MEASURING ULTRA LOW-LEVEL MUSICAL AMPLIFIER SIGNALS
This may be the most important letter I have ever received from a reader (which is why it is located here and not below). The letter speaks for itself, and is unedited, with my bold. (My "personal comments", plus a sceptical and cautionary perspective from another source, will follow below).
"My name is Steve Keiser, the 'K' of B&K components, and presently design engineer with Luminance Audio. I have developed a measurement system which is able to quantify distortion levels at micropower quantities down to 1/1000th of a watt. These measurement techniques are unprecedented, and have revealed a number of revelations of amplifier distortion characteristics, at micropower levels, which are in direct opposition with traditional and scientific assumptions up unto this point.
The emphasis of my work is to definitively quantify low level signal linearity measurements of power amplifiers, and attempt to correlate these measurements with subjective listening results, as well as establishing the significance of low level distortion. Conventional test equipment generally does not resolve meaningful distortion measurements below 100mw, since the measurements become predominated by noise.
I have modified a spectrum analysis software program, which uses time-averaging to effectively cancel out noise products, leaving an identifiable signal and its related harmonics. This time-averaging approach is to identify extremely weak signals from spacecraft, amid a very high noise ambient background. Using this method, I am able to resolve a standardized total harmonic distortion measurement down to 1/1000 watts, and an approximate measurement down to 1/500,000th of a watt.
My measurement results oppose common engineering supposition, in that it is commonly believed that very low signal linearity is essentially 'virtually perfect', and that only high level signal linearity is a relevant parameter. My measurements indicate exactly the opposite is true of this common held assumption, particularly for amplifiers employing solid state devices.
To give you an example: the Halcro DM58 amplifier measures .007%THD at 2 watts, whereas at 1/1000th watts, THD measures 8.9%! By contrast a Wavac SH833 measures .57%THD at 2 watts and 1.6%THD at 1/1000th watts. The tube Wavac exhibits significantly lower THD at low signal levels by orders of magnitude than the Halcro. I have measured numerous amplifiers, both solid state and tube, which I will provide to you as well as any other information you may want pertaining to this work.
Correlative Listening Tests
...I will now provide a supplementary addition regarding correlative listening tests with a panel of 5 evaluators. Some of the tests were conducted using a blind A/B comparison method, in order to satisfy militant objectivists. The two amplifiers compared were a Wavac SH833 and Halcro DM58. In 10 trials, with listeners blindfolded, every listener on the panel preferred the Wavac by several orders of magnitude, with commentary such as describing the Halcro as sounding: transistory, thin, harsh, dark, closed in spatially, as well as having poor sound floor resolution.
Every listener described the Halcro as being 'unlistenable', while the Wavac enjoyed universal positive accolades. These listening tests correlate exactly with the comparative measurements I outlined.
My research into this characteristic is currently ongoing, and I would enjoy sharing my results with audio enthusiasts, editors, and designers. If this correlation between measurement technique and listening impressions holds up consistently, it could mean a whole new approach to audio engineering could be opened up resulting in significant breakthroughs in design performance. The main idea is to let our ears continue to be the final arbitrator of component performance and allow objective science to enhance our subjective appreciation."
The results of these measurements, and listening tests, are not surprising to me. They confirm what I, and many thousands of other audiophiles and music lovers, have heard since the 1960's. I just wish someone had discovered Mr. Keiser's breakthrough method of measurements a long time ago. From my perspective, Steve Keiser may eventually deserve some sort of "Audio Nobel Prize" for his work, at least after it is verified by other objective audio scientists and technicians.
This new measurement procedure may also be a vindication for me personally. Back in 1999, on this website, I wrote a short article about "The Problems With Measurements", which is concerned with a closely related issue. It is still posted in My Audio Philosophy.
Further, these new measurements, as important as they may turn out to be, still don't address what musical information is actually "lost" at low signal levels. The added distortion appears to explain the Halcro's (anecdotal) sonic problems of "harshness", but I believe the "leanness" is caused by a loss of harmonics, which are still unmeasurable, as far as I know.
A Cautionary Viewpoint (From a Surprising Source)
To gain another perspective, I asked Israel Blume, of Coincident Speaker Technology, to respond to the posted results of Steve Keiser's THD measurement experiments. As a manufacturer of tube amplifiers, and a SET model in particular, Blume's response below was somewhat of a surprise to me. Here it is, with some minor editing and my bold (with my response below):
"It is common knowledge that currently applied amplifier measurement techniques are not sufficiently refined to correlate to what we hear. The revelation here is that THD measurements are now possible at the micro power level (as low as 1/1000 of a watt), which the auteur of this technique claims has not been previously possible.
The question now is: Will this new measurement scheme provide greater insight into the subjective sound of the unit being tested? If previous THD testing is any indication, the answer would be "no". This category of measurement has been clearly shown to bear no correlation to the sound quality of an amplifier (with the exception of ridiculously excessive amounts of measured THD).
In analyzing the utility of the micro power testing, a significant factor to be considered is the power rating of the unit tested and its design goals. All things being equal, an amplifier designed to perform optimally at low power levels, should be more linear and exhibit lower levels of distortion than one designed for high power applications*, with regard to its measurements at the micro power level. The quality of sound produced by the lower powered amp might therefore be superior, if it was not pushed to beyond its available power limits.
The sensitivity of the speaker used, and the ease with which the speaker can be driven, would be of enormous significance in comparing the sound quality of these 2 amplifiers. In another system setting, where a difficult to drive speaker is being employed, the reverse subjective results might likely ensue due to the power demands placed upon the low powered amplifier, notwithstanding the large power amplifier's higher levels of THD at the micro power level.
For the new THD measurements to have validity, a large sampling of every type of amplifier topology and power rating will need to be tested. The subjective analysis will require a speaker system easy enough to drive, that the power delivery of the amplifiers will not be a variable under test. Only the quality of sound should matter. In an extreme example, a SET amplifier, even of the lowest distortion measurable, will not sound right on a 82 db sensitive speaker with an impedance that dips to 1 ohm, while a high current, large power amplifier with much higher levels of micro power THD, will sound superior on the same speaker.
I would like nothing more than to have a set of measurements available that could accurately predict the sound of an audio component. As a designer, life would become infinitely easier. I further would be thrilled to have said measurements verify what I firmly believe- No amplifiers in existence can compare sonically to the best SET amplifiers when used on easy-to-drive speakers. For now, much more testing will be required before those assertions can be made."
*Actually, Michael Fremer, and some other mainstream proponents of "1,000 watt amplifiers", have claimed they have absolutely no sonic downsides at low listening levels. (Example- See Fremer's review of the Musical Fidelity Titan amplifiers, in the June 2009 issue of Stereophile, which is still posted on-line on their website.)
Personal Notes- This contrarian response from Blume is highly ironic, since he should be extremely thrilled at these results. However, I believe these results, even if still preliminary, are more important than Blume describes. The discovery that an amplifier so highly touted, the Halcro, has much higher measurable distortion than their own specs and reputation, at volume levels which are definitely audible, is an important breakthrough in my opinion.
Just as important to me, a completely different technology (Wavac/Tube/SET), considered "obsolete" by so many "audio experts", ends up having much better measurements in this very same area. The blind listening tests, while not surprising to me, are also important, as well as the actual sonic descriptions of the two models.
In the final analysis, I agree with Blume that many more experiments are in order, which I will post as they are received, no matter what they are. This may be only the beginning, but the amplifiers chosen, and the results so far, couldn't be more edifying. Let's hope that future results are equally surprising and revelatory.