sound similar to you." He points out depending on the impedance, in his own blind tests he's level matched via volts, but sometimes found via subjective listening that the levels didn't actually match. So, paraphrasing, he says to make sure in addition to matching the levels electronically, to make sure subjectively the sound levels sound the same since that is the test we are going to run.

(In my test, I couldn't tell any level difference after matching with two different decibel meters).

You may differ somewhat from Amir on the worth of home single blind tests, but hey we all got opinions

 

[beefkabob]

Those are broad propositions, but you aren't directly addressing my particular test as far as I can tell (for instance, which of those issues plausibly impact my blind test results).

Further, it's worth pointing out that your level of skepticism of home blind tests seems to be at odds with Amir, who produced a video explaining the worth of doing home blind tests, and that they don't have to be double blind to be informative.

Also note in Amir's video when it comes to matching levels, he suggests using a voltmeter, HOWEVER he also says even AFTER matching levels with a voltmeter "play some music and do the A/B test and make sure they (levels) sound similar to you." He points out depending on the impedance, in his own blind tests he's level matched via volts, but sometimes found via subjective listening that the levels didn't actually match. So, paraphrasing, he says to make sure in addition to matching the levels electronically, to make sure subjectively the sound levels sound the same since that is the test we are going to run.

(In my test, I couldn't tell any level difference after matching with two different decibel meters).

You may want to dismiss Amir's view, but hey we all got opinions

I don't know your full setup, and there are still ways for things to go completely wrong, but I applaud your efforts. Also, for those reasons, I push for measurements. I'm not saying that tubes cannot sound different from each other. I'm saying that a home test without measurements isn't going to convince me of anything.

 

[MattHooper]

I don't know your full setup, and there are still ways for things to go completely wrong, but I applaud your efforts. Also, for those reasons, I push for measurements. I'm not saying that tubes cannot sound different from each other. I'm saying that a home test without measurements isn't going to convince me of anything.

I completely understand that and I especially find your last sentence completely reasonable! **

This is why I didn't present the test as establishing anything with some scientific certainty. It's also why I've always felt that doing blind testing of equipment at home is ultimately for the individual's use, at best.


It reminds me of one of the first blind tests I did, in the 90's. (Which I've mentioned before here). There was lots of knocking of heads between the "objectivists/subjectivists" on the usenet forums, with the "objectivists" being knowledgeable in electronics. The theme on CD players/DACs was that any competently designed player should sound identical to the next. That made total sense to me, yet I mentioned that I had 2 different CD players and a DAC and they seemed to sound distinctly different to me. Knowing about sighted bias I had my father in law (an engineer) help me do a blind test shoot-out.

I easily identified between the players/DAC and posted the results. Some people accepted them, but some more skeptical engineers rightly pushed back and said "here's how you can tighten up your test." (I included a voltmeter in the next test). So I repeated the tests with their suggestions, and once again easily discerned between the CD players/DAC and posted the results. Again discussion ensued, some accepting the results, but some (e.g. Arny Kruger) absolutely dug in their heels and tried to suggest ways the test may have been faulty (none of which were actually plausible, given what I knew of how we performed the tests).

But I completely understood the position of those remaining skeptical. All they had was someone's claims on a usenet forum. They weren't there for the test themselves, so how do they know there wasn't fibbing somewhere, or exaggeration, or an imperfect description of what actually happened during the tests, or that there wasn't some obvious problem that I wasn't aware of that they would have spotted. Not being able to confirm any of this for themselves, if they hold the claim to be dubious to begin with, I understand remaining skeptical. I've seen some subjectivist-type audiophiles claim they did blind testing at home hearing differences between AC cables. Frankly...their say-so isn't enough for me, given what I think I understand about the subject.

When the average audiophile does blind tests at home, it seems the best you can do is look to critiques of the test, do your best to infer how likely the test suffered a problem, and make your tentative personal conclusions. You aren't going to automatically convince skeptics who weren't there.

I know some others on this site have said they've done their own blind tests on gear in the past, though we don't have their methodology/results to interrogate. But I understand their experience is part of what formed their views on certain subjects, just as mine have influenced my own.

** (btw, in my blind test thread there was a link to stereophile's measurements of the CJ preamp, and stereophile also measured the Benchmark preamp).

 

[MattHooper]

Not related to tubes, but somewhat apropos of my post above about attempting blind tests, here's a link to another thread where I just posted a link to a previous blind test I attempted with the AVSForum members:

Master Thread: Are measurements Everything or Nothing?

