On Monday, July 26, 2021 at 7:22:23 AM UTC-4, oben.. at gmail.com wrote:
Skin effect - a second chance at an audio effect
The usual message about skin effect and its impact on high frequencies is that these effects are way above audio frequencies so there is no effect on sound reproduction. But skin effect also occurs as a result of transients (1). Music play back can be full of short fast transients and these transients will push current towards the conductor’s surface.
Yes, music CAN and often DOES contain transients, quite often in fact.
What's important to your thesis is whether those transients contain
sufficient high frequencies as you claim.
Do they? From where do those transients originate, as far as the skin
effect being a significant factor?
Well, this one is easy: they originate from the source of the musical
material being conveyed (at ANY step in the reproduction process). Let's examine the whole chain:
1. The acoustical source: let's pick something that's VERY "transienty":
something like a percussive hit, such as a snare drum, or a triangle. If
we measure, using the WIDEST POSSIBLE BANDWIDTH INSTRUMENTATION
AVAILABLE, yes, we find oodles of high frequency information in that
acoustical sound: someone else mentioned this in terms of the Fourier or
Laplace transform of the music.
But, before we go ANY further, let's examine the caveat I shouted out
above: by "THE WIDEST POSSIBLE BANDWITH INSTRUMENTATION AVAILABLE," what
that means is very wide bandwidth laboratory instrumentation microphones
(they exist, e.g., The Bruel&Kjear 1/8" microphone (type 4138) had a
bandwidth to 140 kHz. These beasts are exceedingly rare, and exceedingly
expensive and pretty noisy to boot: I know of not a single commercially
available recording of ANY type of music using these microphones.
2. The recording process: once out of the rarified laboratory situation
mentioned above, music is recorded with more down-to-earth microphone.
Especially in popular, minimalist high-end reocrding, it's not at all
uncommon to use large- diaphragm condensor microphone: seldom to these
make it to 20 kHz. Even more conventional microphones often used for
spot pickups of drums and the like are typically 1" or 3/4" mics, and
they die pretty much at 20 kHz and above.
I know of no commonly used dynamic microphones that can make it to 20
kHz without a struggle.
THEREFORE: Any high-frequency information resulting from an acoustically
sourced transient is eliminated by the limited bandwidth of the
microphone. A reasonable facsimile of the transient will be captured,
but it will be devoid of any of the high-frequency information you seek.
Let's look at the recording devices:
a. Analog. You have tape, and you have direct to disk. Without going
into the details, these media are limited for all intents and
purposes to about 20 kHz. ANY high-frequency information resulting
from the transient that exceeds that bandwidth that managed to make
it through the microphone will be eliminated by the storage and
subsequent processing stages.
b. Digital: while tthere are recording systems that support
high-bandwidth digital encoding and storage, what's more important
is what's finally delivered to the listener. And unless you can
point to any currently available delivery media that is an
excpetion, all such media is limited to a 20 kHz or so bandwidth.
THEREFORE: Any high-frequency information that makes it through the
bandwidth limited microphone is eliminated by the limited bandwidth of
the recording/processing/delivery media. A reasonable facsimile of the
transient will be recorded, but it will be devoid of any of the
high-frequency information you seek.
The important point here is that this high-frequency information that you
point to as a result of "transient" IS NOT and, indeed, CAN NOT be present
in recording delivered to mere mortal such as you and I that can be played
on equipment owned by mere mortals such as you and I.
Frequency skin effect is not a two state phenomenon but increases with frequency. what if transient skin effects are additive to frequency skin effects.
Unless you have an alternative view of such things such as Fourier and
Laplace mathematics, or have reason to cast serious doubt on basic
principles such as the time-frequency dality of real signals, then you need
to understand, simply, that:
TIME effects ARE FREQUENCY effects,
and
FREQUENCY EFFECTS are TIME EFFECTS.
If you can demonstrate that this is NOT the case with the same kind degree
of mathematical and scientific rigor used by the likes of Fourier, Laplace, Shannon, and others then, sorry, you're faced with the cold hard fact that
the material you're playing IS band-limited, the transients in such
material IS band limited, and the transient, which may be present, have all their energy within the combined bandidth of each and every step of the process.
If so would music frequency skin effects, under certain conditions, now be noticeable at audio frequencies? Do transients shift the skin effect and frequency curve?
No, transients are NO different than EVERY OTHER MUSICAL SIGNAL.
Skin effects might deteriorate sound reproduction for the human ear - high frequency smearing.
This is conjecture, at best.
The effect would be music dependent as some music will be more transient
rich compared to other music.
This does not loigically follow from a premise which itself is not
logically supported.
And some transients will be larger.
There is NO difference between a transient whose bandwidth is limited to a certain amount and ANY other signal within those same constraints as far as
the effect on frequency-dependent phenomenon such as skin effect are
concerned.
And before you run back to your thesis, the ultimate amount of high
frequencies on must concern themselves with at ANY point is ALWAYS
constrained by the bandwith of the material presented to that stage,
REGARDLESS of the nature of the signal. That you have a recording/delivery chain whose bandwidth is limited to, oh, 20 kHz, even when presented at
it's input with a "transient", will not delivered or be required handle ANY information outside of that 20 kHz bandwidth.
UNLESS: the system is driven into non-linearity (distortion, i.e., the
system is broken).
This suggests that the sound deterioration will appear inconsistent
between users and within user systems.
Logically, it does not, because your premise that somehow transients in a band-limited system have components outside of that bandwidth is simple
false.
Can we hear skin effect smearing? And what does it sound like? Is the
impact of skin effect more prevalent with tinned or silver plated wire in
the circuit? the plating material will have different electrical properties compared to the wire and this would argue for a greater effect. Sourcing
the internet
Not what one would consider a good source for information on signals, transients, bandwidth, hearing, effects on hearing and verifying such
effects.
I would use the curent thread as an existence proof of such.
- a number of unprompted posts
unvetted, uncontrolled, unverified, unsupported, unrepeatable
draw attention to the sound signature differences between plated and
unplated wire - mainly that silver plated wire sounds bright and hard in
the upper frequencies.
If one searches the internet for posts regarding silver-plated wire and
bright or hard high frequencies, guess what: you'll find posts regarding silver-plated wire and bright or hard high frequencies.
What does this prove? That you can find ANYTHING you want on the internet.
Dick Pierce
--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)