etheory - a bunch of projects in progress

[evilcat]

Yep, I've seen it and loved it! From the technical AND artistic point of view. I would have a lot of good things to say about it but in french, I'm not good enough in english!!! I don't know if it's correct in english but I found it very bold.

 

[etheory]

We pick up the keys today to our apartment and move in tomorrow. So freaking excited!
I've already figure out how to set up my man-cave....
Not long now before I can dust off the soldering iron and go nuts 8)

 

[etheory]

And..... We've moved.
2 weeks to wait for our Internet connection to be re-established in the new place, and by that point I should have my soldering station set up in the garage downstairs and the music studio set up and I can get back into making some stuff. Until then, more unpacking! More soon....

 

[etheory]

Almost back up and running again.
Internet is connected.
Half my computer is up but without audio.
This weekend I'll do the audio interconnect and the set up of my new soldering space in our garage, and then I'll be back into it.

 

[mig27]

Hope I'm not getting on your nerves but I'm still eagerly hoping for your e274 project to come alive.
Is this still happening?

Thanks!

 

[iprovlek]

Hope I'm not getting on your nerves but I'm still eagerly hoping for your e274 project to come alive.
Is this still happening?

Thanks!

Me toooo!

Waiting for E274 and w295b

 

[jordan s]

Bring it back! I gotta build those comps.

 

[mig27]

Sorry for necro-posting, but I really want to give this a last bump.
We're all very keen here on your e274 project..

Thaaanks!


Michael

 

[etheory]

Hi everyone!

I'm really humbled by the interest in these projects.

I just wanted to assure you that I haven't binned any of them, merely re-prioritized my life a little.
After obtaining a mortgage, I decided that I needed to divert my attention to projects to make me the money I need to pursue the electronics projects (which cost money to test and purchase parts for).

So in the meantime I've taught myself much more in-depth electronics theory, and started to model VSTs of circuits in unprecedented detail, to make commercial VSTs that I will be selling to raise the funds to finish these electronics projects.

You can find a few early examples of this work here:
https://soundcloud.com/evolutionarytheory/evolutionary-theory-ms-20-lpf-late-model-vst-plug-in
https://soundcloud.com/evolutionarytheory/evolutionary-theory-4-pole-roland-inspired-filter-zero-delay-feedback-tpt
https://soundcloud.com/evolutionarytheory/evolutionary-theory-ms-20-hpf-late-model-vst-plug-in
https://soundcloud.com/evolutionarytheory/evolutionary-theory-etheorycompressor-vst-test-01

My main goal is to model, with extreme accuracy (based on careful analysis of the real circuits + proper physical modelling techniques that I've been improving over time), some analog synthesizer filters, a compressor, and a few distortion circuits.

So far the results have been nothing short of spectacular. I've also done some research into new mathematical/numerical methods which I believe could be re-used for many different purposes, and after the release of these plug-ins I'll be releasing a number of math papers detailing the new techniques used in these plug-ins (all of which are novel, and invented by me).

I expect the first commercially-available releases to be before the end of this year, after which, if I manage to raise some funds through sales, I intend to fully rekindle all of my analog audio projects, along with a bunch of others (including the compressors and filters detailed throughout this thread).

Thanks so much for the support, I definitely haven't stopped doing stuff! On the contrary it's full-steam ahead!

cheers.

 

[opacheco]

I finally just realized why the value of C3 sitting near the diode bridge is the value it is.
I raised it to 22uF in my circuit thinking it wouldn't make much difference, initially thinking it was just to smooth out the negative feedback and Ts1 bias signal, but then, after changing it to be the same ratio (4.7:1) as it was in the original circuit, a wonderful thing started to happen.

The weird bias + the 10uF C3 capacitor, removes a LOT of the CV feed-through in that circuit in a very clever way.

I reduced C1, C2 to 10uF for a much faster attack time, which means that C3 I'll set to 2.2uF to get a close-ish ratio to the original (which used 47uF to 10uF).

