perceptible audio distortion

[Don Klipstein]

This varies widely with the type of distortion.

For "harminic distortion", this varies widely with which harmonics are
formed and what frequency range is involved.

For a "1-size-fits-all" figure, a gentle nonlinearity that causes a
sinewave to take on mostly second harmonic becomes audible around 1% THD
as far as I have heard. A very sharp clipping becomes audible at much
lower THD - maybe .2% on most voice and music signals.

Sinewaves with distortion generating mainly second harmonic will have
audibility threshold of distortion vary widely with frequency. I have
seen sources saying awfully high figures over 10% for very low audio
frequencies.

Some frequencies of sine waves with sharp clipping may have distortion
audible at under .1%.

If all distortion harmonics are ultrasonic, then "harmonic" distortion
is inaudible except for any audible shift in the strength of the
fundamental.

- Don Klipstein ([email protected])

 

[Don Klipstein]

"Glenn Gundlach"


** Really ?

Maybe only the uA709 with zero bias on the output followers.

I have heard *very audible* crossover distortion with LM324. At least
with National Semiconductor ones that Rat Shack sold in the early 1990's.

- Don Klipstein ([email protected])

 

[Don Klipstein]

My immediate thought was the TLO-74, but I'm not sure why.
Probably some piece of crap I had to repair in my earlier years....

I have sometimes had TLO74 act up, probably more with capacitive loads
such as shielded cables. Often a 150 or 330 ohm resistor between an
op-amp output (with unity or otherwise low wired gain) and the conductor
of a shielded cable works wonders!

I have had TLO84 behave better when TLO74 did not, though TLO74 was said
to be less noisy, at least in the 1980's.

TLO7* and TLO8* have evolved over the decades. YMMV!!!

- Don Klipstein ([email protected])

 

[Don Klipstein]

It's very subjective and depends highly on the harmonic structure of the
distortion. Odd higher order harmonics tend to sound the worst. 2nd may
be virtually inaudible at several percent.

Although 2nd is often inaudible at several percent, I find it to be
"repeated conventional wisdom" that odd-order is worse than even-order.

I speak from experience trying various softnesses and hardnesses of
diode clippers on voise, music in general, and guitar signals with
clipping/squashing on both positive and negative peaks and also on peaks
of only one polarity. I did this with 1N34A and 1N4148.

- Don Klipstein ([email protected])

 

[Don Klipstein]

AT FULL POWER typically. The numbers aren't that simple. There really is
so much to it. Simple THD is almost a superficial measure now.

At average listening levels they WILL be better and some MUCH better than
others.

The clue ? You can still hear different amps sounding seriously different
even with the same speakers. Fact.

Also, loudspeakers mostly distort a sinewave with 2nd and 3rd harmonic
(until pushed really badly) while amps generally do their audible
distortion with higher order harmonics.

2nd is generally both less audible and more tolerable. 3rd is
significantly worse than 2nd but a lot more tolerable than 6th and 12th,
13th and 14th when those are at audible frequencies.

- Don Klipstein ([email protected])

 

[Phil Allison]

"Don Klipstein

"Glenn Gundlach"

I have heard *very audible* crossover distortion with LM324. At least
with National Semiconductor ones that Rat Shack sold in the early 1990's.

** The LM324 is no audio grade op-amp.

It was designed for single supply, low current applications - ie battery
operation.

The output stage is a real odd ball - requiring direct coupling of the
load ( no caps ) if x-over distortion is to be avoided.

See page 8 of http://www.national.com/ds/LM/LM124.pdf

Even then, its overall poor performance counts it out, except maybe for
guitar stomp boxes.


..... Phil

 

[Phil Allison]

"Don Klipstein"

Also, loudspeakers mostly distort a sinewave with 2nd and 3rd harmonic
(until pushed really badly) while amps generally do their audible
distortion with higher order harmonics.

** Complete bollocks.

2nd and 3rd harmonics are the DOMINANT ones in nearly all audio amplifiers -
both tube and SS.

Higher order harmonics are typically at such low levels ( ie -90 dB) they
fail to reach levels that would be audible through a speaker by anyone in a
normal room.



..... Phil

 

[Don Klipstein]

In <1453596a-a2d0-4a3f-86b8-d55c72e3bd98@c65g2000hsa.googlegroups.com>,

Once a speaker is involved, all bets are off regarding comparing sound
since the damping factor will most likely be different between
amplifiers.

