Misleading Measurements

[lucretius]

6 hours ago, March Audio said:

Just on point 1, its not quite correct, sorry :).  The waveform shape makes no difference as it is the peak voltage level that is 0dBFS.  With a sinewave the RMS level will be - 3.01dB

 

But doesn't the standard say to set this max RMS level as the reference -- 0 dbfs?  That is to say, the 0 dB reference for either peak OR RMS measurement is that of a sinewave at full scale.  (And isn't the case that waveform shape does make a difference when it comes to computing RMS?)  Seems to be a lot of discussion about this on the net, mostly confusing.

 

 

 

 

[lucretius]

3 hours ago, bluesman said:

We need to be honest and clear when discussing things based on a well defined standard.  First, #1 is not your observation - you're quoting definition 3.4 in the very AES standard under discussion.  Here's a link to the actual document, so everyone can read it.

 

Actually, I looked a a bunch of documents but I didn't have access to the AES standard. Thanks for the link.

 

3 hours ago, bluesman said:

Second, while it's true that a square wave contains more energy than a sine wave of the same frequency and amplitude, the standard specifically tells us in the ssentence after the one you offered as your observation that only a sine wave is to be used, specifically:  "Square-wave signals at this level are not recommended because tilt or overshoot introduced by any filtering operations will cause clipping of the signal".

 

Yes.  Only a sine wave is to be used for setting the zero reference.  But music is not sine waves. So the point of bringing up the square waves is to show that the meter could read higher than 0 (using an RMS meter?).

 

3 hours ago, bluesman said:

Third, "full scale amplitude" is defined clearly as the "amplitude of a 997-Hz sine wave whose positive peak value reaches the positive digital full scale, leaving the negative maximum code unused". It isnnot defined as the RMS value of that signal - it's defined on the basis of peak value.

 

Yes.  Then the standard also says this:

 

5.4

Input for full-scale amplitude

 

     NOTE The characteristic to be specified is the analog signal voltage required to reach digital clipping under normal device settings.

 

In systems where the output is accessible in the digital domain, the input for full-scale amplitude shall be the r.m.s. voltage of a 997-Hz sine wave that shall be applied to the input to obtain a digital signal whose positive peak value reaches the positive digital full scale.

 

and this:

 

6.3 Output amplitude at full scale

 

In systems where the input is accessible in the digital domain, the output amplitude at full scale shall be the r.m.s. voltage that results from a sine wave whose positive peak value reaches the positive digital full scale under normal settings of gain controls.

**************************************

 

Color me confused.  Perhaps someone can explain to me what is meant in 5.4 and 6.3 above.

 

[stereo coffee]

4 hours ago, March Audio said:

Yes, but (sorry) as discussed at length earlier in the thread there are no domestic dacs that output at -10dBV.  Most are close to 2 v rms at 0dBFS. This is +6dBV. (RCA output)

Correct, and incorrect ,  the players certainly have that capability,  but the media is held back deliberately presently to sensibly allow dynamics in music we buy. This has nothing to do with how much level you CAN then add at your DAC or whatever, but is simply the level chosen by the music industry so that music we buy is music and not distorted audio. 

 

It is described here:

https://en.wikipedia.org/wiki/Loudness_war

 

The loudness war (or loudness race) is a trend of increasing audio levels in recorded music, which reduces audio fidelity and — according to many critics — listener enjoyment.

 

Companding is the future 

The larger picture is if you want to have greater dynamics without distortion , you need to use companding to do so, just as recording studios have done to preserve ... let me repeat that ... preserve  dynamics since 1965 - Dolby A - and always one step ahead DBX.  The DBX type 4 White paper gives good overview of the history of companding  + improvements that can be made to digital recording.   https://warehousesound.com/dbxtiv.htm

 

Media we buy would in theory contain the ORIGINAL companding used during recording, and would play back expanded the same audio - companding being a two part process.  It is though possible to enjoy companding real time, to experience this connect a late 1980's early 1990's DBX 150x to a 16 bit  CD player and compare it to a 20 bit player - you should find the 16 bit player is every part ( i was going to say bit )  as good as any given 20 bit CD player.  

 

 

 

 

 

 

 

 

 

[bluesman]

2 hours ago, lucretius said:

Color me confused.  Perhaps someone can explain to me what is meant in 5.4 and 6.3 above.

