Dear all,

Can someone please confirm the difference between amplitude and power?  
I understand that doubling the amplitude gives 6 db increase and  
doubling the power or intensity gives 3 db increase.

Power from my understanding is the measure of the maximum amplitude  
within a unit of time, is that correct?

And amplitude is a measure of peak or rms?

Best,

P



Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

There are more subtleties in this question than can be covered in a 
short answer. This page gives a good overview of the subject, doing the 
important task of relating it all primarily to physical properties, and 
extending it to electrical domains later on:


http://www.indiana.edu/~emusic/acoustics/amplitude.htm

Unfortunately, the fact that we can contrive to obtain both amplitude 
and power values for a single sample ("instantaneous") can easily trip 
us up. We use the dB value  directly with reference to a pair of 
amplitudes (i.e it is a ratio, or comparison), and as that page 
indicates, we need to understand the distinction between intensity and 
amplitude. In digital audio we tend to compare amplitudes, so that a 
doubling of power is still 6dB. RMS is yet another related story, 
covered at the top of that page, and is a reminder that most talk about 
amplitude is about the average amplitude over time, rather than the 
instantaneous value of a waveform at one point. RMS is a practical way 
to express that average, given that typically a signal is symmetrically 
bipolar, and taking a literal average will give a zero result.

The bottom line is that to be properly precise, we always have to define 
our terms completely, whether "peak amplitude" or "rms amplitude" or 
"rms power" etc; they all come in handy in particular contexts. We can 
often get away (most of the time!) just referring to "amplitude" or 
"power", because we are within a known domain of digital audio, but 
properly we should be a little more precise about these things. The 
decibel especially, being a raw ratio between [any] two values, exists 
in a wide range of forms, of which the 0dBFS definition is but one.


Richard Dobson



On 12/09/2010 09:42, Peiman Khosravi wrote:
> Dear all,
>
> Can someone please confirm the difference between amplitude and power? I
> understand that doubling the amplitude gives 6 db increase and doubling
> the power or intensity gives 3 db increase.
>
> Power from my understanding is the measure of the maximum amplitude
> within a unit of time, is that correct?
>
> And amplitude is a measure of peak or rms?
>
> Best,
>
> P
>
>
>
> Send bugs reports to the Sourceforge bug tracker
> https://sourceforge.net/tracker/?group_id=81968&atid=564599
> Discussions of bugs and features can be posted here
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>
>




Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

Arggh, sent mid-edit! I meant to write:

In digital audio we tend to compare amplitudes, so that a doubling of 
amplitude is 6dB, while a doubling of power is still 3dB.

And I meant to add that one example of this is the "hole in the middle" 
effect of a linear stereo pan.

On 12/09/2010 11:23, Richard Dobson wrote:


> amplitude. In digital audio we tend to compare amplitudes, so that a
> doubling of power is still 6dB.




Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

Thanks very much indeed Richard. That's a very clear explanation.

Best,

Peiman

On 12 Sep 2010, at 12:18, Richard Dobson wrote:

> Arggh, sent mid-edit! I meant to write:
>
> In digital audio we tend to compare amplitudes, so that a doubling  
> of amplitude is 6dB, while a doubling of power is still 3dB.
>
> And I meant to add that one example of this is the "hole in the  
> middle" effect of a linear stereo pan.
>
> On 12/09/2010 11:23, Richard Dobson wrote:
>
>
>> amplitude. In digital audio we tend to compare amplitudes, so that a
>> doubling of power is still 6dB.
>
>
>
>
> Send bugs reports to the Sourceforge bug tracker
>           https://sourceforge.net/tracker/?group_id=81968&atid=564599
> Discussions of bugs and features can be posted here
> To unsubscribe, send email sympa@lists.bath.ac.uk with body  
> "unsubscribe csound"
>



Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

Ah, yes, that's the proper explanation of the "hole in the middle", thanks :-)

On a somewhat related subject: Do you also have a hint on explaining
equal power crossfading between two sources ?
If one crossfades between two identical signals (yes, why would one do
that, exept for testing...) we get a constant perceived loudness when
using a linear crossfading curve. But when crossfading between two
different signals (as we would normally do), I find that the same
curve as commonly used for equal power panning (e.g. sqrt(kpan)) does
make it sound like the output maintains the same power during the
crossfade. This of course creates a digital signal with a higher
amplitude than each of the input signals (approx. 1.4 the original amp
midway in the crossfade).

