Very interesting paper on Tube vs. Solid State perceptual differences
May 6, 2005 at 6:04 PM Thread Starter Post #1 of 18

Glassman

Headphoneus Supremus
Joined
Mar 22, 2003
Posts
1,830
Likes
11
really interesting reading, you won't regret the time spent reading it I promise
wink.gif


http://milbert.com/tstxt.htm

also look here, various devices tested in similar configuration with distortion spectra plotted.. especially notice the absence of >3rd harmonic in case of simple triode and the noise floor is nice too..

http://www.spectrum.ieee.org/select/0898/tubesb3.html
 
May 6, 2005 at 6:56 PM Post #2 of 18
Maybe I missed something, but the article only seemed to talk about the sounds the components made when they were over worked. What would account for the difference in sound if both a tube and solid-state amp were working within completely normal limits?
 
May 6, 2005 at 7:59 PM Post #3 of 18
Quote:

Originally Posted by jpr703
Maybe I missed something, but the article only seemed to talk about the sounds the components made when they were over worked. What would account for the difference in sound if both a tube and solid-state amp were working within completely normal limits?


psychology.
 
May 7, 2005 at 12:11 AM Post #4 of 18
Differences are easily summed by what was said when I walked into my favorate hifi shop looking for some 6922 for my next project:

"Do you stock any vacuum tubes?"
"Do we look like we belong to the flat earth society?"
 
May 7, 2005 at 12:13 AM Post #5 of 18
Quote:

Originally Posted by jpr703
Maybe I missed something, but the article only seemed to talk about the sounds the components made when they were over worked. What would account for the difference in sound if both a tube and solid-state amp were working within completely normal limits?


Primarily harmonic distortion.
 
Jan 14, 2008 at 12:22 AM Post #6 of 18
I can't resist. I did read the paper and I find it VERY interesting. Actually I should have been asleep and it's the paper's sake I'm not :p

To sum up, the paper says: Tubes have strong even order harmonics, in solid state systems those are much weaker (opamps don't almost exhibit any). However solid state systems have strong third harmonics. As the paper states, even order harmonics (an octave higher than fundamental tone) sound musical to us and give the tone "body" - odd order harmonics rather remove it and contribute to a dull sound.
The point of this is that tube amps distort more but that distortion sounds pleasant to the ear. Solid state amps are missing even order harmonics (compared to tube designs) thus missing the... harmonic sounding distortion that would compensate disharmonic taste of third (and 5th and 7th) order harmonics.
The paper states one nice thing - opamp's distortion is so obviously ugly sounding that only 5dB of it are detectable by listeners whereas transistors have to distort 10dB over the fundamental signal to be detectable and tubes may distort even up to 20dB without being noticed by untrained listeners.
That means you can load a tube amp 15 more than equivalent (in power) opamp using solid state one (but I've got a question... aren't solid state amps capable of twice more power handling than tube amps at the same price range?

Microphone preamps ARE often driven into clipping - that's why the study isn't pointless.

However for music reproduction the matter of clipping characteristicsis less important as most of us don't overload our amps.
The distortion present in recording audible as shrill or closed sound character can't be removed by the means of adding more distortion (hovewer a bad recording CAN overdrive your amp - that's where amp design becomes more important - Portishead 's song Dummy is a good example of a bad recording. I hate it :p)

That's why I think the paper is especially relevant to studio engineers but not to listeners.
 
Jan 14, 2008 at 9:27 PM Post #7 of 18
wow, what you posted, d.phens, explains clinically well why I dislike digital sound but would never know how to describe it except to say that it sounds 'flat' and cold to me. It's great that there is a science behind something like this.
 
Jan 14, 2008 at 11:41 PM Post #9 of 18
Quote:

Originally Posted by d.phens /img/forum/go_quote.gif
To sum up, the paper says: Tubes have strong even order harmonics, in solid state systems those are much weaker (opamps don't almost exhibit any). However solid state systems have strong third harmonics. As the paper states, even order harmonics (an octave higher than fundamental tone) sound musical to us and give the tone "body" - odd order harmonics rather remove it and contribute to a dull sound.
The point of this is that tube amps distort more but that distortion sounds pleasant to the ear.



It's a nice theory, but it is also demonstrably wrong in two easy ways. 1. Build a push pull tube amp. This will cancel this even harmonic distortion, yet it will still sound like a tube amp. 2. Use your computer to inject even harmonic distortion into your signal and listen through your opamps. It still won't sound like tubes.