The entire document and its intent are confusing.  It refers to an analog input driving a digital device, which is already a bizarre and conflicting circumstance for a document that purports to be for the "measurement of digital audio equipment".  And it describes the input signal as "... the maximum analog signal that may be applied to the device for correct operation".   Pure digital devices do not have analog inputs, but the document specifies analog signal generators.  So there are ADCs and DACs in the mix to provide a digital input and an output signal with measurable voltage.  Remember that this document is 23 years old and was last revised in 2004.  Also remember that it is not a true standrad - it's neither enforceable nor used consistently throughout the industry.

 

Pure digital domains exist in which to generate, manipulate, and measure digital audio signals, e.g. MATLAB.  I don't know the current status of digital audio standards regarding pure digital platforms and devices, but MATLAB and similar schema really should be the platform on which future digital audio standards are developed and measured.  In this setting, 0 DB FS would truly be the maximum amplitude digital signal handled by the equipment under test (EUT).  It would not be measured in volts - it would be measured in digital code.

 

For example, in a 16 bit integer domain, digital signal value in dB FS = 20*log10(abs(value)/32768) because 16bit signed has values between -32768 and +32767, e.g. 2^15=32768 expressed in MATLAB code.

[pkane2001]

4 hours ago, lucretius said:

5.4

Input for full-scale amplitude

 

     NOTE The characteristic to be specified is the analog signal voltage required to reach digital clipping under normal device settings.

 

In systems where the output is accessible in the digital domain, the input for full-scale amplitude shall be the r.m.s. voltage of a 997-Hz sine wave that shall be applied to the input to obtain a digital signal whose positive peak value reaches the positive digital full scale.

 

and this:

 

6.3 Output amplitude at full scale

 

In systems where the input is accessible in the digital domain, the output amplitude at full scale shall be the r.m.s. voltage that results from a sine wave whose positive peak value reaches the positive digital full scale under normal settings of gain controls.

**************************************

 

Color me confused.  Perhaps someone can explain to me what is meant in 5.4 and 6.3 above.

 

 

Didn't read the document, but from your quote, it's just saying that analog input/output required to produce a full-scale signal (at input or output) is to be reported as the RMS voltage of a sine wave, rather than its peak value. For example, when 0dBFS is reported to require 2v input, this means 2v RMS, not peak. The actual voltage that produces 0dBFS is then the peak value of about 2.8v.

 

[March Audio]

4 hours ago, stereo coffee said:

Correct, and incorrect ,  the players certainly have that capability,  but the media is held back deliberately presently to sensibly allow dynamics in music we buy. This has nothing to do with how much level you CAN then add at your DAC or whatever, but is simply the level chosen by the music industry so that music we buy is music and not distorted audio. 

 

It is described here:

https://en.wikipedia.org/wiki/Loudness_war

 

The loudness war (or loudness race) is a trend of increasing audio levels in recorded music, which reduces audio fidelity and — according to many critics — listener enjoyment.

 

Companding is the future 

The larger picture is if you want to have greater dynamics without distortion , you need to use companding to do so, just as recording studios have done to preserve ... let me repeat that ... preserve  dynamics since 1965 - Dolby A - and always one step ahead DBX.  The DBX type 4 White paper gives good overview of the history of companding  + improvements that can be made to digital recording.   https://warehousesound.com/dbxtiv.htm

 

Media we buy would in theory contain the ORIGINAL companding used during recording, and would play back expanded the same audio - companding being a two part process.  It is though possible to enjoy companding real time, to experience this connect a late 1980's early 1990's DBX 150x to a 16 bit  CD player and compare it to a 20 bit player - you should find the 16 bit player is every part ( i was going to say bit )  as good as any given 20 bit CD player.  

 

 

 

 

 

 

 

 

 

Sorry  but again this is all incorrect.

 

In an earlier post I asked if you could provide a list of albums for which I would post the amplitude statistics.  I wanted you to specify the albums so I couldn't be accused of being selective.

 

Most albums are normalised close to 0dBFS to maximise the signal to noise ratio.

 

I will post some examples later.

[March Audio]

5 hours ago, lucretius said:

 

But doesn't the standard say to set this max RMS level as the reference -- 0 dbfs?  That is to say, the 0 dB reference for either peak OR RMS measurement is that of a sinewave at full scale.  (And isn't the case that waveform shape does make a difference when it comes to computing RMS?)  Seems to be a lot of discussion about this on the net, mostly confusing.

