> Extended response MATTERS <

No it doesn't.

Seriously, if the response is truly accurate all the way to ~20 KHz, with minimal ringing and no high Q peaks etc, that's all anyone needs. Consumer level A/D conversion has been up to that task for at least 5 to 10 years now.

--Ethan

This type of thing might be relevant in a studio. High frequencies (above 20KHz) that are not needed can get "folded" in (Nyquist sampling theorem)if they are not properly filtered out before sampling for A to D conversion. One of the benefits of a higher sampling rate (being able to handle more bandwidth) is the reduction in the possibilitiy of artifacts (from a steep filter or from parasitic stuff folded into the audible range)

Fortunately, in reverse it is relatively simple to "upsample" CD data in order to make filtering much easier. Clearly the A to D process is the most delicate.

If the sound is inaudible, however, there is really no value in producing it out the speaker in playback. Contrary to what the article says...a tranducer that accurately produces 100 KHz is actually LESS LIKELY to be good in the much more important 3 KHz to 12 Khz range. If the insanity of this article had any validity then we would use tweeters instead of woofers to produce bass frequencies accurately.

You are correct and some amps cannot handle those HFs either.

Ethan, you surprise me; nibbling at one of DUP's baited hooks?

If you respond to his garbage, you just encourage him; he needs to be totally ignored!

Thanks for the Strad info, by the way; very informative (and that was an excellent article you wrote).

Ignore who? Never heard of this "DUP" fellow.

I heard about extended HF response some years ago, and I believe Van Alstine mentioned it as well. So I did my own investigation, using a passive RIAA tube design I was working on. I wanted to see if what they were saying was true.

Although the article's last half concentrated on CD systems, I wanted to see what would happen in analog territory.

RIAA circuits have TCs of T3=3180us, T4=318us, T5=75us. ("On RIAA Equalization Networks" by Stanley Lipshitz)

Another T6 occurs because of the cutters and lack of infinite gain and room to cut. I believe the turnover frequency was approx 55-65khz.

I manipulated the -1db point from 200khz to 150khz (I know, much higher than the turnover frequency used in studios). Changing T6 involved adding approx 50pf to the passive circuitry.

Anyway, the Point is that making this change, between 200khz and 150khz -1db, resulted in a dramatic sonic change, affecting the perceived sound of the highs, and the perceived sound down to the bass. I would have guessed a very minor change in the highs, nothing as dramatic as what I actually heard.

Another problem in the early development of global feedback amplifiers revolved around some circuits having resonant peaks well above 20khz, 60khz or more, because of poor designing. In any case, depending on the amplitude of the peak, lowering those resonant peak resulted in distinct changes in perceived sound, manifesting itself as less brightness.

Another good point already raised is getting rid of artifacts and minimizing ringing.

The point is that I think we should be careful before making any final cut and dry decisions.