If, indeed, harmonic distortion is the issue (and it may not be, word on the street is that Hirata distortion may be more important -- my technical knowledge is way too poor to have an opinion) maybe it is only the high odds that matter with the evens being unimportant? That is to say, maybe even order distortion is inconsequential, while odd isn't. So, it is the solid state stuff that is colored, and the way to make it uncolored, like tubes are naturally, is to try to rid them of this odd order distortion. You'll periodically her people talk about certain tube circuits sounding solid statey, and these are inevitably ones that use tubes with high odd harmonic distortion -- like the 6922.

Quote:

Originally Posted by regal /img/forum/go_quote.gif
IMO where a tube excels is in playing a recording that was recorded with loads of tube distortion (60-70's electric guitar.)


Which suggests to me that tubes are more linear, and thus better able to reproduce the rich harmonic content of these recordings. The similarities between a tube guitar amp, and a tube hifi amp, are pretty minimal.

Fwiw, the "color" that people talk about with tube amps generally has less to do with the tubes, and more to do with the electrolytic caps used. Build a tube amp without electrolytics in the signal path and it will not only sound better than a solid state amp, but it will also not have that mushy warm sound that gave these amps a bad name in the first place.
 
Jan 15, 2008 at 3:05 AM Post #10 of 18
Thanks glassman (belated) for an excellent link and read. The question just won't go away.
wink.gif

In the later 70's TIM (transient intermodulation distortion) differences between tube and solid state were investigated in the interest of quantifying what was heard but never explained. That was followed later by Hirata's investigations, mentioned in dsavitsk's post, and that is finally being resurrected again with good cause.

Those inclined may find this piece of interest as an minor adjunct to the excellent Milbert piece linked by glassman......
An 'priceless' excerpt on conventional testing from the link:
"Unfortunately, square-waves are not actually much more like music than is a simple sinewave. As a result, although such tests are useful for some purposes, they also do not seem to be a useful guide to musical performance in many cases."
Amplifier distortion masurements

For those still interested, further insight may be gained by seeking out more info on Hirata distortion as mentioned by dsavitsk Also in his fwiw, he brought up a very valid point about caps and "tube sound" which leads some to (falsely) describe some SS amps as sounding 'Tube Like'.

Personally, I would have rather tubes had stayed in the obscure minority because the new 'interest' in tubes has really screwed up prices.
 
Jan 15, 2008 at 7:22 AM Post #11 of 18
I think the circuit topology has as much to do with the differences as the devices themselves. High voltage, high impedance, transformer coupling, zero feedback. Etc...

(some) Tubes have low enough distortion that they can work well without feedback. Things like NOS 45's are probably among the lowest distortion devices ever made. The big PP DHT tube amps achieve extremely low distortion considering the total lack of feedback.

Look at things like the Pass Zen, which sounds WAY more like a tube amp than a normal SS amp & 2nd harmonic injection will. I know Gary Pimm has built a high-voltage PP transformer-coupled solid state amp that sounds a *lot* like the PP tube amps (maybe better?).
 
Mar 10, 2008 at 5:54 AM Post #12 of 18
Quote:

Originally Posted by dsavitsk /img/forum/go_quote.gif
word on the street is that Hirata distortion may be more important


Reference to an explanation of said distortion?
 
Mar 10, 2008 at 10:28 AM Post #13 of 18
Quote:

Originally Posted by d.phens /img/forum/go_quote.gif
However solid state systems have strong third harmonics. As the paper states, even order harmonics (an octave higher than fundamental tone) sound musical to us and give the tone "body" - odd order harmonics rather remove it and contribute to a dull sound.


Strong? Take for instance AD8620. It has a THD of 0.0006%. It's a bit strong to call that strong. It still sounds distorted. There must be another explanation.
 
Mar 10, 2008 at 11:00 AM Post #14 of 18
Third harmonics are not that audible. Odd harmonics really become a problem with 5th and above.
John Curl (of Parasound, and one of the most well known audio designers) will tell you the same thing. If third harmonics were a real problem, cassette tapes would have been absolutely unlistenable (that harmonic there can reach 15%). Basically, second through fourth harmonics shouldn't have much attention paid to them.
 
Mar 11, 2008 at 4:20 PM Post #15 of 18
Crowbar,

I had not thought about Joseph Conrad for a while, but that quote from Orwell hooked me. I agree, but wonder if it will still be true in, say, 20 years when women doing any occupation will be routine. Will that change the way they-and men-think about what is significant in life?

F
 

Users who are viewing this thread

Back
Top