The RF power meter is very interesting. I have often wondered how that is or was done. Mearuring the temperature to get a power measurement in this case is almost a sort of isolator connecting the microwave world to the low frequency world. I love to see such out of my ordinary world things like this.
And the Vari-Mu thing is fascinating as well.
And the Vari-Mu thing is fascinating as well.
So for the rectifier circuit, a nice oscilloscope graphic is posted showing input / output. What would the log amp output be, say in response to a triangle modulated level of an audio "carrier"? Just to see what it does in response to a linearly increasing / decreasing signal level? For comprehensional purposes, as I'm unable to imagine it just looking at the circuit.
In the case of the Omnipressor the log amp is the rectifier. You can think of it as a logarithmic full wave rectifier that's peak detected. Filtering it to produce an RMS response would slow it down and you would no longer have an Omnipressor. The rectifier is logarithmic due to the exponential input to the Blackmer VCA that needs to be canceled to a linear control voltage. The complete system gives the same response as an OTA with a linear rectifier. It's all simpler than it seems.
Got it - thanks! I was remembering the words "linear decibel companding" from the DBX literature and thought the control had to be logarithmic because DBs are. So with double linear CV you get apparently half as loud.
I would love to see something like that. Like this is something i miss at home ( lots of synth and instrumentsvin here).
A problem is, where do you get the good sounding springs these days?
This guy has more than 10 available; https://www.ebay.com/itm/403123524477 But is it a good sounding unit?
And, like all things audio, there's always the money is no object offering https://www.ebay.com/itm/326151513136
Guess what T de Paravicini implemented in his version of 660? Cascode of pcc189/188. And he redesigned the output transformer too.
I didn't try chip amp but an amp with a 100v/70v line output transformer match the level needed to drive time constant circuitry ( original). The issue being you need to hp the sidechain input if you don't want to run into nasty amp issues... tubes were 6k4p-ev used in triode mode ( 12ba6a russian equivalent, more or less). Curves are close to 6386 in triode modes: Manley used this tubes on original 'varimu' when original 6386 stocks gone scarce and unreliable: Tbar mod.
And speaking of Manley's VariMu, David used solidstate in sidechain to drive 6386. Not as much compression range as a 670 ( less voltage in drive signal) but it sound reminiscent for low comp ratio.
De Paravicini's use solidstate in sidechain too. Very clever circuit imho. Worth taking a look at it.
I would be scared of chipamp behavior driving an output transformer. But i might be overscared i don't know? That said i managed to kill a chanel of the amp i used for experiment... he didn't liked ouput transformers saturation it seems.
One would think there's no magic in the transformer that simply drives the bridge rectifier in that time constant circuit. I bet you could approach a 70V level from two, 180 deg out of phase chipamp outputs, doing a bridge tie to the bridge rectifier input.
In my experience none of it is necessary, but fun nonetheless. Meaning the tubes, the voltage, the transformers etc., etc.. It can all be done with an 8 volt single supply and a small handfull of transistors and op-amps. How to do it exactly does take some serious knowledge but quite feasible.
The big hardware was simply a necessity in the 1950s when the Fairchild was developed. Now the same effect can be achieved by simply building a very good feedback compressor while carefully leaving some of the low order distortion intact and try not to outperform the Fairchild too much because now it can all be done faster and with much lower distortion etc,. That can ruin the magic in some cases.
It's certainly fun of course on a hobby level to try it all and see what you get. My hobby just happens to be related to making tiny but analog versions of classic things and see what I can get away with. It's also about discovering how those old records were produced and when the old techiques are something that should be preserved or not.
The big hardware was simply a necessity in the 1950s when the Fairchild was developed. Now the same effect can be achieved by simply building a very good feedback compressor while carefully leaving some of the low order distortion intact and try not to outperform the Fairchild too much because now it can all be done faster and with much lower distortion etc,. That can ruin the magic in some cases.
It's certainly fun of course on a hobby level to try it all and see what you get. My hobby just happens to be related to making tiny but analog versions of classic things and see what I can get away with. It's also about discovering how those old records were produced and when the old techiques are something that should be preserved or not.
So I did some measurements on my tube amp with the Vactrol attenuators, to see what the raw CV does in terms of their response. There are two, one for each of a ~50V RMS (+) and (-) sinewave signal. I have their illuminating LEDs connected in series, current limited by a 100k resistor. These are in parallel with the input to a 250K pot, which is the post volume. I'll guess ~50K as the source impedance...
