Hi Mark, I really didn't understand what you called "Static Compression
Curve", but anyway... I would try a burst signal, sine wave. Then,
compute the envelope and take the parameters.
Regards,
C.Herrera
________________________________
________________________________
De: "m...@gmail.com"
Para: a...
Enviadas: Ter-feira, 14 de Abril de 2009 13:17:37
Assunto: [audiodsp] Test Signal to get Static Compression Curve?
What is the ideal test signal to use to get a Static Compression Curve for a
dynamics processor (compressor) ?
Can you assume that the processor operates the same across 20Hz-20kHz?
In that case, then a simple 1 kHz sin waveform that is gradually increased in db
amplitude should work as long as its slow enough to not trigger the attack and
release sections of the compressor?
The compressor I am looking to characterize is a peak FET-type, feedback with
relatively medium to fast attack and release. (All way within 1 second).
_____________________________________
Res: Test Signal to get Static Compression Curve?
Started by ●April 15, 2009
Reply by ●April 16, 20092009-04-16
A Static Compression Curve is the Input-Output relationship of a Dynamics
Processor ignoring all Attack and Release settings. It gives the exact details
of the compressor's threshold ratio. For example, a 2:1 ratio could be
implemented as a hard knee, or a soft knee ratio, both having very different
results.
I found the following from UA's webzine in the sept 08 article, "Usually, static compression curves are measured by recording the response of a compressor to a sinusoidal input at a single frequency. The input is held at constant amplitude until the output settles into steady state, and then the amplitude of the output is recorded. Output level is plotted as a function of input level, on a dB-dB scale. Static compression ratio is defined as the reciprocal of the slope of the static compression curve. For example, if an increase in input level of 4 dB results in an increase in output level of 1 dB, then the ratio of compression is 4:1."
This is sort of what I was thinking, but wanted to see if anyone had any real expirience, and strategies on how to get the final curve? Maybe i could use a single sinusoid and change by 1 db every 1 second. Then only use the 2nd half of the sinusoid output for my plot?
Hi Mark, I really didn't understand what you called "Static Compression
>Curve", but anyway... I would try a burst signal, sine wave. Then,
>compute the envelope and take the parameters.
>
>Regards,
>C.Herrera
>
>________________________________
>________________________________
>De: "m...@gmail.com"
>Para: a...
>Enviadas: Ter-feira, 14 de Abril de 2009 13:17:37
>Assunto: [audiodsp] Test Signal to get Static Compression Curve?
>
>What is the ideal test signal to use to get a Static Compression Curve for >a dynamics processor (compressor) ?
>
>Can you assume that the processor operates the same across 20Hz-20kHz?
>
>In that case, then a simple 1 kHz sin waveform that is gradually increased >in db amplitude should work as long as its slow enough to not trigger the a>ttack and release sections of the compressor?
>
>The compressor I am looking to characterize is a peak FET-type, feedback wi>th relatively medium to fast attack and release. (All way within 1 second).
>
>_____________________________________
I found the following from UA's webzine in the sept 08 article, "Usually, static compression curves are measured by recording the response of a compressor to a sinusoidal input at a single frequency. The input is held at constant amplitude until the output settles into steady state, and then the amplitude of the output is recorded. Output level is plotted as a function of input level, on a dB-dB scale. Static compression ratio is defined as the reciprocal of the slope of the static compression curve. For example, if an increase in input level of 4 dB results in an increase in output level of 1 dB, then the ratio of compression is 4:1."
This is sort of what I was thinking, but wanted to see if anyone had any real expirience, and strategies on how to get the final curve? Maybe i could use a single sinusoid and change by 1 db every 1 second. Then only use the 2nd half of the sinusoid output for my plot?
Hi Mark, I really didn't understand what you called "Static Compression
>Curve", but anyway... I would try a burst signal, sine wave. Then,
>compute the envelope and take the parameters.
>
>Regards,
>C.Herrera
>
>________________________________
>________________________________
>De: "m...@gmail.com"
>Para: a...
>Enviadas: Ter-feira, 14 de Abril de 2009 13:17:37
>Assunto: [audiodsp] Test Signal to get Static Compression Curve?
>
>What is the ideal test signal to use to get a Static Compression Curve for >a dynamics processor (compressor) ?
>
>Can you assume that the processor operates the same across 20Hz-20kHz?
>
>In that case, then a simple 1 kHz sin waveform that is gradually increased >in db amplitude should work as long as its slow enough to not trigger the a>ttack and release sections of the compressor?
>
>The compressor I am looking to characterize is a peak FET-type, feedback wi>th relatively medium to fast attack and release. (All way within 1 second).
