Propeller Head Plaza Technical and scientific discussion of amps, cables and other topics.

---

[ Follow Ups ] [ Post Followup ] Thread: [ Display  All  Email ] [ Propeller Head Plaza ]

---

[ Alert Moderator ]

---

Looking at it backwards

Posted by Jon Risch (B) on April 11, 2009 at 12:12:20

In Reply to: I think Doug Self just flipped me posted by BillyBuck on April 10, 2009 at 14:25:52:

See:
http://www.audioasylum.com/audio/cables/messages/30013.html

There IS a limit to how much distortion, noise, low level nonlinearities, etc. can be added to the music before it suffers audibly. When the signal has been savaged by the studio to such an extent, there is often very little "distortion headroom" left over for the playback system.

So looking at the distortions incurred in the studio as some sort of license to commit further mayhem to the signal is just the wrong way to look at it.
Rather, given what is done to the signal when it is recorded means that we must take every care not to further degrade the signal any more than it has already been degraded.

Finally, regarding objectivist POV, you might want to think about this:
most objectivist concepts involve bench testing as the ultimate metric. How well does a power amp perform while hooked up to load resistors, how well does it measure on THD, IM, S/N, etc.

This SEEMS entirely reasonable, and SEEMS to be solidly based on what has gone before in terms of hearing research, psychoacoustics research, etc.

But several things get overlooked, and these are not the only things that are not being taken into account.

1. Bench tests do not include getting vibrated by the speakers that are in the same room. While this might seem trivial, let me cite an example.

A digital amp prototype sounded pretty rough, the louder the music got, the fuzzier the sound. Bench tests looked great, very low IM, low THD, etc. Come to find out, the designer had used a piezoelectric resonator for the digital clock. Vibrate a piezoelectric device, it generates a spurious signal that jitters the clocking frequency it is concerned with generating.
Put a sandbag on the resonator, sound improved noticeably. Replaced the resonator circuit with a quartz crystal (also piezoelectric, but with much more defined vibrational modes, and a physical mounting system more resistant to outside vibrations) based one, and the sound improved even more.

Bench measurements remained the same however, with virtually no difference between the two versions.

2. Power amplifier measured good on bench into resistive loads, low THD, low IM, all other specs good. Hooked up to actual loudspeakers, it had severe issues with stability: intermittent incipient oscillation, severe clipping distortion (sticking, hanging, oscillation, DC offset, etc.) The reactive load of the loudspeakers was feeding back-EMF into the power amp output, bollixing up the negative feedback circuits and circuit servo's.

Once the reactive loading was taken into account, the circuit could be further stabilized and tweaked. With just a resistive bench load, none of these problems was even seen or anticipated.

In both of these instances, the unit measured just fine, but sounded bad.

Many parts are microphonic or self-microphonic, and not just tubes.

Capacitors, both electrolytic and film caps ( SEE: http://www.icwltd.co.uk/claritycap/download/news2.pdf ), transistors and other semiconductors, resistors, cables, and other parts.

What we hear ALWAYS trumps what we measure, because what typically gets measured is a very limited and tunnel-vision version of the entire reality of what is actually going on.

Then there are the seeming contradictions of what is routinely heard versus what we are SUPOSSED to be able to hear.

If you ever get a chance to talk to recording studio personnel, ask them what their favorite A to D is, and what dither algorithm they prefer for digital recording and playback. Each dither algorithm SOUNDS quite different, they each have their own sonic signature, and some prefer one over the other because of this. So what?
{Dither algorithms have names like: Sony Super Bit Mapping, Apogee UV22, POW-r, etc.)

The action on the signal for ALL of these dither algorithms occurs at levels that are at the -90 dB and below level. If you ask most hard-core objectivist's, they will be glad to tell you that "signal distortions below -60 dB can not be heard" (this is 0.1% or less distortion), yet the very subtle differences between the various dither algorithms are easily noticeable. For reference purposes, -90 dB is equivalent to 0.00316% distortion.

There is no such thing as a meter reading for soundstage depth, there is no such thing as a meter reading for Boogie factor, etc. No one has the beginning of a clue as to what to measure for, or how to do it.

So don't be too quick to hang your hat on the Objective hat rack, there are just too many things that don't get included in the model, included in the measurements, to ignore the subjective input altogether.


Jon Risch

• http://www.audioasylum.com/audio/cables/messages/30013.html   (Open in New Window)

This post is made possible by the generous support of people like you and our sponsors:

Kimber Kable

---

[ Top of Page ] [ Contact Us ] [ Support/Wish List ] [ Copyright © 2025, Audio Asylum, all rights reserved ]

---

[ General ] [ Speakers ] [ Tubes ] [ Vinyl ] [ Digital ] [ Hi-Rez ] [ Video Asylum ] [ Cables ] [ Tweaks/DIY ] [ Music ] [ Films ]

---