OT: audio levels
jb-electronics
I am thinking of building a simple all-in-one VU meter using a
microcontroller to visualize the audio signal on a headphone line (say,
I feed the headphone output of my PC to my circuit, and then I install a
pass-thru to connect to my headphones).
What is the typical voltage amplitude I should expect? I do realize this
question is quite general, but perhaps you can point me in the right
direction? Many thanks!
Best wishes
Jens
David Forbes
--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send an email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/2a4dfbef-7430-06ac-b272-973d84c6dae8%40jb-electronics.de.
For more options, visit https://groups.google.com/d/optout.
johnk
1). Don't call it VU unless you match the actual ballistics of the ORIGINAL VU meter
2). Europe tends to use a PPM [quasi-peak] system. Peak Program Meter.
3). Consumer and pro levels are very different. They are often misunderstood and hence mounds of garbage appear on the web.
4). dB is a ratio and a reference must be stated. Thus you must choose what reference your 0VU is.
5). Your headphone out MIGHT be the Line Out. There are [quite a few] standards for Line Level.
6]. A useful fact relates to the maximum level that can be digitally encoded. A CD can produce a certain maximum digital value. That gets converted to some analogue value. You can work empirically by using [say] standard test tones from the web OR an audio program like Audacity to produce known digital levels. 0 dBFS is the max level... BUT there can be output results higher than this [long topic], so work on using either -0.3dB or even -3dB below (or do I mean over? Semantics ) this as maximum.
7). In "recent" years the "loudness wars" and misuse of Digital Audio Workstations has resulted in CDs and digital audio having bugger-all dynamic range. Magnetic tape was automatically a limiter [curved]. People using digital should be working [say]20dB down from max.... but the signal level looks so tiny to them in "oscilloscope" mode. They think that they are getting better results by working "hot" and getting the best sig/noise ratios. Their loss of dynamic range is not a good outcome. I mention this because it DOES relate to the choice of 0 dB[ref].
BTW.. some digital audio actually clips! The master has too much dynamic range for the "cutting" [vinyl, cd , web etc] engineer. He wants it to sound HOTTER. Sigh.
8). Talking the voltage levels. The term dBU refers to a bridging voltage measurement. These days you definitely do NOT want to try expecting 600 ohm impedances. The "usual" outputs are these days exceptionally low... eg way under 20 ohms. This way you can feed a great number of inputs from one output. Big topic. The problems start here for your question. Many consumer outputs place a series resistor for protection purposes. This doesn't affect Bridging (high Z meter) readings much, BUT the input impedance of the loads does. And the number of such loads connected.
There are many manufacturers who conform to certain industry standards, and many more who don't.
9). Read Bob Katz book Mastering Audio. Downloadable, although I didn't check which editions. I use 2nd edit for recording because I do not do web stuff. His later edition(s) cover the "modern" approach. [read Amateur Fiddling for modern :-)) ]
10). I suspect that the most useful meter for your purposes is one which relates well to the maximum signal level and has fast response. BUT VU is so averaged and hence slow. There are many so-called VU that are fast reacting and hence display peaks. And be careful of the definitions here when you see rms stated. As David said, work with peak voltages that you can see on a scope. This is where you need to start looking at the Orban meters. I haven't checked but they were downloadable. Apart from the damping factor for Peak, his special meters that relate to perception are interesting. He covers the standards for TV advertising versus program content. And that leads to discussions of loudness [not volume].
Reminder... VU can ONLY be used when the meter has the defined ballistics. See the BTSJ papers from the 1940s iirc. And the original meter HAD/HAS to have the attenuator incorporated. The use of the meter without recognising the lack of sensitivity led to the actual zero level of +4dBm. The original switched attenuator had zero attenuation in the 4 position. It takes +4dBm to drive the meter to the zero indication. Big topic too. Watch out for web misinformation. You can check my version by referring to the papers. I can not [easily] access all my digital archives at present so I am not providing full details.
