^^This.
I keep telling people you really don't have to know anything technical. All you need to know is there is a thriving vacuum-tube industry with more US tube equipment manufacturers now than there were in 1958. If the transistors and high feedback really did sound better, that industry would not exist. So its apparent that something is being missed. My pet theory is that we design for good numbers that don't happen to have a lot to do with how we perceive sound.
Quite a few do. Certainly mine, Kevin Kennedy's, Frank deGrove's, Pete Millett's, Tubelab's... Some people like to design hot glowing effects boxes, and if that's their goal, great. Ditto the solid state forum, there's a mix of approaches, not just transparent amps.
Nelson is profoundly correct- it's all about entertainment.
Not to be contrary, I should have been more accurate - talking about power amps here. We routinely discuss PPM distortion in SS, not even close in tubes at the same power level. Similar situation with bandwidth. I DO follow that forum, in fact the last three years most of the design work I have been doing is on hybrid amps.
I do agree that it is about entertainment. And not only the entertainment of the material being reproduced, but of tinkering with ideas, and - on a more consumer oriented, yet more problematic level - audiophiles like to play. There are miriad combinations of sources, amps, speakers, cables, voodoo to try. Here's proof: try to market an active system that isn't made up of separate components. It works in HT where people are fed up with many boxes and tons of cable, but not in high-end and no matter how good it is.
what I take from lossy perceptual codec theory is that there's lots we don't hear
critical band theory, frequency masking is fundamental to lossy perceptual codecs ability to use less than 20% of the Shannon channel capacity and be scored as audibly transparent
typically everything within one critical band is represented with 6-7 bits - leading to the idea that things like distortion products with frequency component amplitudes less than ~1% of other music signal in the same critical band aren't heard
with test tones with large frequency separation though it is apparently possible to hear something as low as 80 dB down from another component far away in in our hearing critical band scale
its not surprising then that technical errors may be inaudible in many music reproduction cases
while at the same time if we don't appeal to complex Psychoacoustics and only use perceptual noise floor as a engineering goal to keep all distortion below the technical performance constraints look a bit more extreme
critical band theory, frequency masking is fundamental to lossy perceptual codecs ability to use less than 20% of the Shannon channel capacity and be scored as audibly transparent
typically everything within one critical band is represented with 6-7 bits - leading to the idea that things like distortion products with frequency component amplitudes less than ~1% of other music signal in the same critical band aren't heard
with test tones with large frequency separation though it is apparently possible to hear something as low as 80 dB down from another component far away in in our hearing critical band scale
its not surprising then that technical errors may be inaudible in many music reproduction cases
while at the same time if we don't appeal to complex Psychoacoustics and only use perceptual noise floor as a engineering goal to keep all distortion below the technical performance constraints look a bit more extreme
Last edited:
Measurable differences, sure. Audible? Probably not, other than the usual boring stuff like frequency response. That's why Clark was willing to put up money in his test, which does not exclude tube amps.
I would tend to agree with that. which is a shame. Meridian have had some success at selling active over the last 25 years, but they were into lifestyle systems pretty much at the beginning. And of course Sonos sell millions of active boxes. Likewise the desktop audio market is doing well with active.
So some light at the end of the tunnel, but not much.
And yes I am going active...
No crossover distortion at all ? , that will be very good point&advantage for start to design HQ-PP-SS amp ! ,only if those rare SIT`s will be easy&cheap available now , but reality is so different ....
BTW , I know for one GNFB free HQ-OTL-PP tube power amp design where output power triodes never cut off to .🙂
You may want to narrow your claim down to audio and stage amps as there were a lot of industrial tube products made by a lot of companies back in 58, as well as TVs.
You can buy land cameras again. Doesn't mean they are good, in fact the non pro polaroids were carp even in the day. In fact my film camera is worth more now than it was 10 years ago.
Fixed wheel bikes are a daft notion, especially in town, but most cities are full of hipsters trying to shorten their lives.
So I don't see the correlation, but I am very pleased that there is choice and that enough people buy things to keep some of these fred in a shed (and larger) operations going. Stops the world getting boring and homogenous with everyone having an iphone and a sonus and a spotify account.
I don't follow this argument either. There are more class d amps on the market now then there were 10 years ago and probably more class d amps purchased in the last few years than tube amps. By your logic, that must mean they are "superior"...
Your problem is you believe it is a zero sum game. You think one topology is "superior" and "sounds better" in an absolute sense, for everyone. You seem to forget there is such a thing as personal preference. I will use your argument to support my belief that people's perceptions are not all the same: if they were, we wouldn't have successful ss, tube, and class d companies- everyone would own tubes. We wouldn't have all the wide variety of speaker types, or cable wars, or vinyl vs digital.
So you have found your holy grail. Congrats. Mine happens to different and every bit as valid.
Not absence of distortion, as the output is not linear, but there is nothing like a easily discernible crossover region - it's de facto spread from full + to - output voltage. Result is less objectionable HD and IMD, easier to correct via NFB - or if desired, not corrected at all (see Nelson's SIT amps).
Yes crossover distortion is a widely-encompassing term. The prototypical one is where complementary output devices turn off near zero output, leaving a region where the system has essentially no gain. But the gm-doubling (or gm-halving) stuff can be much smoother and less annoying/easier to correct. And there are ways to mitigate it even without global feedback.
That's exactly my point from the beginning. One could say the peculiarities/features of each technology produce circuits that measure quite differently for the same perceived accuracy. Which tells us there are different ways to get to that point. Eg, one would hardly insist on -100dB THD products on a tube amp (not that it's strictly impossible) which would in fact be still quite sensible on an SS amp.
There's a lot of gamesmanship surrounding getting the numbers low. And that's fine, it's interesting from an intellectual standpoint. It doesn't change the sound, but we all have various reasons for doing what we do. I'm interested in getting -100dB out of a tube power amp just for the sheer challenge.
That is, if you assume a resistive load. Add reactance and it shifts away from zero. Inductive loads can be special fun, even if it's only a capacitive load isolation cell (parallel RL in series with the output) 😛.
BTW here's an interesting thing to try, along the lines of being aware that amps work on a linear scale while the ear works on a log scale, and that average listening levels fall below 1W and often into the mW region - adjust output offset voltage and measure/listen to the amp at low levels.
