Analog signal compression

[Pipila]

Hi everyone,

I'm into the audio effects for the guitar and I want to understand how to compress the guitar signal to make a compressor.

I've found several diagrams and I know they'll work but can't understand how they work.

For compression I understand that the large wave length is reduced to a smaller one and the short wave length is increased. In other words low frequencies get high and high frequencies get low.

This is because if you compress the amplitude you turn down the volume so thats what I know for now, everything I have found is about coding and compressing digital signals.

Thanks in advance.

 

[MrAl]

Hi,


Provide a couple links so we can check them out. Maybe someone here can help.

 

[Pipila]

Hi,

Thanks for your interest, but links to what? A commercial compressor diagram?

DOD 280 compressor

Thats the one I'm planning to do but I want to understand the way it treats the signal.

 

[ericgibbs]

Hi,

Thanks for your interest, but links to what? A commercial compressor diagram?

DOD 280 compressor

Thats the one I'm planning to do but I want to understand the way it treats the signal.

hi,
My PC is getting a download warning from that link.!!!

 

[Torben]

Hi everyone,

I'm into the audio effects for the guitar and I want to understand how to compress the guitar signal to make a compressor.

I've found several diagrams and I know they'll work but can't understand how they work.

For compression I understand that the large wave length is reduced to a smaller one and the short wave length is increased. In other words low frequencies get high and high frequencies get low.

You're confusing wavelength with amplitude. If you lower the wavelength (raise the frequency) of an audible tone you raise its pitch; if you raise the wavelength (lower the frequency), you lower the pitch. So a device which did this would make a bass guitar sound like a banjo and vice versa.

A compressor provides an amplitude cutoff point above which the signal is attenuated. It's not for affecting frequencies--that's what EQs are for.

This is because if you compress the amplitude you turn down the volume so thats what I know for now, everything I have found is about coding and compressing digital signals.

Thanks in advance.

Now's you're on track: amplitude, not frequency. But it's not quite as simple as just turning down the volume. It's more like turning down the maximum volume that can be reached. So you can crank up the guitar (or whatever) volume to increase the volume of quieter stuff, and then stick a compressor on it to keep the louder stuff from getting too loud.


Regards,

Torben

 

[Gary B]

What you are interested in is the first part of a compander like we use in the telephone industry. The analog ones work by biasing an amplifier (transistor) so it operates close to or in its nonlinear range. This can be near saturation or cut off. Small signals don’t push it out of the linear area, but, larger ones do.

 

[Pipila]

hi,
My PC is getting a download warning from that link.!!!

Sorry, I thought it would redirect you to the pdf file.

DOD 280 compressor

That's the one.

You're confusing wavelength with amplitude. If you raise the wavelength of an audible tone you increase its pitch; if you decrease the wavelength, you lower the pitch. So a device which did this would make a bass guitar sound like a banjo and vice versa.

A compressor provides an amplitude cutoff point above which the signal is attenuated. It's not for affecting frequencies--that's what EQs are for.

Yes, I was on the wrong way. I read a little bit more and saw this video:

DOD Compressor

He raises the signal but the compressor attenuates the amplitude when it passes the high point of the dynamic range (I read that this is how it's called) and when it is below the range it's amplified.

This last event should explain that it's called compressor/sustainer it compresses large amplitudes and amplifies the lower ones and it keeps amplifying until there is no signal to amplify (the guitar string stops vibrating)

Am I right?

And last question, in the above circuit the attenuation is made by Q2?

What you are interested in is the first part of a compander like we use in the telephone industry. The analog ones work by biasing an amplifier (transistor) so it operates close to or in its nonlinear range. This can be near saturation or cut off. Small signals don't push it out of the linear area, but, larger ones do.

I didn't understand this, large signals are clipped when they pass through the transistor and only small signals pass through without distortion?

Thanks a lot for your answers.

 

[Externet]

The loud audio levels make the transistors Q1 and Q2 to conduct more than when the audio is of lower amplitude.

The amount of conduction of the transistors turn on a led more for loud audio or less for quiet passages.

That led shines its light onto a light dependent resistor in the feedback gain circuit of the amplifier.

More loudness --> more Q1, Q2 transistor conduction --> more light on the led --> more conduction on the LDR --> more attenuation of the amplifier.

Q1, Q2 do not lower the gain, they command the amplifier to lower the gain by lowering its feedback resistance.

 

[Pommie]

If you raise the wavelength of an audible tone you increase its pitch; if you decrease the wavelength, you lower the pitch.

Isn't that the wrong way around? Longer wavelength=lower pitch.

Mike.

 

[Torben]

Isn't that the wrong way around? Longer wavelength=lower pitch.

Mike.

Damn. Yep. I mixed up frequency and wavelength. I plead writing while tired.


Thanks,

Torben

[Edit: I've edited the original post in case later readers don't read the thread.]

 

[MrAl]

Hi there,


Not sure if this topic is still of interest, but the main idea with signal compression is to make a measurement of the output (feedback system) or input (feed forward system) and make a change in the gain based on if this measurement is within a given range. When the gain changes the output changes, and it is arranged so that when the output goes higher the gain goes lower and vice versa.
The resulting DE comes out a bit different than a standard feedback system because the gain is allowed to change.

 

[Pipila]

Hi everyone,

I appreciate your help, now I have a little idea of how this works but I know nothing about op-amps and feedback. I guess I'll start to read about it.

Thanks again for everything hope this helps someone else.

 

[audioguru]

I think the LDR has an action that is too slow to be used in an audio comprssor.

 

[Nigel Goodwin]

I think the LDR has an action that is too slow to be used in an audio comprssor.

LDR's (and light bulbs) are perfectly suited to audio compressors, attempting to use photo-transistors and LED's is where the problems arise.

The LDR/bulb combination gives a good attack/decay timing.

 

[audioguru]

I was mistaken.
The article about a Fast Peak Limiter on Rod Elliot's site says that an LDR/LED combination is far too slow to prevent an amplifier from clipping but is perfectly adequate for a guitar.

 

[MrAl]

Hi again,


I was going to say, if the feedback is too fast then it causes audio distortion, so there is actually a limit to what the speed of response can be, and it's usually set with an RC network.
For example, if the audio coming in is 100Hz and the time constant is only 0.01 seconds, significant sine distortion would result. Thus it's a tradeoff of compression speed vs distortion and usually the distortion takes precedence. The drawback is the 'breathing' effect.
Using digital signal processing this problem can be alleviated to some degree.

A feed forward system is very simple to analyze, so i'll try to get back here with an example some time.

 

[audioguru]

if the feedback is too fast then it causes audio distortion

If the release time is too short then low frequencies are distorted.
If the attack time is too slow then a loud sound blasts through extremely loudly with severe clipping distortion until the slow circuit "catches up".

My local TV station recently renovated their live news studio. They also replaced the sound man with a horrible compressor. Its attack is too slow and its resting gain is too high so that the first syllable of every word spoken by the anchor people is loud and hits my head like a hammer. When it compresses then it cuts the high audio frequencies maybe to reduce the distortion but then everything spoken is not clear and is barely understood.
Commercials and off site reporters sound normally crisp and clear with no noticeable compression.

 

[MrAl]

Hi again,


Yes unfortunately there are various pitfalls to various compression techniques, but some can be handled similar to regular feedback systems. For instance by including at least the first derivative in the compression equation. A system can be made to closely track inputs that vary more quickly by including the rate of change of input rather than simply the change itself.
The rate of change is an indicator of just how high the input might go, giving a sort of advance notice of how high it might really go. The second derivative of course could then be used to estimate just how high the first derivative might go, and make the system behave even smoother.