That only affects frequency response (to the extent loudspeaker
impedance varies with frequency) and resonant character (to the extent
tghe loudspeaker has resonances with impedance peaks/dips).

Suppose amplifier A has a damping factor of 20, amplifier B has a
damping factor of 500, and you try each with a loudspeaker having a peak
impedance 5 times nominal and a dip impedance 75% of nominal.

Amp A compared to Amp B accentuates frequency response at the
impedance peak by about .323 dB and de-accentuates frequency response at
the impedance dip by about .17 dB. In the likely event that the impedance
peak in this case is resonant in character, Q of the resonance is
increased by about 3.79%.

I would think that if one amp had THD .2% in higher order harmonics
and did so at frequencies as low as 500 Hz and the other had THD .02%,
then the harmonic distortion and intermodulation distortion would outweigh
effects of damping factor.

- Don Klipstein ([email protected])

 

[Don Klipstein]

Ironically, I get more bass from my tube amps than from a low-Z solid state
amp. These speakers are vintage Magnavox cabinets from the '50s, so the old
paper suspension midranges really resonate a lot. The bass is overwhelming
when driven with a constant-current amp (Z ~ 100 ohms)! Probably
underdamped at low Z.

Most loudspeakers with actual bass response have an impedance peak at
some bass frequency at which they have useful frequency response. Lower
output impedance amplifiers emphasize the loudspeaker's frequency response
where such an impedance peak occurs.

One example is "Marshall" type and similar 4-loudspeaker closed-box
guitar loudspeaker cabinets. Amplifiers whose output impedance rises
when pushed into distortion (due to loss of
output-impedance-lowering-effect of negative feedback) give actual bass
effect with these cabinets when pushed into distortion. Such amps are
often tube amps. Solid state amps often fail to make "Marshall"
"traditional" type/style (my words) loudspeaker cabinets get similarly
bassy. I have found a fix to be putting a power resistor (or bank
thereof) with sufficient wattage rating and resistance around or somewhat
over the loudspeaker cabinet's nominal impedance to "fix" this at least
largely - though lowest notes when "clean" get "boomed up" somewhat.

- Don Klipstein ([email protected])

 

[Don Klipstein]

"Don Klipstein


** The LM324 is no audio grade op-amp.

It was designed for single supply, low current applications - ie battery
operation.

The output stage is a real odd ball - requiring direct coupling of the
load ( no caps ) if x-over distortion is to be avoided.

See page 8 of http://www.national.com/ds/LM/LM124.pdf

I was already aware of the LM324 having a high tendency to only behave
well without crossover distortion in single-supply circuits.
I wonder why you did not cite this along with ua709 before I mentioned
it from actual experience working with it.

- Don Klipstein ([email protected])

 

[Don Klipstein]

"Don Klipstein"


** Complete bollocks.

2nd and 3rd harmonics are the DOMINANT ones in nearly all audio amplifiers -
both tube and SS.

Higher order harmonics are typically at such low levels ( ie -90 dB) they
fail to reach levels that would be audible through a speaker by anyone in a
normal room.

You try pushing a low-THD SS amp (.05% or less) with a sinewave to 2 dB
past having THD twice rated and see if harmonic content past 5th is so
much as 60 dB down!

- Don Klipstein ([email protected])

 

[Phil Allison]

"Don Klipstein"

I was already aware of the LM324 having a high tendency to only behave
well without crossover distortion in single-supply circuits.

** You have it wrong yet again.

The x-over distortion issue is with cap coupling, not single supply.

I wonder why you did not cite this along with ua709 before I mentioned
it from actual experience working with it.

** The OP said "some old opamps" - which the uA709 certainly is, having
been around since 1965 and one of the first ever made. OTOH - the LM324
is relatively modern, having appeared in 1974 plus, as I said, not used in
hi-fi audio.

FYI - the popular audio grade op-amp appeared as follows.

uA741 = 1968
LM101 = 1968
LM301A = 1969
uA748 = 1969
RC4558 = 1974
LF355 = 1975
TL081 = 1975
TL071 = 1976
NE5534 = 1977

Most are way older than folk imagine.



....... Phil

 

[Phil Allison]

"Don Klipstein"

You try pushing a low-THD SS amp (.05% or less) with a sinewave to 2 dB
past having THD twice rated and see if harmonic content past 5th is so
much as 60 dB down!

** Try sticking your fat head down a dunny

- you pathetic, know nothing, old bullshit artist.




