Amplifier reproduction of instrument Timbre
Drew -- I think that all sounds right, but when expressed to that extreme is probably applicable mainly to mid-fi and lo-fi mass market products. I think that at this point most high-end designers are aware of these issues, and it's more a matter of the degree to which the particular design can minimize the use of feedback without unreasonably degrading the parameters that are helped by feedback. Regards, -- Al |
I couldn't find the comments about transient issues in DAC design, however, I did find a lot of information on TIM. I thought this comment was particularly interesting given the constant debate about measurable characteristics of amplifiers, I will leave the company unnamed, "We have developed a high quality, precision audio hi-fi amplifier designed with a completely different set of criteria from conventional commercial hi-fi amps. As most naive hi-fi buyers compare harmonic distortion specs between amplifiers and influence their choice on little else, manufacturers compete by using large amounts of negative feedback to reduce their total harmonic distortion (THD) to absurdly low levels, typically less than 0,05%, sacrificing transient response. This is actually ridiculous, as the ear finds it impossible to hear harmonic distortion below about 5%, and in fact creates significant (and measurable) internal harmonic distortion of its own. Even excellent loudspeakers produce 1 to 5% THD at low frequencies! The result of all this excessive negative feedback is that when a transient signal with a short rise-time is applied to the amplifier, the input stage(s) overload for a brief period of time, until the transient arrives at the output and the correction signal is fed back to the input. For a simple transient, such as a step function, the result is merely a slowing down of the step at the output. If, however, the input consists of a continuous tone, plus a transient, then the momentary overload will cause a loss of the continuous tone during the overload period and the creation of instantaneous dynamic intermodulation products. These specs are never specified in commercial hi-fi amplifiers, as they are difficult to measure, usually very poor, and sound terrible.. Some designers now hold the view that in current amplifier designs, harmonic and intermodulation distortion levels are so low that transient effects are the main cause of audible differences between designs, and the area in which the most improvement can be made. " |
Drew -- Yes, lack of phase coherence and time alignment between the drivers of multi-way speakers is also a frequent cause of problems relating to clean transient response. But excessive negative feedback in amplifiers is a separate and comparably significant cause of similar symptoms. Just Google "transient intermodulation distortion" and you'll find lots of interesting reading. I'm not particularly aware of specific dac implementations being prone to TIM problems -- perhaps others will comment. And yes, the character of harmonic distortion in tube designs is widely regarded as consisting of low order even harmonics, while solid state harmonic distortion, at least at high power levels, tends toward being primarily odd harmonics. Perhaps in turn that leads many solid state designers to apply negative feedback too liberally, reducing harmonic distortion but leading to TIM and problems handling transients cleanly. Regards, -- Al |
"I have just been through 5 different amps using the same DAC, speakers and cables (Monarchy DAC24, Green Mountain Europa, anti-cables) and am struck with how differently each amp reproduces instrument timbre." The amp question aside, try tube rolling your Monarchy dac. This dac is very good in my opinion but voicing by using different tubes can effect the signature greatly. |
Regarding the Haas effect, the thing that interested me most about the amplifier differences was that they were all perceive through the same listening conditions and speakers which are presumably time/phase coherent. I would think any phase or time shift problems would be introduced by the speakers and not the amplifier. Unless of course, there is some fundamental design flaw in the amp. Al - your explanation of transient intermodulation distortion is very interesting - I understand that this problem can also be introduced by the DAC in certain implementations or maybe that is a different issue. Regarding even vs odd harmonic distortion, maybe this is old information but I understand that tubes introduce harmonic distortion on low-even harmonics and transistors on high-odd harmonics. This is why tubes have long been considered to be more "musical" than SS. I, however, prefer clean SS and find that the harmonic distortion is not a primary determinant in the naturalness of timbre. I must admit I have not spent a lot of time listening to very good (expensive) tube gear so I could just be naive. Thanks for the responses, this is the kind of information I am looking for. drew. |
Hi Drew, I think that your question is put in an exceptionally perceptive manner. One major factor that I think is relevant would be the Haas Effect: http://en.wikipedia.org/wiki/Haas_Effect Basically, our hearing mechanisms "latch on" to the leading edge of transient waveforms, and give them disproportionate emphasis relative to what may follow a few milliseconds or a few ten's of milliseconds later. We evolved that capability to aid localization of the source of sounds that may arrive at our ears via both a direct path and (slightly later) via reflections. The Ayre's no-feedback design approach would, everything else being equal, reproduce the leading edges of transients more cleanly than an amplifier that uses significant amounts of feedback. That is because of the non-zero amount of time it takes for the signal to propagate around the feedback loop. During that amount of time the input may have changed significantly, and the correction being applied by the feedback process may be based on out-of-date information, so to speak. What is called transient intermodulation distortion is one of the adverse effects of that. Creating an amplifier that uses little or no feedback while still performing well in other respects (linearity, low output impedance, low harmonic distortion, etc.) requires both very high quality parts, and a lot of care and intelligence in the design. Also, for any given amount of harmonic distortion that the amplifier may produce, its structure is very significant. It's well recognized that low order even harmonics (such as 2nd and 4th) are euphonic and musical, while odd harmonics (3rd, 5th, etc.), and especially high order odd harmonics (9th, 11th, etc.) are the opposite. Harmonic distortion is primarily the result of non-linearity in the transfer function of the amplifier, linearity meaning that a plot of output amplitude vs. input amplitude should be a straight line, within the amplifier's power range. I'm sure others will contribute lots of other thoughts on your question, because there are obviously many factors that are involved, but in my case these are the ones that come to mind first. Regards, -- Al |