11). As David implied... I think you will need a sensitivity adjustment. Users can calibrate the zero to represent whatever they want. The meter characteristic/ballistic HAS to be known by a serious user. There are examples on the web.
Whatever you do has to provide for the whole audio chain.. there has to be headroom everywhere so that there is no clipping.
12). Here is a typical online statement - this one form Audio Technica :-
"Line level refers to the typical level (strength or amplitude) of the audio signal from mixers, signal-processing equipment and other consumer and professional audio gear. There are two line level types: consumer and professional. Consumer line level is generally thought of as a signal whose level is at -10 dBV (0.316). CD players and DVD players are examples of consumer line level equipment. Professional line level is generally thought of as a signal whose level is at +4 dBu (1.23 volts or significantly higher). Signal-processing equipment and professional mixing consoles are examples of professional line level equipment."
Done quickly, apologies for errors and omissions,
John Kaesehagen
Australia
jb-electronics
Thank you very much for your help! I am not trying to build a
professional VU meter; rather, what I want is to create a
microcontroller-based LM3916 alternative. These are now obsolete LED bar
driver ICs with a logarithmic output that are quite handy for building a
simple (not professional) VU meter.
Lots of things to read for me!
Best wishes
Jens
johnk
I wonder what the various guys who have mentioned making audio meters ended up doing?
Even though you want it for a headphone level indicator, it might get used by others for different purposes.
The problem with any averaging meter in today's audio environment is the fact that digital systems clip when overdriven.
I mentioned tape recording [analogue] earlier. Tape has a natural compression effect [I said limiter last time]. It was the lack of realisation of this that caused so much of the argument when digital workstations and digital storage arrived. The tape recording produced more punch because they couldn't afford to operate down in the linear region. [Signal/noise].
A VU meter needs an audio system with 14 to 18dB headroom. Because the meter is slow it does not indicate peaks. Taking voice intelligibility into account [some clipping allowed], voice peaks are generally 4 to 14dB above what is indicated.
A PPM can be used simply - "don't go into the red". The zero [ie start of red section] indicates that clipping is occurring. As I mentioned before 0dBFS should be a little down from the red [big topic]. FS means Full scale in the digital arena - the biggest number that can be represented by your 16 or 24 bits (or whatever).
For your headphone use you could make the red zone be the onset of distortion caused by the headphones overdriving OR it could represent the max safe listening level [assuming you don't have level limited phones already].
Sound card specs used to give good levels info. I just looked at my motherboard handbooks from the last 15 years and they have zilch.
I am mentioning all of this because if you go to the trouble of coding something then you have the opportunity to build in features.
Another reason for going a bit overboard is this: I believe that when shortcuts and liberties are taken with standards it should be with the knowledge of what has been done. It should be decision based - not a result of ignorance.
I am a bit wary of pointing to a wiki article, but you have probably already been there....
https://en.wikipedia.org/wiki/Line_level
An aside: now that we are talking audio, I would love to get an argument going about Watts RMS !
I say this: the watts specified for an amplifier output are intended to be average power and not rms power.
When Watts rms is stated it is NOT intended to be read as that. It is intended to mean "Watts [rms derived"]. There is a lot of difference.
Any takers?
I invite sceptics to do the math or draw a sinewave power waveform and inspect it carefully. Try 1V rms and 1A rms sine wave resistive.
Having looked at it and tipped the top half into the trough [proving what I proposed], go ahead and perform an rms calculation on that power waveform. You might be surprised.
If that didn't get your attention, look at it his way. Even the respected name companies use the W RMS term. So, how can I be right?
John K
Charles MacDonald
> An aside: now that we are talking audio, I would love to get an argument going about Watts RMS !
> I say this: the watts specified for an amplifier output are intended to be average power and not rms power.