...... Phil

 

[Les Matthew]

Oh yeah. Now I remember.
I think it was a BGW guitar amplifier.

One of the worst designs I ever had to repair!!.

I always found BGW's easy to repair, mind you it was back in the
eighties and were mainly 750's.

les...

 

[[email protected]]

TL084's were my "jelly-bean" OpAmp when I did discrete designs.

There is a trivial solution to driving capacitive loads. I believe
I've addressed that issue on my website: SED section.

...Jim Thompson

This?

http://www.analog-innovations.com/SED/LM324StabilityWithCapLoad.pdf

Michael

 

[[email protected]]

I love reading the really serious-sounding "dip your stuff in LN2"
sites. For a while at least there was a site that sold specially
preselected, matched vacuum tubes that had been dipped.

I mean, it's just intuitively obvious that taking a tube that's been
stored for 40 years and dipping it in LN2 is going to make it sound
better, right?

--

Tim Wescott
Wescott Design Serviceshttp://www.wescottdesign.com

Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" gives you just what it says.
See details athttp://www.wescottdesign.com/actfes/actfes.html

And not just dipped in LN2. Dipped how many times? For how long?
Was the wire instantly insulated to prevent any oxygen from diffusing
in?

Michael

 

[[email protected]]

Ironically, I get more bass from my tube amps than from a low-Z solid state
amp.  These speakers are vintage Magnavox cabinets from the '50s, so the old
paper suspension midranges really resonate a lot.  The bass is overwhelming
when driven with a constant-current amp (Z ~ 100 ohms)!  Probably
underdamped at low Z.

Tim

More bass is generally what you get. But is it accurate?

 

[[email protected]]

<[email protected]>=  TROLL






We will call that your opinion. Tube amps are known for their low
damping factor since their loop gain is generally not as high as solid
state designs.

** Knew the fuckwit TROLL was gonna shift the context like that.

Facts:

Tube amps with unusually low damping factors are simply not hi-fi amplifiers
AT  ALL  !!!   They are either guitar amplifiers or replicas of antique tube
junk from the pre hi-fi era of the 1930s or 40s.

Virtually all SS hi-fi amps have high damping factors and so do the vast
majority of  *genuine hi-fi* tube amps.

.....    Phil

Really? I assume that anyone who puts "fuckwit TROLL" in a post knows
nothing.

 

[[email protected]]

In <1453596a-a2d0-4a3f-86b8-d55c72e3b...@c65g2000hsa.googlegroups.com>,






  That only affects frequency response (to the extent loudspeaker
impedance varies with frequency) and resonant character (to the extent
tghe loudspeaker has resonances with impedance peaks/dips).

  Suppose amplifier A has a damping factor of 20, amplifier B has a
damping factor of 500, and you try each with a loudspeaker having a peak
impedance 5 times nominal and a dip impedance 75% of nominal.

  Amp A compared to Amp B accentuates frequency response at the
impedance peak by about .323 dB and de-accentuates frequency response at
the impedance dip by about .17 dB.  In the likely event that the impedance
peak in this case is resonant in character, Q of the resonance is
increased by about 3.79%.

  I would think that if one amp had THD .2% in higher order harmonics
and did so at frequencies as low as 500 Hz and the other had THD .02%,
then the harmonic distortion and intermodulation distortion would outweigh
effects of damping factor.

 - Don Klipstein ([email protected])

The ear is more sensitive to changes in amplitude for wideband signal
sources than say sine waves. You can demonstrate this with a pink
noise generator.

There have been double blind tests regarding if people can tell the
difference between amplifiers. [Double blind in the sense that the
person switching the amps is behind a curtain so that no visual clues
are given.] The tests were "same" or "different". Midband amplitudes
were made equal. The results were that some people can tell the
difference between amplifiers on a basis that exceeds statistical law
of chances.

 

[Eeyore]

I have always wante to run the following experiment:

Gather the CEO's of all these cable companies together in a room.
Play some music using their expensive audio-fool cables.

Then, substitute those expensive cables with long strips of cut-up,
old-style, metal garbage can lids pop-riveted together to form a
suitably long enough "cable". Even better if the lids are grimey,
with a few spaghetti sauce stains here and there, or whatever...

Then, ask them to tell the difference. (Adjusting only amplitude if
needed to overcome lid loss, ohms only, not response).

I would bet you that "most" would not make the distinction any better
than random chance!

Studio Sound once found that rusty nails made acceptable binding posts.

Graham