I use my system for films, music (CD, vynil...), videogames and some TV. Neutral and transparent is the most convenient I can do due to the variety of sources I play. My system has a neutral orientation as well,it's the safest way to go. But I do have plans for a little set-horn system in the...

www.audiosciencereview.com

 

[srkbear]

Brian May's signature tone was not from a tube amp. It was from a 1.5 watt Germanium transistor amp that was retrieved from a dumpster. Read about its origins here https://guitar.com/features/gallery/the-story-of-brian-mays-deacy-amp/
Clapton reportedly used a 5 watt Fender Champ - 6V6 output tube - for Layla.
Weber Speakers made a tube amp kit called "smokin Joe" in honor of Joe Walsh IIRC. 15 watt EL84 push-pull amp - https://www.tedweber.com/sjii-c-kt/

Generalizations are rarely accurate.

Actually May’s primary amp has always been a Vox AC30, which is indeed a tube amplifier. The amp you’re referring to is the “Deacy” transistor-based amplifier designed by Queen’s bassist, John Deacon, who had formidable skills as a technical engineer himself—and May used it primarily to simulate odd sounds reminiscent of string instruments on songs such as “Lazy on a Sunday Afternoon”.

But it was not responsible for May’s “tone”, which was largely created by the unique constellation of both in-phase and out-of-phase Burns Tri-Sonic single coil pickups, wired in series, coated with epoxide adhesive to cool off the sharpness of the single coil wiring. The somewhat nasal sound is a result of the neck and middle pickups being out of phase—his Red Special was tricked out with an array of switches allowing him to attain a vast array of tones. His sharp attack was aided by his use of a sixpence coin instead of a typical plectrum.

All this was augmented by his use of his choice of British tube amplifiers. If you think the soaring roar he created on his iconic solo in “We Will Rock You” was the result of a tiny transistor amp powered by a 9 volt battery, you’re kidding yourself—that was all Vox AC30 with Tung-Sol EL84s.

I have actually built several Fender Tweed ‘57 Champ amplifiers myself, with a variety of tube configurations. Clapton did indeed use this amp to record the main riff in “Layla”, turned up to full volume and enclosed in a wooden enclosure with close mic placement—and the stock tubes were indeed Tung Sol 6V6s. Walsh has used any number of amps during his career and rarely were they solid state.

What you meant by your closing “generalizations are rarely accurate” statement remains a mystery to me, but if it was in the service of sanctimony and officiousness, I’m afraid that your satisfaction is in no way a result of me feeling intimidated. Cheers!

 

[srkbear]

Slow clap.

You have figured out my evil plan.

Distortion does make guitars sound nice. Settings and equipment can definitely lead to a certain sound. However, those certain sounds are likely easier and cheaper to create with software and a bit of knowhow, as opposed to being a gear junky chasing the next magical tube.

I have no idea whatsoever what your point is. Slow clap? My post was a performance? I meant it sincerely, please sit down. What’s up with the gratuitous one-upmanship that is pervading this site of late?

 

[srkbear]

I see no parallel between guitar amps and hifi amps except they draw power and often make noise.

A guitar amp is intended to alter sound coming from a pickup (which also alters sound) to a speaker. The two become a musical instrument, making artistry and virtuosity into various kinds of music (or noise). Screwing around with tubes in guitar amp sockets will alter the 'thing' that the guitar amp does. Kinda like adding a guitar fuzz pedal or such.

A hifi amp is intended to make small signals from its input into large and powerful signals that are closely representative of the input. No artistry is intended in that amp, and the virtuosity is left to the source (and too often the speakers).

Messing around with the tubes in a hifi amp messes around with all sorts of parameters that likely were never intended. Screwing with tubes will not make your Harry Phibeetz into Jimi Hendricks. It will make Harry into a possible train wreck that appeals to you. If your hifi tube amp is any good at all, its designer took pains to place each tube into its optimum operating circumstance for accurate reproduction. Changing tube types and manufacturers simply rolls the dice on operating points. Is it a worthwhile goal to make Jimi's guitar sound more like a banjo? A harpsichord?

Are you inclined to buy a solid state amp, learn soldering, and change transistors to tickle forth your wet dreams? It'll happen, but it requires some repair bench skills.

If you think a tube is bad, replace it with the same type and reset the bias if you're able. Tube rolling won't make you a composer. Or a musician. Or even a producer. It will make you kind of cranky.

The point was that in guitar amps tubes are driven well beyond clipping to produce distortion, which is indeed musical and the tubes are as instrumental to the performance as the pickups are on a guitar. Tube applications in guitar playing are only analogous to their use in Hifi applications in that they both provide amplification—but their implementations are entirely distinct. Clipped transistors are not music in any sense to most musicians. Are you a guitarist?

 

[roderickvd]

I think that “generalizations are rarely accurate” is a wonderful and deeply philosophical oxymoron.