Audio levels and testing Passive Volume Controls

[stereo coffee]

There seems to be incorrect procedure when measuring passive attenuators at ASR.

In the recent review here :https://www.audiosciencereview.com/...s/schiit-freya-s-preamplifier-review-2.11543/
we see " It performed really well in passive mode, providing fully transparency to my ultra low distortion and and noise audio analyzer"

In truth there really was no test at all, of the passive attenuator, because the unit was arranged to have maximum volume - and therefore we can presume all that was being measured was the units internal wiring, as there was effectively a short circuit from input to output. No wonder it measured well.

Suggestions for future passive measurement
Passives need I suggest to be measured, from where audio first becomes apparent and suggested then at 5 different incremental settings upward. We would then see in the presumed ethos of the forum, "audio science" the effect resistance and possibly where present contact distortion, or other distortions are occurring; https://www.pinkfishmedia.net/forum/threads/alps-potentiometers-gain-and-distortion.68076/ and digital variants https://www.diyaudio.com/forums/analog-line-level/227693-digitally-controlled-pot-distortion.html

RMS levels
Secondly the RMS voltage level ASR uses is unrealistic IMO to what any known CD contains with level. This then can change the need of objective assessment to be awkwardly distanced from what we hear. Rather 4V RMS may be the potential a analog output can swing into a given load, .... but the actual CD itself, I would argue can never deliver this figure.

Audacity -lets check some CD's
Readers can with minimal expense and I hope will indeed prove this to themselves, by installing Audacity then allowing that programme to reveal the RMS level shown as an example in the image. The image is a flac file of vangelis 1492 track 12, loaded into audacity.

The figure of measurement being input with a passive should by all means attain the maximum as the last of 5 volume measurements - but the maximum should be related to the present industry standard applying to the level applied when digitising of a recording occurs, which as I see is usually one in the same as the recording, unless specifically level adjusted by the artist during mastering. We can see this for ourselves using audacity. https://www.audacityteam.org

Questions arise like is CD level and streaming level the same ?

Lets check Audacitys scale system first
However firstly we need to assess how close audacity is itself to being correct with its scale shown as 1.0 ,0.5 ,0.0 and the same in the lower side direction to reveal a AC signal. Audacity does not state the scale to be voltage - but by all accounts it is exceedingly close, in this case as a
measure of 2v , being 1v positive and 1v negative checks, with a AC volt meter measuring to be the same.

If we apply levels which are used every day that we listen to , we then get closer to aligning objective vs subjective assessment as being the same.

 

[Blumlein 88]

I'd think the most important level in a passive volume to measure is at the -6 db point. This will be maximum output impedance which will cause maximum FR difference when it interacts with the following devices input impedance and cable capacitance.

I've done reasonable measures like in your link and have not found distortion to rise like is being described. I have found distortion in some pots that were faulty as in dirt contaminated or oil contaminated. With a pot testing at -6 db should give a reasonable idea.

I did prefer switched resistors anyway.

 

[JohnYang1997]

I can only agree on the attenuation part. Not so much for the rest.

 

[JohnYang1997]

I'd think the most important level in a passive volume to measure is at the -6 db point. This will be maximum output impedance which will cause maximum FR difference when it interacts with the following devices input impedance and cable capacitance.

I've done reasonable measures like in your link and have not found distortion to rise like is being described. I have found distortion in some pots that were faulty as in dirt contaminated or oil contaminated. With a pot testing at -6 db should give a reasonable idea.

I think -1, -3, -6, -20, -60 are good five steps. -6db can be good for single dashboard but these 5 steps are more informative than just 1. -1db for example will very likely to introduce distortion due to voltage coefficient of resistor.

 

[jsy]

In truth there really was no test at all, of the passive attenuator, because the unit was arranged to have maximum volume - and therefore we can presume all that was being measured was the units internal wiring, as there was effectively a short circuit from input to output. No wonder it measured well.

I asked a similar question, and Amir pointed out that he had tested the Freya at several attenuation settings in the initial review. Did those tests not apply to the unit in passive mode?

RMS levels
Secondly the RMS voltage level ASR uses is unrealistic IMO to what any known CD contains with level. This then can change the need of objective assessment to be awkwardly distanced from what we hear. Rather 4V RMS may be the potential a analog output can swing into a given load, .... but the actual CD itself, I would argue can never deliver this figure.