I do recall that the merchants is North America got into a bit of a
"Horsepower Race" with Power ratings on consumer audio equipment.
rather than accept a low number the marketing engineers, started doing
things like testing with Pulse inputs to use the electrical capacity of
the Filter caps. even further some only tested One channel, (so the
power supply was idling on the other.) or connecting a big lab supply as
Music is different than sine waves. You started seeing Peak Music Power
ratings.
the FTC in the States started cracking down and said that any power
claims had to be based on teh RMS output (.707 of the peak) using a sine
wave with Specified distortion. AND that the unit had to be conditioned
at almost maximum power (75% if I recall, or Perhaps 90%)for several
minutes before taking the measurement. SOME units sudenly could not
show ANY power as the pre-conditioning was enough to cause them to shut
down or Melt down.
I recall seeing a sears catalogue (here in Canada) which showed a 100W
stereo, with a small text box titled Truth in Audio stating the the unit
was good for 2W under the US FTC method.
So we can blame the US for inventing "RMS Power" - although they were
really trying to do the right thing.
--
Charles MacDonald Stittsville Ontario
cm...@zeusprune.ca Just Beyond the Fringe
No Microsoft Products were used in sending this e-mail.
Tomasz Kowalczyk
W dniu piątek, 29 marca 2019 08:46:37 UTC+1 użytkownik charles napisał:
On 2019-03-28 10:09 p.m., johnk wrote:
the FTC in the States started cracking down and said that any power
claims had to be based on teh RMS output (.707 of the peak) using a sine
wave with Specified distortion. AND that the unit had to be conditioned
at almost maximum power (75% if I recall, or Perhaps 90%)for several
minutes before taking the measurement. SOME units sudenly could not
show ANY power as the pre-conditioning was enough to cause them to shut
down or Melt down.
I like to know maximum constant RMS output, because it tells me a lot about the amplifier and its capability. The peak power or "music power usually" lacks a definition (I can easily imagine a design that would allow for short pulse of higher power, but power supply/amplifier would fry up if it was to deliver this peak power for longer time).
johnk
Well, how do I say this Thomas?
It is NOT to be called Watts RMS !
They left the vital word out – it is Watts [RMS derived]. They left out “derived”.
The RMS volts and RMS amps that you mention when multiplied together produce Watts. Just plain Watts.
These Watts are actually the average power of the power waveform that resulted from your two sinewaves.
Remember too, that the RMS value of the voltage waveform gives the DC voltage that provides the same heating effect. And that is average power.
Gee, I didn’t say that well. I have just spent a while fighting with Win 10 and drivers for CH340 on Arduino clone boards – I haven’t recovered !
You might think that I am nit picking. However you did say this, “RMS power of a sine wave is 0,5 times peak power.”
And the power waveform isn’t really a sinewave in the way we mean it. The values of interest here lie in the area under the curve (notice it is twice the frequency too?]. In a sine wave [like the Voltage one] the areas of interest are the equal sized ‘lobes’ above and below the zero line.
I invite you to draw out the two sinewaves [Volts and Current] and the resulting power waveform and perform an actual root-mean-square calculation on it to prove your statement. [Graphically is more reliable because it shows the workings J ]
Spoiler: you will NOT get 0.5 x pk as the answer.
The VERY rough sketch that I sent Charles shows what I mean about the average value [the green bit tipped over into the trough].
(Rough because I was on a new touch screen laptop and NOT in tablet mode. I was experimenting; made it tricky to draw with the pen. )
I know that you know what you mean when you refer to amplifiers this way, but you could add the extra word and be ‘more right’ J
Thanks for nibbling on the hook. But, I really do wish that someone had been willing to do the graphical maths thing. Someone must want to prove me wrong, surely.
John Kaesehagen
Australia
neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] On Behalf Of Tomasz Kowalczyk
Sent: Tuesday, 16 April 2019 20:30
To: neonixie-l
Subject: Re: [neonixie-l] OT: audio levels
--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/7c2c3bfe-e43d-4701-86c9-f9ecb39e735a%40googlegroups.com.
johnk
Thomas, the audio levels standards do exist. But I agree with you that there isn’t just one.