I tried to plot db in attenuation versus control voltage; thanks to my wife for the excel formulas that would take me all day otherwise. The Vactrols remain virtually open circuit until CV reached about 3.1 and I'm 6 db down by the time CV has increased to about 5.5V . Even with the LED current exact, they dont track too well at the higher attenuation...for the two of three in a bag I purchased off an ebay seller.
My question is, is the attenuation range "suitable" or "applicable" for a guitar compressor function? It sounds like it's certainly doing something, after offsetting the 3V "threshold" and driving it down with a textbook op-amp full wave rectifier, where the gain has been adjusted to provide a perceptible effect.
Just wondering if I need to use a different topology to get even more attenuation, or would more than 6db never be used practically by a guitar player anyway, as it would squash the sound down well beyond a generally musically useful level? Like, if you did an average across all guitar players who routinely use compressors, what would the average and +3 sigma attenuation in db be? I have no idea -
I tried to plot db in attenuation versus control voltage; thanks to my wife for the excel formulas that would take me all day otherwise. The Vactrols remain virtually open circuit until CV reached about 3.1 and I'm 6 db down by the time CV has increased to about 5.5V . Even with the LED current exact, they dont track too well at the higher attenuation...for the two of three in a bag I purchased off an ebay seller.
My question is, is the attenuation range "suitable" or "applicable" for a guitar compressor function? It sounds like it's certainly doing something, after offsetting the 3V "threshold" and driving it down with a textbook op-amp full wave rectifier, where the gain has been adjusted to provide a perceptible effect.
Just wondering if I need to use a different topology to get even more attenuation, or would more than 6db never be used practically by a guitar player anyway, as it would squash the sound down well beyond a generally musically useful level? Like, if you did an average across all guitar players who routinely use compressors, what would the average and +3 sigma attenuation in db be? I have no idea -
This is why Vatrols are normally used for feedback compressors. The control voltage is then part of global sidechain feedback loop. This tends to normalize the curve and come up with something useful. Feedback compressors work very well with guitars but Vactrols are slow and why they're mostly used for vocals.
My version of OTA based feedback compressors can be very fast however. The one used on the Cowgirl In The Sand test recording measures in at about 20 µS. Less than one cycle at 20 kHz. The artifacts are next to nothing. This is a big part of how that recording manages to blend guitar and vocal 3 feet from an ordinary volcal mic and my wife doing dishes about 12 feet behind the mic in it's dead zone. None of these things mute the other in the recording. It's pretty remarkable really. In all of my busy time I have only made one and it was embedded in one of my guitar amps and sold. Maybe that compressor would make an interesting discussion on feedback compressors or more specifically how to make Dynacomps and Guyatones etc. look a tad lame. That's what it did to me.
My version of OTA based feedback compressors can be very fast however. The one used on the Cowgirl In The Sand test recording measures in at about 20 µS. Less than one cycle at 20 kHz. The artifacts are next to nothing. This is a big part of how that recording manages to blend guitar and vocal 3 feet from an ordinary volcal mic and my wife doing dishes about 12 feet behind the mic in it's dead zone. None of these things mute the other in the recording. It's pretty remarkable really. In all of my busy time I have only made one and it was embedded in one of my guitar amps and sold. Maybe that compressor would make an interesting discussion on feedback compressors or more specifically how to make Dynacomps and Guyatones etc. look a tad lame. That's what it did to me.
Response speed is certainly one defining character of a compressor. I'm trying to grok what it does to the signal amplitude, in a quantifiable way, that I can the relate to how the effect sounds in that particular dimension. I'm all for how it sounds BTW, just want to understand it better. And be able to speak the language, from perception to circuitry.
I visited the reddit "guitar pedal" page, where someone did a decent write up. They made statements like a 6 to 1 compression ratio means for a 6 db increase in amplitude coming in, it gets compressed down to a 1 db - or practically imperceptible - increase in amplitude coming out. They also gave some common ratios actually used in practice, none of which were anything like 10 or 20:1, most well under 8:1, so I'm getting a feel for what a practical range for these things is. That is, if your VCA can attenuate 8 db or even 6, nobody's going to use more in general practice - unless they're doing some strange, one off effect.
Am I about right?