 

 

 

 

Yes 2 volts rms.  However the peak voltage value of 2 volts rms is 2.8 volts

 

It is the peak voltage we have to accommodate.

[pkane2001]

1 hour ago, bluesman said:

The entire document and its intent are confusing.  It refers to an analog input driving a digital device, which is already a bizarre and conflicting circumstance for a document that purports to be for the "measurement of digital audio equipment".  And it describes the input signal as "... the maximum analog signal that may be applied to the device for correct operation". 

 

It is clear that the document is referring to devices that have at least an analog input or an analog output, including devices accepting or producing a digital representation of the signal. In other words, analog devices, as well as, devices containing ADCs and/or DACs. 

[March Audio]

1 hour ago, bluesman said:

The entire document and its intent are confusing.  It refers to an analog input driving a digital device, which is already a bizarre and conflicting circumstance for a document that purports to be for the "measurement of digital audio equipment".  And it describes the input signal as "... the maximum analog signal that may be applied to the device for correct operation".   Pure digital devices do not have analog inputs, but the document specifies analog signal generators.  So there are ADCs and DACs in the mix to provide a digital input and an output signal with measurable voltage.  Remember that this document is 23 years old and was last revised in 2004.  Also remember that it is not a true standrad - it's neither enforceable nor used consistently throughout the industry.

 

Pure digital domains exist in which to generate, manipulate, and measure digital audio signals, e.g. MATLAB.  I don't know the current status of digital audio standards regarding pure digital platforms and devices, but MATLAB and similar schema really should be the platform on which future digital audio standards are developed and measured.  In this setting, 0 DB FS would truly be the maximum amplitude digital signal handled by the equipment under test (EUT).  It would not be measured in volts - it would be measured in digital code.

 

For example, in a 16 bit integer domain, digital signal value in dB FS = 20*log10(abs(value)/32768) because 16bit signed has values between -32768 and +32767, e.g. 2^15=32768 expressed in MATLAB code.

This whole conversation is about how the digital data interfaces with the analogue world so ADC and DAC are an intrinsic part of the conversation.

[stereo coffee]

It is explained here in terms of streaming services, that levels are backed off to allow for dynamics in music. 

 

 

 

[danadam]

9 hours ago, March Audio said:

there are no domestic dacs that output at -10dBV.  Most are close to 2 v rms at 0dBFS. This is +6dBV. (RCA output)

My take on this: the -10 dBV is a line level, which is a different thing than max level. If wikipedia is to be believed, line level is:

Quote

A line level describes a line's nominal signal level ...

and following to nominal level:

Quote

The nominal level is the level that these devices were designed to operate at, for best dynamic range and adequate headroom.

[...] the headroom as the difference between nominal and maximum output.

If we assume a headroom of 16 dB, then -10 dBV line level (i.e. nominal level) + 16 dB headroom = +6 dBV = 2 V max level. So in my understanding domestic DACs have both -10 dBV line level and 2 V max level.

 

But to be clear, I'm not saying that they should be tested at -10 dBV instead of 2 V or that the music content never reaches near those 2 V :)

 

[March Audio]

28 minutes ago, danadam said:

My take on this: the -10 dBV is a line level, which is a different thing than max level. If wikipedia is to be believed, line level is:

and following to nominal level:

If we assume a headroom of 16 dB, then -10 dBV line level (i.e. nominal level) + 16 dB headroom = +6 dBV = 2 V max level. So in my understanding domestic DACs have both -10 dBV line level and 2 V max level.

 

But to be clear, I'm not saying that they should be tested at -10 dBV instead of 2 V or that the music content never reaches near those 2 V :)

 

There is no nominal level, it has to be related to a specified digital level and specified signal.

[March Audio]

6 hours ago, stereo coffee said:

Correct, and incorrect ,  the players certainly have that capability,  but the media is held back deliberately presently to sensibly allow dynamics in music we buy. This has nothing to do with how much level you CAN then add at your DAC or whatever, but is simply the level chosen by the music industry so that music we buy is music and not distorted audio. 

 

It is described here:

https://en.wikipedia.org/wiki/Loudness_war

 

The loudness war (or loudness race) is a trend of increasing audio levels in recorded music, which reduces audio fidelity and — according to many critics — listener enjoyment.