>
>_____________________________________
Reply by ●April 16, 20092009-04-16
Mark-
> A Static Compression Curve is the Input-Output relationship of a Dynamics Processor ignoring all Attack and Release
> settings. It gives the exact details of the compressor's threshold ratio. For example, a 2:1 ratio could be
> implemented as a hard knee, or a soft knee ratio, both having very different results.
>
> I found the following from UA's webzine in the sept 08 article, "Usually, static compression curves are measured by
> recording the response of a compressor to a sinusoidal input at a single frequency. The input is held at constant
> amplitude until the output settles into steady state, and then the amplitude of the output is recorded. Output level
> is plotted as a function of input level, on a dB-dB scale. Static compression ratio is defined as the reciprocal of
> the slope of the static compression curve. For example, if an increase in input level of 4 dB results in an increase
> in output level of 1 dB, then the ratio of compression is 4:1."
>
> This is sort of what I was thinking, but wanted to see if anyone had any real expirience, and strategies on how to get
> the final curve? Maybe i could use a single sinusoid and change by 1 db every 1 second. Then only use the 2nd half of
> the sinusoid output for my plot?
What about a series of swept sine waves, each with increasing amplitude? Then you could plot a 3-D spectrograph
showing both amplitude response varying with your input, and also frequency response. You would know a) whether your
amplitude response is as expected (hard or soft knee, etc) and b) that the compressor algorithm didn't do anything
unusual w.r.t frequency.
-Jeff
> Hi Mark, I really didn't understand what you called "Static Compression
>>Curve", but anyway... I would try a burst signal, sine wave. Then,
>>compute the envelope and take the parameters.
>>
>>Regards,
>>C.Herrera
>>
>>________________________________
>>________________________________
>>De: "m...@gmail.com"
>>Para: a...
>>Enviadas: Ter-feira, 14 de Abril de 2009 13:17:37
>>Assunto: [audiodsp] Test Signal to get Static Compression Curve?
>>
>>What is the ideal test signal to use to get a Static Compression Curve for >>a dynamics processor (compressor) ?
>>
>>Can you assume that the processor operates the same across 20Hz-20kHz?
>>
>>In that case, then a simple 1 kHz sin waveform that is gradually increased >>in db amplitude should work as long as its slow enough to not trigger the a>>ttack and release sections of the compressor?
>>
>>The compressor I am looking to characterize is a peak FET-type, feedback wi>>th relatively medium to fast attack and release. (All way within 1 second).
> A Static Compression Curve is the Input-Output relationship of a Dynamics Processor ignoring all Attack and Release
> settings. It gives the exact details of the compressor's threshold ratio. For example, a 2:1 ratio could be
> implemented as a hard knee, or a soft knee ratio, both having very different results.
>
> I found the following from UA's webzine in the sept 08 article, "Usually, static compression curves are measured by
> recording the response of a compressor to a sinusoidal input at a single frequency. The input is held at constant
> amplitude until the output settles into steady state, and then the amplitude of the output is recorded. Output level
> is plotted as a function of input level, on a dB-dB scale. Static compression ratio is defined as the reciprocal of
> the slope of the static compression curve. For example, if an increase in input level of 4 dB results in an increase
> in output level of 1 dB, then the ratio of compression is 4:1."
>
> This is sort of what I was thinking, but wanted to see if anyone had any real expirience, and strategies on how to get
> the final curve? Maybe i could use a single sinusoid and change by 1 db every 1 second. Then only use the 2nd half of
> the sinusoid output for my plot?
What about a series of swept sine waves, each with increasing amplitude? Then you could plot a 3-D spectrograph
showing both amplitude response varying with your input, and also frequency response. You would know a) whether your
amplitude response is as expected (hard or soft knee, etc) and b) that the compressor algorithm didn't do anything
unusual w.r.t frequency.
-Jeff
> Hi Mark, I really didn't understand what you called "Static Compression
>>Curve", but anyway... I would try a burst signal, sine wave. Then,
>>compute the envelope and take the parameters.
>>
>>Regards,
>>C.Herrera
>>
>>________________________________
>>________________________________
>>De: "m...@gmail.com"
>>Para: a...
>>Enviadas: Ter-feira, 14 de Abril de 2009 13:17:37
>>Assunto: [audiodsp] Test Signal to get Static Compression Curve?
>>
>>What is the ideal test signal to use to get a Static Compression Curve for >>a dynamics processor (compressor) ?
>>
>>Can you assume that the processor operates the same across 20Hz-20kHz?
>>
>>In that case, then a simple 1 kHz sin waveform that is gradually increased >>in db amplitude should work as long as its slow enough to not trigger the a>>ttack and release sections of the compressor?
>>
>>The compressor I am looking to characterize is a peak FET-type, feedback wi>>th relatively medium to fast attack and release. (All way within 1 second).