Then there's the https://sengpielaudio.com/calculator-loudness.htm?source=https://vi-control.net/community page, which helps to quantify db changes relative to signal and power level changes. I didnt even know it takes 10 db, or going from a 10W amp to a 100W amp, to get just twice as loud perceptibly. ( I thought it was 6...) Anyway, from there I gleaned that a 6 db drop is perceived as ~2/3 as loud. So, using a 6:1 ratio, stomps the dynamic of, say a hard struck chord, down to 2/3 of what it would have been. Seems reasonable.
Still, I'm not yet at the level of understanding something like how someone with a 30W amp using a 6:1 compression ratio can sound as loud as someone with a 100W amp and no compression. I'll guess that's an effect happening in the temporal dimension, which is energy = watts X time.
BTW, I find your demos inspiring. "Tuesday Afternoon" caused me to find, format to a single page, print and put in my song book and start to learn to play / sing. "Cowgirl in the Sand" inspired me to work on that one more to refine. Funny how an interest in audio compression technology causes me to improve my repertoir - maybe that's what I'm supposed to get out of participating in this thread. Got anything else to demo that I can do? ;') I mentioned "Quiet Nights", which - Jobim - is a level of chording I'll never be able to do in this lifetime. Happily challenged just moving from F#m to Bm and back.
I visited the reddit "guitar pedal" page, where someone did a decent write up. They made statements like a 6 to 1 compression ratio means for a 6 db increase in amplitude coming in, it gets compressed down to a 1 db - or practically imperceptible - increase in amplitude coming out. They also gave some common ratios actually used in practice, none of which were anything like 10 or 20:1, most well under 8:1, so I'm getting a feel for what a practical range for these things is. That is, if your VCA can attenuate 8 db or even 6, nobody's going to use more in general practice - unless they're doing some strange, one off effect.
Am I about right?
Then there's the https://sengpielaudio.com/calculator-loudness.htm?source=https://vi-control.net/community page, which helps to quantify db changes relative to signal and power level changes. I didnt even know it takes 10 db, or going from a 10W amp to a 100W amp, to get just twice as loud perceptibly. ( I thought it was 6...) Anyway, from there I gleaned that a 6 db drop is perceived as ~2/3 as loud. So, using a 6:1 ratio, stomps the dynamic of, say a hard struck chord, down to 2/3 of what it would have been. Seems reasonable.
Still, I'm not yet at the level of understanding something like how someone with a 30W amp using a 6:1 compression ratio can sound as loud as someone with a 100W amp and no compression. I'll guess that's an effect happening in the temporal dimension, which is energy = watts X time.
BTW, I find your demos inspiring. "Tuesday Afternoon" caused me to find, format to a single page, print and put in my song book and start to learn to play / sing. "Cowgirl in the Sand" inspired me to work on that one more to refine. Funny how an interest in audio compression technology causes me to improve my repertoir - maybe that's what I'm supposed to get out of participating in this thread. Got anything else to demo that I can do? ;') I mentioned "Quiet Nights", which - Jobim - is a level of chording I'll never be able to do in this lifetime. Happily challenged just moving from F#m to Bm and back.
Peak reduction.
If there is no ( or low) peaks and you use the whole 30w availlable of power, your rms spl level( average level, the one our auditory system use to 'feel' level) will be the same ( or higher) than a 100w amp without the peak reduction ( original 'wide dynamic' input signal).
I'm not sure if you guys know about this site, but here it is: https://sengpielaudio.com/Calculations03.htm
This guy died some time ago and his son keeps the site maintained and it's a wonderful thing. I'm not even sure when I first used it. Maybe 20 years ago or even more. It's one of my favorite tools to be lazy. I am just too busy to look up some of this that I don't use often enough to get off the top of my head.
This guy died some time ago and his son keeps the site maintained and it's a wonderful thing. I'm not even sure when I first used it. Maybe 20 years ago or even more. It's one of my favorite tools to be lazy. I am just too busy to look up some of this that I don't use often enough to get off the top of my head.
I think I figured out why I like compression and what it means to me to "ride" on the effect. I can get louder without having to sing louder or "as hard". I did the Moody Blues song Tuesday Afternoon and I distinctly ran right out of breath on that "Explain it all with a sigh ahhh ahhh" part, where it goes to the Ab and then C. I even ran short the second time knowing it was coming up again. I bet if the compressor was up and somewhat deep, I would have been able to use less breath across that span of sound / time.