best
Oeyvind

2010/9/12 Richard Dobson :
> Arggh, sent mid-edit! I meant to write:
>
> In digital audio we tend to compare amplitudes, so that a doubling of
> amplitude is 6dB, while a doubling of power is still 3dB.
>
> And I meant to add that one example of this is the "hole in the middle"
> effect of a linear stereo pan.
>
> On 12/09/2010 11:23, Richard Dobson wrote:
>
>
>> amplitude. In digital audio we tend to compare amplitudes, so that a
>> doubling of power is still 6dB.
>
>
>
>
> Send bugs reports to the Sourceforge bug tracker
>           https://sourceforge.net/tracker/?group_id=81968&atid=564599
> Discussions of bugs and features can be posted here
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>
>


Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

Hmm, well I think the short answer is that the constpower pan is 
specific to the situation of two loudspeakers - in making a phantom 
image the two acoustic sources sum as powers, not as amplitudes (if that 
is the right way of putting it). I have also heard explanations in the 
form: ten singers singing a passage are not ten times as loud as one singer.

As it happens, non-linear mono crossfades are far from unusual - very 
common options in samplers defining loop points. The case of "identical" 
signals is a little moot, as ostensibly if you happen to combine them 
out of phase, cancellation effects can be significant. When doing a 
"morphing" kind of crossfade,  my first instinct is to do a const-power 
mix. Certainly in implementing a wet/dry mix control in a plugin, I 
found const-power sounded the smoothest. But the whole thing is full of 
nuances, not least the vexed question of the non-linear response of the 
ear. To merge sounds with peak energy in different frequency bands, 
almost by definition, short of getting abstrusely technical about it 
all, you just find the law that sounds the most "right". As a developer 
of course I can pass the buck and offer the user a choice - then I am no 
longer responsible for the decision. Which is probably why good 
mixing/mastering engineers do in the end earn their fee.

Richard Dobson

On 12/09/2010 21:52, Oeyvind Brandtsegg wrote:
> Ah, yes, that's the proper explanation of the "hole in the middle", thanks :-)
>
> On a somewhat related subject: Do you also have a hint on explaining
> equal power crossfading between two sources ?
> If one crossfades between two identical signals (yes, why would one do
> that, exept for testing...) we get a constant perceived loudness when
> using a linear crossfading curve. But when crossfading between two
> different signals (as we would normally do), I find that the same
> curve as commonly used for equal power panning (e.g. sqrt(kpan)) does
> make it sound like the output maintains the same power during the
> crossfade. This of course creates a digital signal with a higher
> amplitude than each of the input signals (approx. 1.4 the original amp
> midway in the crossfade).
>



Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

Thanks a lot for the input on this, Richard.
Oeyvind

2010/9/12 Richard Dobson :
> Hmm, well I think the short answer is that the constpower pan is specific to
> the situation of two loudspeakers - in making a phantom image the two
> acoustic sources sum as powers, not as amplitudes (if that is the right way
> of putting it). I have also heard explanations in the form: ten singers
> singing a passage are not ten times as loud as one singer.
>
> As it happens, non-linear mono crossfades are far from unusual - very common
> options in samplers defining loop points. The case of "identical" signals is
> a little moot, as ostensibly if you happen to combine them out of phase,
> cancellation effects can be significant. When doing a "morphing" kind of
> crossfade,  my first instinct is to do a const-power mix. Certainly in
> implementing a wet/dry mix control in a plugin, I found const-power sounded
> the smoothest. But the whole thing is full of nuances, not least the vexed
> question of the non-linear response of the ear. To merge sounds with peak
> energy in different frequency bands, almost by definition, short of getting
> abstrusely technical about it all, you just find the law that sounds the
> most "right". As a developer of course I can pass the buck and offer the
> user a choice - then I am no longer responsible for the decision. Which is
> probably why good mixing/mastering engineers do in the end earn their fee.
>
> Richard Dobson
>
> On 12/09/2010 21:52, Oeyvind Brandtsegg wrote:
>>
>> Ah, yes, that's the proper explanation of the "hole in the middle", thanks
>> :-)
>>
>> On a somewhat related subject: Do you also have a hint on explaining
>> equal power crossfading between two sources ?
>> If one crossfades between two identical signals (yes, why would one do
>> that, exept for testing...) we get a constant perceived loudness when
>> using a linear crossfading curve. But when crossfading between two
>> different signals (as we would normally do), I find that the same
>> curve as commonly used for equal power panning (e.g. sqrt(kpan)) does
>> make it sound like the output maintains the same power during the
>> crossfade. This of course creates a digital signal with a higher
>> amplitude than each of the input signals (approx. 1.4 the original amp
>> midway in the crossfade).
>>
>
>
>
> Send bugs reports to the Sourceforge bug tracker
>           https://sourceforge.net/tracker/?group_id=81968&atid=564599
> Discussions of bugs and features can be posted here
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>
>


Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"