 

Companding is the future 

The larger picture is if you want to have greater dynamics without distortion , you need to use companding to do so, just as recording studios have done to preserve ... let me repeat that ... preserve  dynamics since 1965 - Dolby A - and always one step ahead DBX.  The DBX type 4 White paper gives good overview of the history of companding  + improvements that can be made to digital recording.   https://warehousesound.com/dbxtiv.htm

 

Media we buy would in theory contain the ORIGINAL companding used during recording, and would play back expanded the same audio - companding being a two part process.  It is though possible to enjoy companding real time, to experience this connect a late 1980's early 1990's DBX 150x to a 16 bit  CD player and compare it to a 20 bit player - you should find the 16 bit player is every part ( i was going to say bit )  as good as any given 20 bit CD player.  

 

 

As I mentioned here are a few examples of the signal levels on various albums.  All are at or coming very close to 0dBFS.  This means on a typical dac the RCA peak output voltage will be getting close to 2.8 Volts.

 

Bob Marley - Baylon By Bus

 

The audio data is usually normalised to come close to 0dB to maximise signal to noise ratio.

 

As previously mentioned this should really be looking at the true peak level which accounts for inter sample overs and can be higher than the actual sample value.  Unfortunately it seems that many mastering engineers dont understand this and ram it right up to 0dB.  To be safe you really should back off to maybe around -3dB to ensure there are very few clipped samples.

 

Bill Evans Live at the Montreau jazz Festival

 

Daft Punk Random Access Memories

 

Copland fanfare Fort The Common Man

 

 

[March Audio]

1 hour ago, stereo coffee said:

It is explained here in terms of streaming services, that levels are backed off to allow for dynamics in music. 

 

 

 

Sorry this is wrong.  As I keep demonstrating to you the signal levels are normalised to peak close to 0dB

 

You have got to stop thinking about this in terms of music.  Music cannot be used to quantify signal levels.  You have to use consistent signals of known value.

 

A quiet piano concerto will have very different RMS level to a highly compressed death metal track.  They will boith peak close to 0dBFS.

 

 

Just to add to this you may choose to adjust signal levels to make different tracks/albums etc have a similar perceived volume level, but this is not relevant to the discussion here.

[bluesman]

2 hours ago, March Audio said:

This whole conversation is about how the digital data interfaces with the analogue world so ADC and DAC are an intrinsic part of the conversation.

...which is exactly why AES 17-1998 is confusing many in this thread. Although entitled “measurement of digital audio equipment”, it is based entirely on measuring analog parameters at the ADC input and DAC output.  The amplitude of a pure digital signal is determined by its coded content, which is read rather than measured.  A 0 dB FS digital signal in a digital domain is defined as the highest level achievable by the EUT.  For a 16 bit signal that would be the coded word 32768, which is what constitutes the signal that enters the DAC stage. A steady state sine wave or music waveform at 0dB FS would be a chain of signal samples along with the coded value for maximum signal strength.

 

This will be translated into an analog signal by a DAC in the course of making sound from it. But the voltage of that signal will be determined by the design of the DAC, not by an AES standard and not by the digital values in the signal.  It is usually in the range bandied about in this portion of the thread.  But it differs from device to device by design, and the range of usual values is higher for commercial equipment than for most consumer devices.

[March Audio]

21 minutes ago, bluesman said:

...which is exactly why AES 17-1998 is confusing many in this thread. Although entitled “measurement of digital audio equipment”, it is based entirely on measuring analog parameters at the ADC input and DAC output.  The amplitude of a pure digital signal is determined by its coded content, which is read rather than measured.  A 0 dB FS digital signal in a digital domain is defined as the highest level achievable by the EUT.  For a 16 bit signal that would be the coded word 32768.

This will be translated into an analog signal by a DAC in the course of making sound from it. But the voltage of that signal will be determined by the design of the DAC, not by an AES standard.  It is usually in the range bandied about in this portion of the thread.  But it differs from device to device by design, and the range of usual values is higher for commercial equipment than for most consumer devices.

No it's not, its relating digital levels to analogue levels.

 

Also its not a standard.  It's a method that has no formal implementation or AFAIK in recognised formal international standards bodies.

 

As a method it is correct. The dac or adc is actually irrelevant.  The method is applicable to all.  It doesn't specify analogue voltage levels.

 

It's implicit that an audio has to be converted to and from the digital domain so I'm really not sure what your issue with it is.

