In article <103udn0$27r2q$3@dont-email.me>, G6JPG@255soft.uk says...

On 2025/6/30 11:45:39, Philip Herlihy wrote:

In article <103q74h$15dk2$1@dont-email.me>, G6JPG@255soft.uk says...

One of them being at full scale - doesn't matter which one - will
contribute no degradation (the factor is 1). So normally, do that -
unless that means you have to set the other one _so_ low that you get
rounding-error noise.

In practice, assuming you're using 16-bit representation or better,
you're unlikely to be able to hear _any_ effect.

Not my experience. On my ageing Dell Vostro, if I have Windows Media
Player set to maximum output there is clear "clipping", which gives the
sound a harsh edge. Turn down the output control on WMP (compensating
as needed with the speakers' volume control) and the distortion
disappears.

Ah, you're using external speakers, with a physical volume control (a
knob, or up and down buttons)?

Rotary dial. In another sub-thread I've been made to realise that I've overlooked the "system" volume control, which I'll have to play with nex
time I play music!

--
--
Phil, London

--- MBSE BBS v1.1.1 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

In article <103udhu$27r2q$2@dont-email.me>, G6JPG@255soft.uk says...

On 2025/6/30 11:39:26, Philip Herlihy wrote:
[]

It'll depend on the hardware you have, but on my Dell Vostro, I set
Windows Media Player to about 2/3 max volume, and then adjust my
speakers' volume to taste. With WMP on max output, I get that harsh
sound that comes from an overload somewhere. You should experiment.

Interesting. When you say "WMP on max output", do you mean the slider in
the WMP window? Does it still sound like an overload with that on max.,
but with the master volume control (near the clock) set to _less_ than max.?

It's always good to have it pointed out you've overlooked something!
I've ignored the "system" volume control, simply adjusting the WMP
control and the speaker volume (rotary dial). Next time I play any
music I'll experiment!

Thinking about it more, I realise I don't really understand quite what's
going on when clipping is detected. I'm using an external USB DAC
(Cambridge Audio Dacmagic). So the signal is still digital when it
reaches that - the speaker volume control is downstream (analogue
signal). So how does a digital output (however loud it's coded to be) overload a downstream input stage? I wonder if this is a shortcoming in
my DAC - though it certainly delivers better sound than the standard PC
jack.

--
--
Phil, London

--- MBSE BBS v1.1.1 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

On Tue, 7/1/2025 8:36 AM, J. P. Gilliver wrote:

On 2025/7/1 11:27:4, Philip Herlihy wrote:

In article <103udhu$27r2q$2@dont-email.me>, G6JPG@255soft.uk says...

On 2025/6/30 11:39:26, Philip Herlihy wrote:
[]

It'll depend on the hardware you have, but on my Dell Vostro, I set
Windows Media Player to about 2/3 max volume, and then adjust my
speakers' volume to taste.ÿ With WMP on max output, I get that harsh
sound that comes from an overload somewhere.ÿ You should experiment.

Interesting. When you say "WMP on max output", do you mean the slider in >>> the WMP window? Does it still sound like an overload with that on max.,
but with the master volume control (near the clock) set to _less_ than max.?

It's always good to have it pointed out you've overlooked something!
I've ignored the "system" volume control, simply adjusting the WMP
control and the speaker volume (rotary dial).ÿ Next time I play any
music I'll experiment!

Thinking about it more, I realise I don't really understand quite what's
going on when clipping is detected.ÿ I'm using an external USB DAC
(Cambridge Audio Dacmagic).ÿ So the signal is still digital when it
reaches that - the speaker volume control is downstream (analogue
signal).ÿ So how does a digital output (however loud it's coded to be)
overload a downstream input stage?ÿ I wonder if this is a shortcoming in
my DAC - though it certainly delivers better sound than the standard PC
jack.

Unless it's been badly set up, a DAC should not produce clipping, unless fed with a digital signal that already includes clipping. It will produce a signal of a given voltage. That could still be more than whatever it is feeding is capable of accepting.

Ideally, you need something that can produce a signal you can vary the output of, in an analogue fashion: ideally a signal generator, but maybe a CD player playing a test tone, with a volume control. Feed this into your "speaker" (amplified speaker system), and turn it up until you hear distortion; at that point, measure the output with a voltmeter (with it still connected to the speaker input); that is the maximum voltage the speaker can accept. (Doesn't matter if the voltmeter isn't "true RMS" or what frequency you use, though I'd suggest a low one - as long as you use the same signal for all measurements.) Then measure (again, while connected to the speaker load) the output of your DAC playing the same signal. If this is higher than the speaker can take, you've found the problem. (The solution would be a resistive dropper - or, always run the system volume turned down sufficiently; this latter in theory affects fidelity, but in practice I challenge anyone to be able to
hear any degradation.)

Do this when you have the house to yourself if you don't want to irritate people!

If you can borrow an oscilloscope, that will show you fairly quickly if any signal you can get at is clipping, and let you find any relevant levels.

Here is some clipping.