Perhaps I'm missing something, but how would testing a unit at a slightly higher input voltage than normal sources provide have a negative influence on the test results? I understand that increased input signal strength would tend to improve SNR measurements, but wouldn't that effect apply equally across all devices tested?

 

[RayDunzl]

CDs have a voltage?

DACs produce a voltage, the maximum of which varies among devices.

Assuming the maximum specified output of a common DAC is reported as 2V, that is likely RMS, is equal to 2.8284V peak, or 5.6567 peak to peak.

So, by that measure, "1" on Audacity would be 2.8284V. The waveform does not trace RMS, it is instantaneous, and shows peak at the extremes.

0,5 on Audacity represents half-value, etc.

The voltage output of a DAC will bear a linear relationship to the values shown on the Audacity waveform.

---

Actually, I'd say "1" on Audacity, and "-1", represent the maximum swings of 16, or 24, or 32 bit digital values.


What happens after that is dependent upon the analog hardware passing the decoded digital signal.

 

[stereo coffee]

CDs have a voltage?

DACs produce a voltage, the maximum of which varies among devices.

Assuming the maximum specified output of a common DAC is reported as 2V, that is likely RMS, is equal to 2.8284V peak, or 5.,

So, by that measure, "1" on Audacity should be 2.8284V. The waveform does not trace RMS, it is instantaneous, and shows peak at the extremes.

---

Actually, I'd say "1" on Audacity, and "-1", represent the maximum swings of 16, or 24, or 32 bit digital values.

What happens after that is dependent upon the analog hardware passing the decoded digital signal.

The RMS figure in audacity is the light blue trace vs dark blue as peaks

 

[stereo coffee]

I'd think the most important level in a passive volume to measure is at the -6 db point. This will be maximum output impedance which will cause maximum FR difference when it interacts with the following devices input impedance and cable capacitance.

I've done reasonable measures like in your link and have not found distortion to rise like is being described. I have found distortion in some pots that were faulty as in dirt contaminated or oil contaminated. With a pot testing at -6 db should give a reasonable idea.

I did prefer switched resistors anyway.

More easily expressed is where the shunt resistance and series resistance meet each other. This is not necessarily at half volume so needs to be checked by measuring resistance, output to signal ground for the shunt, and input to output for series.

 

[Blumlein 88]

The RMS figure in audacity is the light blue trace vs dark blue as peaks

Yes, and I'm pretty sure Ray knows that. I too am having difficulty in linking the Audacity info with testing or using passive volume controls.

 

[JohnYang1997]

More easily expressed is where the shunt resistance and series resistance meet each other. This is not necessarily at half volume so needs to be checked by measuring resistance, output to signal ground for the shunt, and input to output for series.

What in the

 

[Blumlein 88]

More easily expressed is where the shunt resistance and series resistance meet each other. This is not necessarily at half volume so needs to be checked by measuring resistance, output to signal ground for the shunt, and input to output for series.

In a conventional pot that will be -6 db. If you have resistors in some other network of attenuation, then you'll get different answers.

 

[stereo coffee]

What in the

Unrelated to the signal and only considering the passive resistances formed, the stated output impedance of a passive is where series and shunt resistances meet each other.

 

[Blumlein 88]

Unrelated to the signal and only considering the passive resistances formed, the stated output impedance of a passive is where series and shunt resistances meet each other.

Just so we'll all be on the same page.

Take a hypothetical 10k ohm pot.
At -6 db, you'd have in input impedance of 10 kohm. You'd be tapping that at 5 kohm for the reduced output signal. The output impedance would be 2500 ohm. Any other setting has lower output impedance.

 

[mansr]

CDs have a voltage?

DACs produce a voltage, the maximum of which varies among devices.

Assuming the maximum specified output of a common DAC is reported as 2V, that is likely RMS, is equal to 2.8284V peak, or 5.6567 peak to peak.

So, by that measure, "1" on Audacity would be 2.8284V. The waveform does not trace RMS, it is instantaneous, and shows peak at the extremes.

0,5 on Audacity represents half-value, etc.

The voltage output of a DAC will bear a linear relationship to the values shown on the Audacity waveform.

---

Actually, I'd say "1" on Audacity, and "-1", represent the maximum swings of 16, or 24, or 32 bit digital values.


What happens after that is dependent upon the analog hardware passing the decoded digital signal.

That's a very long-winded way of saying "uh-huh."

 

[RayDunzl]

The RMS figure in audacity is the light blue trace vs dark blue as peaks

Yes, and I'm pretty sure Ray knows that.

No, I didn't.