Each of the areas has its own standard. So, you do have a known standard for consumer electronics. You do have known standards for radio/entertainment, studio and professional work [pre DAWs]. And you do have a ‘new’ standard for recent era digital equipments. That is why I mentioned Bob Katz’ book when Jens asked his question. I had to work with a Dutch mixing desk that attempted to be able to be employed in ‘both’ eras radio and recording studio environments. They didn’t manage it – they threw away a lot of head room. [Scorpius] The metering issues created extra problems because ideally you want the standards there to match the sensitivities too – but there is conflict. And it all depends whether you back the Loudness Wars. Remember too, generally you aren’t trying to make an equipment that is to serve in more than one of the environments.
I say – Long Live Dynamic Range !
John Kaesehagen
Australia
From: neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] On Behalf Of Tomasz Kowalczyk
Sent: Tuesday, 16 April 2019 19:54
To: neonixie-l
Subject: Re: [neonixie-l] OT: audio levels
W dniu piątek, 29 marca 2019 08:46:37 UTC+1 użytkownik charles napisał:
On 2019-03-28 10:09 p.m., johnk wrote:
based on teh RMS output (.707 of the peak)
That would be 0,707 (1/sqrt(2)) voltage output on a resistive load. RMS power of a sine wave is 0,5 times peak power.
I like to know maximum constant RMS output, because it tells me a lot about the amplifier and its capability. The peak power or music power usually lacks a definition (I can easily imagine a design that would allow for short pulse of higher power, but power supply/amplifier would fry up if it was to deliver this peak power for longer time).
It is well known that for most time amplifier delivers a little percentage of its maximum power, but having continous Watts RMS into specified load rating is a very solid rating with solid definition.
On the topic: it really bothers me that there is no definition of one voltage standard. It would make perfect sense to create a standard of a CD output, for example - a 2Vpp signal, in which +1V would correspond to maximum digital value DAC can give (65535 for standard 16-bit) and -1V would correspond to a value of 0. It would make designing amplifiers much easier - the sensivity would be always same.
Currently I'm building a vacuum tube amplifier and I'm mad at the fact that I need to make the sensivity on 200mVpp level (my phone output), but most of other sources will have much higher signal. I think I'll need to go with a noisy method of applying extra resistor in series with volume potentiometer for the higher inputs... or apply the resistor and a "+20dB" switch that will short it.
At least it is a vacuum tube amplifier, so noise will be an issue anyway.
--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/6418d010-db45-4c62-8f64-12bdc99a5520%40googlegroups.com.
johnk
Gee Whizz. Sorry for not spelling your name correctly in the two posts Tomasz.
That was very careless and rude of me – please accept my apologies.
John K
From: neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] On Behalf Of Tomasz Kowalczyk
Sent: Tuesday, 16 April 2019 20:30
To: neonixie-l
Subject: Re: [neonixie-l] OT: audio levels
--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/7c2c3bfe-e43d-4701-86c9-f9ecb39e735a%40googlegroups.com.
Tomasz Kowalczyk
W dniu wtorek, 16 kwietnia 2019 15:57:35 UTC+2 użytkownik johnk napisał:
Well, how do I say this Thomas?
It is NOT to be called Watts RMS !
They left the vital word out – it is Watts [RMS derived]. They left out “derived”.
The RMS volts and RMS amps that you mention when multiplied together produce Watts. Just plain Watts.
These Watts are actually the average power of the power waveform that resulted from your two sinewaves.
Remember too, that the RMS value of the voltage waveform gives the DC voltage that provides the same heating effect. And that is average power.
Gee, I didn’t say that well. I have just spent a while fighting with Win 10 and drivers for CH340 on Arduino clone boards – I haven’t recovered !