[Picture]

https://i.postimg.cc/GmZSBPBn/clipping.gif

And that is sending a sine wave played in windows Media Player Legacy
on one side of the room, to the Line_In on a second PC. The second PC
is using Audacity to record the Line_In signal. Since the tone is 440Hz
and the ADC is 44.1KHz, each (damaged) sine wave is sampled 100 times
per wave. The screen representation fits a continuous curved line to the discrete dots, giving the impression of a much much higher sampling rate.

One channel is clipping, the other is not, due to some Balance
setting that is not in the center of the scale. That saves me from
taking two pictures.

A square wave, is a summation of sine waves. The components of the
clipped waveform, now include at least two sine waves, and quite
possibly a lot more harmonics of the fundamental (all with unique
phase and amplitude to make the waveform what is is). These
off-tones, serve to give the 440Hz clipped sine wave tone a
"raspy sound that is not quite right".

If I connected the Line_In level computer speakers to that signal,
the volume knob on the speaker would need to be turned down. Maybe
the computer speakers move the cones to full excursion, when only
a 0.07 VAC input signal is received. By setting the volume pot to
1/14th of full scale, I can avoid blowing the speakers. The volume
pot could be log or linear, I did not take note of the details
while building the kit. The speakers on the test machine, the knob
is NEVER MOVED. It's what you do with quality pots :-) There is a pot
for Left channel, and a pot for Right channel. In fact, the amp
doesn't even have an on-off switch. It's always on.

Paul

--- MBSE BBS v1.1.1 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

In article <1040ko8$2or9a$2@dont-email.me>, G6JPG@255soft.uk says...

Unless it's been badly set up, a DAC should not produce clipping, unless
fed with a digital signal that already includes clipping. It will
produce a signal of a given voltage. That could still be more than
whatever it is feeding is capable of accepting.

I know what clipping sounds like - it's very characteristic. We have a
street preacher who visits our town square who has a habit of belting
out his story with something being overloaded, and the sound is even
less appealing than his admonishments.

My system WMP, PC, USB, DAC clips (irrespective of overall volume level)
on some CDs if the WMP level control (and the system level) are at 100%.
If I turn down the WMP control, I get clean sound, which I can then
amplify to considerable levels without loss of quality. Yes, I'd be interested to know quite which stage is being overloaded.

--
--
Phil, London

--- MBSE BBS v1.1.1 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)

On Tue, 7/1/2025 1:27 PM, Philip Herlihy wrote:

In article <1040ko8$2or9a$2@dont-email.me>, G6JPG@255soft.uk says...

Unless it's been badly set up, a DAC should not produce clipping, unless
fed with a digital signal that already includes clipping. It will
produce a signal of a given voltage. That could still be more than
whatever it is feeding is capable of accepting.

I know what clipping sounds like - it's very characteristic. We have a street preacher who visits our town square who has a habit of belting
out his story with something being overloaded, and the sound is even
less appealing than his admonishments.

My system WMP, PC, USB, DAC clips (irrespective of overall volume level)
on some CDs if the WMP level control (and the system level) are at 100%.
If I turn down the WMP control, I get clean sound, which I can then
amplify to considerable levels without loss of quality. Yes, I'd be interested to know quite which stage is being overloaded.

https://www.tenforums.com/sound-audio/208334-output-volume.html

Using some of the info there, you can get a dB reading for the master Volume.

[Picture]

https://i.postimg.cc/wjdkbX12/sound-in-db-legacy-control-panel.gif

But the quantification of sound and tracks is endless.

https://www.audiosciencereview.com/forum/index.php?threads/ending-the-windows-audio-quality-debate.19438/page-19

It seems the windows Way, is for Groove Music player to automatically
fight against the user setting :-)

Ultimately, the knob(s) on your external amplifier cascade,
are the final arbiters of what you hear.

*******

In theory, you could use Stereo Mix (AKA "What you hear") as
a means of monitoring the output. That takes a sample of the
output and hairpins it back towards the ADC blocks. But that
is no longer a neutral agent -- I tested Stereo Mix here, and
I could not observe clipping quite the same way as using
a second computer to monitor it.

The HDAudio CODEC has gain and attenuation blocks (popular models
you can find the hardware block diagram for this), in 1.5dB steps.
If you could read out the settings on those, you could do a
more deterministic calculation of the output range possible
for given settings. But if WMPL is using "math for gain",
by scaling the values by some float-value, then looking
at the hardware control settings is useless. And that's the
sort of thing VLC has been doing to various extents. Unless
WMPL labels the slider properly, there's no way to know
exactly what it is doing.

On modern computers, the sound path is not necessarily
HDAudio any more. Which will introduce a whole new set of
variables to study. For example, some laptops have
"digital amp chips" which have a "status", and your
speakers can claim they are "Not Ready". Which I'm
sure leads to customers shitting square bricks of excitement,
when the speakers no longer work.

Paul

--- MBSE BBS v1.1.1 (Linux-x86_64)
* Origin: A noiseless patient Spider (3:633/280.2@fidonet)