:cool:

[Michael Krusch]

 

[NewYorkDave]

Software?

Geez, kids today... :razz:

Along with some not-too-complicated support circuitry, a scope in X-Y mode can be used to measure the input-output transfer function.

 

[Michael Krusch]

 

[Greg]

Here's another simple brute force way at a given frequency... just measure input vs output voltages, and plot them.

I did this so I could "see" the knee on my latest G4000. The plots definitely agree with the behavior I've experience at each ratio. Looks something like this:
http://members.cox.net/capstanrecording/G4000/G4000_Knee.pdf

 

[Michael Krusch]

 

[deveng]

I looked for days to find a cheap or better yet, free, application that uses your soundcard to do this measurment. No luck. The problem is you need one that will sweep the level. I couldn't find one that had this option. I just did it the hard way as well. Set level, measure, plot in Excel. It is slow but works.

Jeff

 

[clintrubber]

[quote author="deveng"]The problem is you need one that will sweep the level. I couldn't find one that had this option.
Jeff[/quote]
Just wondering, if there's a measuring-program that allows you to use a .wav-file als the signal source then you could 'construct' a .wav with different levels.

Or skip the soundcard and program it in your Audio Precision :twisted:

 

[craigb]

I came across this program a while back. It has a pretty well featured signal generator, including level and freq. sweeps.
http://www.ymec.com/products/dssf3e/index.htm

 

[TedF]

I find that the big problem with looking for easy ways of measuring, is that you can lose focus on what you were looking for in the first place.
Measuring the shape of the 'knee', and the ratio of a compressor is easy with a generator and a dB meter, but any compressor worth using for real audio work will have variations in shape of compression response, attack, release and the dynamics of all three. A good old-fashioned 'scope and a way of firing pulses at the compressor is much the best way to see what's happening..... Talk about 'ratio' and 'attack times' is like calling Sophia Loren a 'woman'. (well, it meant something back in the 1060s.) :guinness:

 

[bcarso]

[quote author="TedF"]I find that the big problem with looking for easy ways of measuring, is that you can lose focus on what you were looking for in the first place.
Measuring the shape of the 'knee', and the ratio of a compressor is easy with a generator and a dB meter, but any compressor worth using for real audio work will have variations in shape of compression response, attack, release and the dynamics of all three. A good old-fashioned 'scope and a way of firing pulses at the compressor is much the best way to see what's happening..... Talk about 'ratio' and 'attack times' is like calling Sophia Loren a 'woman'. (well, it meant something back in the 1060s.) :guinness:[/quote]

1060's?? Ted, we're not that old! :razz:

Seriously though, yes, the dynamics are far more interesting than the steady-state. But try to get a handle on those with even a quasi-3D plot!

 

[Michael Krusch]

 

[bcarso]

Thing about it: what we tend to hear the most, for determining/identifying instrument timbres for instance, is the sound of a note's attack. See the work done by J-C Risset, Mathews, and I'm sure by now many others for example.

We test equipment with long duration sinusoids mostly. Then sometimes now we test with tone bursts, quasi-impulses, pseudorandom noise, and various sorts of swept signals (see for example the Farina stuff in the tubular thread nearby).

The way that a compressor's attack/release function works is difficult to quantify and key to how well it preserves, or how it modifies, the source material. The steady-state degree of compression is an important parameter, but won't tell you about the way you got there.

Not to say it can't be done, but it's non-trivial.

 

[Michael Krusch]

 

[deveng]

Clintrubber wrote:

Just wondering, if there's a measuring-program that allows you to use a .wav-file als the signal source then you could 'construct' a .wav with different levels.

I thought about that but didn't get to actually applying it. I guess you could record a sin wave the just do linear fade in with your favorite audio application. Then use the wav file with RMAA for example.

I agree with the other views about measuring compressors but in my case (the C*1 clone) all I wanted to do initially was to verify basic function and plot slope and threshold. There's definatley a time for using a scope and dynamic signals but also for basic verification tools.

I forgot to mention that not only is the swept input an issue but plotting level vs time is not a function I've found on the typical freeware analysis tools. You might check out Freeview_Sound.

http://www.hacker-technology.com/4361/30004.html

It's a capture/plotter type application that utilizes your PC sound card. If you have the time and skill, you can write your own code for it (Boreland Delphi 5, VEE, Visual Basic and C++)

I've got it but haven't had the time to try adapting it to this application. It looks like it could be very usefull though.



Regards,
Jeff