 

 

[bluesman]

48 minutes ago, March Audio said:

It's implicit that an audio has to be converted to and from the digital domain so I'm really not sure what your issue with it is.

You really don’t understand this.   0dBFS (Full Scale) is the clipping point for a signal in a digital audio product. Rather than measuring from the noise floor up, digital signals are measured (or referenced) from full scale down. A 0dB FS (Full Scale) signal contains the maximum amount of digital information that can be used to represent the signal being defined.  
 

The output of a DAC driven with a 0dB FS signal should be at the full electrical potential of the device - this is the voltage drop across the load presented by the next stage’s input.  There is no standard for this value - it is determined by the design of the DAC and it is independent of the digital signal, whose “amplitude” is determined by the digital code it contains.  The same 0 dB FS signal from a digital mixer will drive different DACs to different voltage levels.

 

I’ll say that again - the same 0 dB FS digital signal will produce different output signal levels from different DACs. The digital signal level is proportional to but otherwise entirely independent of the amplitude of the analog signal generated by a DAC.  It is the design of the DAC that determines the voltage of the output it creates from the instructions coded in the digital signal at its input.

[March Audio]

11 minutes ago, bluesman said:

You really don’t understand this.   0dBFS (Full Scale) is the clipping point for a signal in a digital audio product. Rather than measuring from the noise floor up, digital signals are measured (or referenced) from full scale down. A 0dB FS (Full Scale) signal contains the maximum amount of digital information that can be used to represent the signal being defined.  
 

The output of a DAC driven with a 0dB FS signal should be at the full electrical potential of the device - this is the voltage drop across the load presented by the next stage’s input.  There is no standard for this value - it is determined by the design of the DAC and it is independent of the digital signal, whose “amplitude” is determined by the digital code it contains.  The same 0 dB FS signal from a digital mixer will drive different DACs to different voltage levels.

 

I’ll say that again - the same 0 dB FS digital signal will produce different output signal levels from different DACs. The digital signal level is proportional to but otherwise entirely independent of the amplitude of the analog signal generated by a DAC.  It is the design of the DAC that determines the voltage of the output it creates from the instructions coded in the digital signal at its input.

We have been round this already.  It is your misunderstanding.

 

You seem unable to separate the concepts of relating one measurand to another, and one being the other.

 

 

6.3 Output amplitude at full scale

In systems where the input is accessible in the digital domain, the output amplitude at full scale shall be the
r.m.s. voltage that results from a sine wave whose positive peak value reaches the positive digital full scale
under normal settings of gain controls.

 

The design or actual output voltages of the DAC are irrelevent to the statement above.

 

 

[bluesman]

37 minutes ago, March Audio said:

In systems where the input is accessible in the digital domain, the output amplitude at full scale shall be the
r.m.s. voltage that results from a sine wave whose positive peak value reaches the positive digital full scale
under normal settings of gain controls.

But that output amplitude is NOT the same value for every DAC that reads the same 0 dB FS digital signal - there is no specific "equivalent" full scale analog output voltage for a 0 dB FS digital signal because there is no consistent mathematical relationship of any kind between the digital signal being read by the DAC and the analog signal it creates from those digital instructions excepet for the spectral content. Every DAC design outputs the same analog waveform (+/- whatever distortions it intriduces) but at a level determined ONLY by its own design.

 

Whether you think that the analog output signal's peak amplitude is mathematically related to the digital signal or that the analog output signal is the digital signal in a different form (and I can't figure out which of those you actually do believe) doesn't matter because neither is correct.

 

 

[March Audio]

2 hours ago, bluesman said:

But that output amplitude is NOT the same value for every DAC that reads the same 0 dB FS digital signal - there is no specific "equivalent" full scale analog output voltage for a 0 dB FS digital signal because there is no consistent mathematical relationship of any kind between the digital signal being read by the DAC and the analog signal it creates from those digital instructions excepet for the spectral content. Every DAC design outputs the same analog waveform (+/- whatever distortions it intriduces) but at a level determined ONLY by its own design.

 

Whether you think that the analog output signal's peak amplitude is mathematically related to the digital signal or that the analog output signal is the digital signal in a different form (and I can't figure out which of those you actually do believe) doesn't matter because neither is correct.

 

 

 

The AES document *does not* say there is.  It is just relating one to the other.  It relates the analogue voltage peak (whatever that might be) to the digital full scale level.