What happens is that C3 and R13 and R14 form a low pass filter that smooths the compressed signal flowing through the emitter of Ts2, along with providing negative feedback via the input transformer to Ts1, that bias and negative feedback line also contains an inverted filtered signal, consisting mostly of the CV component of the signal, and hence due to the negative feedback, partially cancels it out of the signal path.

When I first found this I couldn't believe it, it's genius. This U274 circuit, even though it's basically just this one (with a few modifications):

is so subtly clever. Those Siemens Sitral guys were on the money for the parts count.

There is now almost no CV feed-through at all. Pretty cool stuff!

If you remove that capacitor entirely, there is no CV feed-through at all, but it comes at the cost of attack speed, which drops significantly.

eTheory, very interesting analysis!!......

I would like to show the Siemens U273 bridge diodes is very similar to these circuits too : This network is shut to the signal too but have variable timing controls (Release Only).  Look like the R6/C3 and R11/C7 are a kind of filter (Shut RC Filter)  for the signal coming thru the D1, D2, D3, D4 (when conducting) to the ground.

I would like to get some ideas in order to modify the attack (maybe with series resistors in the Vcontrol) or Ratio control with some extra gain for the side chain amp (thru R7 or R2 in the original schematic).

Any comment?
Opacheco

 

[etheory]

You are absolutely correct, the U273 is approximately the same idea.

For ratio, you just need to change the gain of the side-chain, since the side-chain gain IS the ratio.

For changing attack and decay times, you could either change the timing capacitors (but that'll change both attack and decay at the same time), or change the resistors in that section of the circuit. I'd just play with it until you like how it sounds. You could always disconnect the diode-bridge VCA from your U273, build-up on on strip-board, and tune it with variable attack and decay until you are happy, then splice your changes in to your full unit.

Happy experimenting!

 

[ruffrecords]

For ratio, you just need to change the gain of the side-chain, since the side-chain gain IS the ratio.

Are you sure about that? Usually, changing the gain of the side chain just changes the threshold.

Cheers

Ian

 

[etheory]

Are you sure about that? Usually, changing the gain of the side chain just changes the threshold.

Cheers

Ian

100% positive.

To change the threshold you need to change the DC offset of the signal that goes into your knee. In many older designs, the knee is the envelope-generator diode in your side-chain - the one that feeds the charging cap and release resistor, via the attack resistor. This diode drop, of around 0.6V (or higher for a tube connected as a diode like in the RS124 etc.) is the effective "knee" of the compressor, since signals below this won't feed the side-chain at all (as the diode will be off).

So in many designs they are kind of connected. In the ideal case, for a perfect diode, the threshold will be some kind of DC offset driven into a diode connected to ground (to chop off the rectified signal), and the ratio is a gain applied to that result.

For a feedback compressor the overall gain during compression is computed as (with some unstated assumptions to keep things simple, like we are working in dB for working units, and the detectors and VCA is logarithmic, but it's close enough to being an instructive example to leave it as is):

y = x - G*y (the output is the input, with negative feedback from the side-chain, through some gain)
i.e.
y*(1+G) = x
i.e.
y = x * (1/(1+G))
i.e. ratio x/y = (1+G)

So your ratio, is just (1+G) where G is the gain of your side-chain.

For feed-forward compression, it's:
y = x  - G*x (the output is the input, with the side-chain removing G of your gain)
i.e.
y = x * (1-G)
i.e. ratio x/y = 1/(1-G)

So your ratio, is just 1/(1-G), where G is the gain of your side-chain.

This all applies for signals above the threshold.

So the ideal compressor can be calculated exactly this way, but in real-world designs things can get mixed up and re-ordered add-infinitum.

Often to compute an accurate ratio, you also need to adjust the knee, since like in the 1176 the knee is a function of the diodes being driven into the envelope capacitor, and the JFET resistive-response curve used as a voltage-divider. So since the gain of the side-chain drives the knee differently, to maintain the same effective knee, you change the DC-offset along with the ratio. Swapping that arrangement around just makes for easier calculations.

