I just had a little tussle with with the local blatherskites over at one of the less hospitable internet audio forums. The actual subject of the exchange is unimportant. However it got me to thinking about distortion and how today the concept of distortion has become seriously misunderstood.
One of the most unfortunate consequences of the “Stereo Wars” of the 1970s was the amount of misinformation and propaganda spread concerning audio quality. And on no subject was this more pronounced than that concerning distortion. We went from the 1950s when we had truly great sounding amps with nominally less then 5% harmonic distortion to the mid 1970s when we had crappy sounding amps with far less than 1% (or even 0.1%) harmonic distortion. But the marketers and salesmen of the time convinced buyers that the overall Total Harmonic Distortion or THD number was some kind of metric by which all amplifiers should be judged..
Some make the argument that it was very high gain amplifiers with massive feedback that drove the reduction in distortion numbers; which is technically true. But Harold Black invented the feedback amplifier in 1927 and published his seminal paper “Stabilized Feedback Amplifiers” in 1934. By the early 1940s we had lots of high gain signal chains with lots of feedback, they just weren’t used for audio amplifiers. But why?
The answer to this question lies in the understanding of distortion. Those designers of the 1940s and 1950s knew that not all distortion was created equal. They knew that odd order harmonic distortion (3rd, 6th 7th, etc.) was the enemy of a musical amplifier and that the even order harmonic distortion (2nd, 4th, 6th, etc.) was the key to enjoyable listening. They understood the action of the human ear and how we perceive sound. They knew that a few percent of even order harmonic distortion (of decreasing magnitude with order) was not only acceptable, but desirable. And this is exactly the type of harmonic structure produced by triodes.
But this is not the type of distortion produced by transistor amps. These amplifiers are rich in odd order harmonics that sound harsh and tiring. So to get transistor amplifiers with even acceptable sound quality, they had to drive the distortion way down. Down, in fact, to far less than 1% THD. And those unscrupulous marketers, salesmen, and designers of the 1970s made THD their whole marketing pitch.
But we know better.
However, due to the decades of misunderstanding and misinformation concerning amplifier performance, the word distortion carries with it a very negative connotation. As such, I have a proposition to make. When addressing the performance of single ended vacuum tube audio circuits, I think we should abandon the entire notion of Total Harmonic Distortion. Instead I think we should talk about two different metrics for our amplifiers.
The first we shall identify as “Color”. The amplifier “Color” should be calculated as the total even order harmonic distortion and reported in percentage (%). The second we shall identify as Total Odd Order Distortion or TOOD. This should be calculated as the total odd order harmonic distortion and reported in percentage (%) as well. What this does is allow us to separate the two different auditory phenomenon associated with harmonic distortion. The TOOD is what we would call “bad” distortion; that which should be minimized. The “Color” is what we would call beneficial distortion; that which we should target just enough to give the amplifier a pleasing richness and fullness. If we base our numbers on the great amplifiers of the 1950s and 1960s maybe we should strive for less than 1% TOOD and maybe around 4% Color.
As an example, let’s look at the 6V6 Marblewood amp. Subjectively, this amplifier sounds much better than one would think based on the THD numbers. It’s been described as “rich”, “full”, and “musical” by those I’ve invited to listen. On the Marblewood amplifier page I posted the distortion characteristics for the amp using the 6V6 power tube. Here is that same table but with the added calculations of “Color” and “TOOD”.
This shows us something very interesting. When the distortion is broken down into “Color” and “TOOD” we see that the numbers say that the discordant distortion is very low and the color is moderate. Does this provide us some more clues as to why single ended tube amplifiers sound so good? Perhaps the “Color” is providing the fullness and richness; keeping the amplifier from sounding sterile or strident. While at the same time, the discordant “TOOD” numbers are far less than 1%.
So what do people think? Am I on to something here?
As always, questions and comments are welcome.
Pingback: The Morning SET Swap | Cascade Tubes
Pingback: Talking about SETs | Cascade Tubes
I love the THD assessment, and I agree 100%. I’ve also argued to simply use our ears not our eyes when considering what audio gear to buy/use. I only recently in the past five years have discovered tubes let alone SET amps. They’ve changed my life and rekindled forever my love of music. Every time I go back to one of my SS amps, be that vintage cap coupled single rail power or modern dual rail, I always go right back to my SET ‘flea powered’ amps. Personally, I think the tube revitalizing movement is awesome and hope it doesn’t loose any steam. Keep up the awesome work Matt!!
You could be on to something worthy of putting to the test objectively and subjectively. If proved and became part of defining amplifier performance, reviews might become more informative. Would transistor based amplifiers operating in class A and single ended have similar colour profile to a triode based amplifier in the same configuration? I’m thinking of the MCA and ZCA for example.
In answer to your question “Would transistor based amplifiers operating in class A and single ended have similar colour profile to a triode based amplifier in the same configuration?” I have to say probably not. Fets and BJTs are generally run in regions where the curves are much more “pentode” like than triode like. Whereas the triode is rich in second harmonic, the SS devices tend to promote a full set of harmonics (even and odd). Not that transistor amps can’t sound good, I’m just saying that they probably won’t sound like a single ended triode amp.
Matt, I will be more specific. If a sawtooth waveform is amplified by a tube amp and by a transistor amp would the output be more round and less sharp leaving the tube amp than the transistor amp? If so, outside of harmonics will this produce a different sound? I know it is a little off topic, but it may account for another difference in the way the two amp topologies sound.
Matt, interesting article. I like your approach to separating even and odd harmonic distortion and defining them as Color and TODD. I do feel that something could be added related to the type of music being amplified. I am curious about slew rate. Does the different slew rates between tubes and transistors affect the overall Color? Any insight will be greatly appreciated.
I did notice that the figures for Color and TOOD don’t add up to THD. I this because you’ve included the 6th harmonic?
That’s because the Total Harmonic Distortion (THD) is the RSS of the individual harmonics, not the sum. I calculated “Color” in the same manner (RSS).
For me, this article does bring perfect clarity to a subject I used to find a little confusing.
As to bringing it in as an industry standard? It would be so sensible, but clear, easy to understand information would be a burden on the salespersons’ feelings of superiority. We’re not supposed to understand certain things. Too much information would make him unnecessary. Many years ago, a friend and I were looking for a demonstration of different turntables in a well known shop in London. The salespersons reply, when we asked was… “What’s there to hear on a turntable?”