I'm generally unconcerned with RMS levels, never tried to find them before.

 

[stereo coffee]

Just so we'll all be on the same page.

Take a hypothetical 10k ohm pot.
At -6 db, you'd have in input impedance of 10 kohm. You'd be tapping that at 5 kohm for the reduced output signal. The output impedance would be 2500 ohm. Any other setting has lower output impedance.

Agreed, the pot itself in such a circuit has a output impedance of 2500 ohm, and certainly lower impedance can be achieved relative to the source by changing resistance level toward the source, but would then not be an expression of the actual pots output impedance.

 

[RayDunzl]

Secondly the RMS voltage level ASR uses is unrealistic IMO to what any known CD contains with level. This then can change the need of objective assessment to be awkwardly distanced from what we hear. Rather 4V RMS may be the potential a analog output can swing into a given load, .... but the actual CD itself, I would argue can never deliver this figure.

Record and play a full-scale sine wave and it will deliver the full RMS value.

 

[restorer-john]

@stereo coffee

Let's face it, passive volume "pots in a can" open up a Pandora's box of frequency response and noise anomalies at anything other than full volume. Personally, I believe they are complete waste of time for anyone with a plethora of source devices they want to use.

Testing such a device with the volume flat out and looping through a line signal will really only show some loss of channel separation and perhaps a tiny bit of noise and THD that wasn't there before. At levels other than full volume, we'll see all the fun stuff and channel mismatching as well. But if Amir makes one test on a passive, it really has to be a maximum volume. Otherwise he needs to make a ton of individual, arbitrary volume positions with all the associated headache of several times the work.

As for CD and level. The standard for CD player output was, since the start of mass production in 1983, 2.0V. That said, many machines hover slightly above that figure and were traditionally called out for that in reviews. The same figure of 2.0V was also the standard for XLR balanced equipped CD players. That figure however has morphed into various numbers and up to 4V to ensure some form of "compatibility" with pro levels. With more and more home gear sporting XLRs and levels creeping up, one could argue 4V is not unreasonable. It is unreasonable for any RCA line level input on a reasonable preamplifier or power amplifier and bears no relationship whatsoever to traditional line levels and amplifier sensitivity levels of the past.

Buy yourself a ACmV meter, load the outputs with say 10K and play a 0dBFS sine and tell us the level in volts. It will be ~2.0V on the RCAs and either the same, or greater on XLRs.

 

[stereo coffee]

@stereo coffee

Let's face it, passive volume "pots in a can" open up a Pandora's box of frequency response and noise anomalies at anything other than full volume. Personally, I believe they are complete waste of time for anyone with a plethora of source devices they want to use.

Testing such a device with the volume flat out and looping through a line signal will really only show some loss of channel separation and perhaps a tiny bit of noise and THD that wasn't there before. At levels other than full volume, we'll see all the fun stuff and channel mismatching as well. But if Amir makes one test on a passive, it really has to be a maximum volume. Otherwise he needs to make a ton of individual, arbitrary volume positions with all the associated headache of several times the work.

As for CD and level. The standard for CD player output was, since the start of mass production in 1983, 2.0V. That said, many machines hover slightly above that figure and were traditionally called out for that in reviews. The same figure of 2.0V was also the standard for XLR balanced equipped CD players. That figure however has morphed into various numbers and up to 4V to ensure some form of "compatibility" with pro levels. With more and more home gear sporting XLRs and levels creeping up, one could argue 4V is not unreasonable. It is unreasonable for any RCA line level input on a reasonable preamplifier or power amplifier and bears no relationship whatsoever to traditional line levels and amplifier sensitivity levels of the past.

Buy yourself a ACmV meter, load the outputs with say 10K and play a 0dBFS sine and tell us the level in volts. It will be ~2.0V on the RCAs and either the same, or greater on XLRs.

Yes measuring passives is entirely different to other devices, and it therefore just like measuring speakers with different parameters to amplifiers, has to be done differently to assess that device as an attenuator, and not as something else.

The machines may hover but the typical CD we purchase, or hear via streaming, has no where near this level. A 0dBFS sine ~2.0V is not what occurs with conventional CD replay, as the audacity graphs show.

The closer we are with objective measurements not being alien or distant from our hearing in origin, the closer we then get to subjective assessments being the same. Objective measurements should reflect close to what we hear, if they are themselves to have any meaning.

 

[Sir Sanders Zingmore]

Perhaps you could share the measurements you’ve made of your passive device .....