You might think that I am nit picking. However you did say this, “RMS power of a sine wave is 0,5 times peak power.”
And the power waveform isn’t really a sinewave in the way we mean it. The values of interest here lie in the area under the curve (notice it is twice the frequency too?]. In a sine wave [like the Voltage one] the areas of interest are the equal sized ‘lobes’ above and below the zero line.
I invite you to draw out the two sinewaves [Volts and Current] and the resulting power waveform and perform an actual root-mean-square calculation on it to prove your statement. [Graphically is more reliable because it shows the workings J ]
Spoiler: you will NOT get 0.5 x pk as the answer.
The VERY rough sketch that I sent Charles shows what I mean about the average value [the green bit tipped over into the trough].
(Rough because I was on a new touch screen laptop and NOT in tablet mode. I was experimenting; made it tricky to draw with the pen. )
I know that you know what you mean when you refer to amplifiers this way, but you could add the extra word and be ‘more right’ J
Thanks for nibbling on the hook. But, I really do wish that someone had been willing to do the graphical maths thing. Someone must want to prove me wrong, surely.
John Kaesehagen
Australia
I've ran a little excel excercise - I calculated voltage and current with peak value equal to sqrt(2) for 360 points of a sine wave. If I run RMS calculation (from definition) on voltage or current, it is equal to 1 (so one times the other is 1W), but running it on voltage*current gives a value of about 1,22W RMS.
I admit this is a mistake to use W RMS.
I'm currently building a vacuum tube amplifier, which works greatly with uncompressed tracks - it compresses them itself via soft clipping! That's the reason 15-20W tube amp can sound like 50W solid state, vocals and most instruments can stay at the same SPL, but vacuum tube amp will compress the short spikes from percussion or other instrument instead of grossly distorting them. Solid state needs that power headroom to prevent going into clipping, because you can easily hear it and it isn't pleasurable.
Of course, I could also just build a simple Gainclone amp with higher output power for more headroom... but high voltage is fun!
johnk
Thank you so much for playing the game Tomasz, and bothering to check the calculations.
It can also be done with DC a meter or scope that shows True RMS when DC is present. [Check by reading a battery on the AC range.]
You can set up a sine wave that is offset by a DC value. You make a voltage waveform that is the same shape as the power waveform that we are talking about. Then you measure it.
There is a useful electronics website.
https://masteringelectronicsdesign.com/how-to-derive-the-rms-value-of-a-sine-wave-with-a-dc-offset/
On that page he shows the mathematics involved. Notice that a very simple formula right at the end/bottom covers what we are doing.
It gives an answer very close to yours too.
Regarding audio and perceived loudness: if you ever want to increase that impact of the sound without “cheating” by using a compressor, then just record to magnetic tape. You could even use a deck that has replay monitor heads immediately after the record heads so that it is close to real time.
To get up out of the noise they had to set up magnetic to work in a rather non-linear region. It automatically compressed. This was one of the things that made full digital audio sound different. They didn’t set out to compress “old” audio but it happened anyway.
That consistent loudness thing really annoys me too. Unfortunately some of the normalising programs cause other problems. Some of the problem is caused by people using the scope view in their digital software. When you have 20dB headroom it looks like you are working so far down that you must be in the noise. They also think that you are wasting that space. This is where metering methods become so important too. And some training/awareness.
Thanks to all the Nixie guys for tolerating this OT discussion too.
Regards,
John K
From: neoni...@googlegroups.com [mailto:neoni...@googlegroups.com] On Behalf Of Tomasz Kowalczyk
Sent: Thursday, 18 April 2019 04:04
To: neonixie-l
Subject: Re: [neonixie-l] OT: audio levels
--
You received this message because you are subscribed to the Google Groups "neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neonixie-l+...@googlegroups.com.
To post to this group, send email to neoni...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/029027cb-4ef1-47cd-b915-cdbaf837f5ca%40googlegroups.com.