 

It says nothing about what voltage levels.  It doesnt need to.  The statement is applicable to any system and any voltage.  It is just defining the relationship.

 

You need to re-read what is says.

 

6.3 Output amplitude at full scale

In systems where the input is accessible in the digital domain, the output amplitude at full scale shall be the
r.m.s. voltage that results from a sine wave whose positive peak value reaches the positive digital full scale
under normal settings of gain controls.

 

 

[lucretius]

7 hours ago, pkane2001 said:

 

Didn't read the document, but from your quote, it's just saying that analog input/output required to produce a full-scale signal (at input or output) is to be reported as the RMS voltage of a sine wave, rather than its peak value. For example, when 0dBFS is reported to require 2v input, this means 2v RMS, not peak. The actual voltage that produces 0dBFS is then the peak value of about 2.8v.

 

 

That's how I read it.  However, if we set 2v RMS* = 0 dB FS, then the peak value is 2.8v only for the 1k sine wave, Music can actually produce a greater peak value than 2.8v, meaning you could have peak values greater than 0 dB on the meter (which of course is measuring the analog peak).

 

*assuming this is max RMS

 

[March Audio]

1 hour ago, lucretius said:

 

That's how I read it.  However, if we set 2v RMS* = 0 dB FS, then the peak value is 2.8v only for the 1k sine wave, Music can actually produce a greater peak value than 2.8v, meaning you could have peak values greater than 0 dB on the meter (which of course is measuring the analog peak).

 

*assuming this is max RMS

 

Its better not to think of it in relation to the RMS of a sine wave.  Just accept the peak value cannot go past 0dBFS which is the highest digital level (+32768 in 16 bit system).  There are no more numbers to describe the input value so it simply cant be higher.

 

Yes a different signal to a sine may have a higher or lower crest factor, but this just means you will have a higher or lower RMS value. A sine wave rms value is 0.707 of the peak, but for a triangle wave the rms value is 0.577 of its peak.

Whatever the peak signal level still cant go past 2.8v, it will just clip. 

 

To look at the statement again:

 

the output amplitude at full scale shall be the r.m.s. voltage that results from a sine wave whose positive peak value reaches the positive digital full scale

 

Its saying that the analogue full scale voltage in RMS is defined by the peak level of a sine wave equalling the maximum digital level.

[pkane2001]

5 hours ago, lucretius said:

 

That's how I read it.  However, if we set 2v RMS* = 0 dB FS, then the peak value is 2.8v only for the 1k sine wave, Music can actually produce a greater peak value than 2.8v, meaning you could have peak values greater than 0 dB on the meter (which of course is measuring the analog peak).

 

*assuming this is max RMS

 

 

As @March Audio said, the RMS value varies with the waveform. Peak value corresponding to 0dBFS is the maximum output.

[lucretius]

4 minutes ago, pkane2001 said:

 

As @March Audio said, the RMS value varies with the waveform. Peak value corresponding to 0dBFS is the maximum output.

 

Actually. it's peak value corresponding to full scale sine wave.

[lucretius]

5 hours ago, March Audio said:

Its better not to think of it in relation to the RMS of a sine wave.  Just accept the peak value cannot go past 0dBFS which is the highest digital level (+32768 in 16 bit system).  There are no more numbers to describe the input value so it simply cant be higher.

 

Yes a different signal to a sine may have a higher or lower crest factor, but this just means you will have a higher or lower RMS value. A sine wave rms value is 0.707 of the peak, but for a triangle wave the rms value is 0.577 of its peak.

Whatever the peak signal level still cant go past 2.8v, it will just clip. 

 

To look at the statement again:

 

the output amplitude at full scale shall be the r.m.s. voltage that results from a sine wave whose positive peak value reaches the positive digital full scale

 

Its saying that the analogue full scale voltage in RMS is defined by the peak level of a sine wave equalling the maximum digital level.

 

Thanks.  I think I am beginning to see it. For 2 vrms (sine wave) which = 2.8 v peak, if we set that to  0 dB FS, then for a square wave, the rms = peak = 2.8 v.  We can never have a peak higher than 2.8v.  However, rms can vary from 2.8 v and downwards for different waveforms. Is this correct?  And do we zero the rms meter? (It would seem that we would need to zero it at 2.8 -- this would mean that both peak meter and rms meter get zeroed at 2.8).