 

[ruffrecords]

You have obviously thought about this a lot and we may be talking a cross purposes but I do not think you are right. Let's stick to a feedback compressor because it is the most common. If you add 10dB of gain into the side-chain before the knee, you simply lower the threshold by 10dB. You do not alter the ratio. Do you agree?

Cheers

ian

 

[opacheco]

eTheory and , ruffrecords,

Thanks a lot for retake this topic again!.....that's very exciting for me because I have take some time to analyse in my humble knowledge.

eTheory the idea of mod the attack changing the R8, R10 and C5 in order to get a more faster (or slower!) attack is really fantastic as long as the total release product (RtotalXC5) be constant in order to NO mod the actual release time!!.

In other hand, I think that extra sidechain  gain look like RATIO mod only but in the attached drawing you can see an idea to mod the Threshold to a REAL Variable Threshold Network; Here Vth is a floating voltage (no ground reference) and here I need any advise because I don't have any idea to its value range (for variable Threshold voltage) and the value R too for limit its current???....I wont to damage any Diode of the Bridge Diodes.

Thanks
Opacheco

 

[etheory]

Let's stick to a feedback compressor because it is the most common. If you add 10dB of gain into the side-chain before the knee, you simply lower the threshold by 10dB. You do not alter the ratio. Do you agree?

If you add 10dB of gain into the side-chain, before the knee, you lower the threshold by 10dB AND simultaneously increase the ratio of the compressor. You have to.

The most intuitive way I can come up with to explain it is this:

The side-chain signal fed to your VCA provides the gain change. The stronger the side-chain signal, the more the compressor VCA will attenuate.

The compressor "ratio" is the ratio of the input signal to the output signal (in dB), so a stronger side-chain signal increases this ratio, by definition.

For the portion of the signal that's above the threshold, the remaining side-chain signal is going to be 10dB stronger than it was (due to the 10dB gain). So the ratio of the compressor will increase by the factor for 10dB = 3.16 approx, and subbing this into the previous computation for the ratio of a feedback compressor: (1+3.16) = 4.16x

So for a feedback compressor, increasing your side-chain gain by 10dB will lower your threshold by 10dB and simultaneously increase the ratio by 4.16x.

 

[etheory]

you can see an idea to mod the Threshold to a REAL Variable Threshold Network; Here Vth is a floating voltage (no ground reference) and here I need any advise because I don't have any idea to its value range (for variable Threshold voltage) and the value R too for limit its current???....I wont to damage any Diode of the Bridge Diodes.

I'm not sure if this will work but it'll be easy enough to simulate in LTSpice, I'll get back to you.

 

[opacheco]

you can see an idea to mod the Threshold to a REAL Variable Threshold Network; Here Vth is a floating voltage (no ground reference) and here I need any advise because I don't have any idea to its value range (for variable Threshold voltage) and the value R too for limit its current???....I wont to damage any Diode of the Bridge Diodes.

I'm not sure if this will work but it'll be easy enough to simulate in LTSpice, I'll get back to you.

That's sound nice!!....The D2 will be conducting when the rectified voltage from the side chain coming from the U transformer (this rectified voltage have its negative polarity always in the terminal near to cathode of D2 and its floating too!) be equal or more than Vth in other words:

Absolute value of (Vcontrol) => Absolute value (Vth)

In this point we will be modifying the threshold level to upper voltage....May be I am wrong!!



Let us to know your simulation results please.
Opacheco.

 

[ruffrecords]

Let's stick to a feedback compressor because it is the most common. If you add 10dB of gain into the side-chain before the knee, you simply lower the threshold by 10dB. You do not alter the ratio. Do you agree?

If you add 10dB of gain into the side-chain, before the knee, you lower the threshold by 10dB AND simultaneously increase the ratio of the compressor. You have to.

The most intuitive way I can come up with to explain it is this:

The side-chain signal fed to your VCA provides the gain change. The stronger the side-chain signal, the more the compressor VCA will attenuate.

The compressor "ratio" is the ratio of the input signal to the output signal (in dB), so a stronger side-chain signal increases this ratio, by definition.

OK, I think we are getting somewhere.  At least you agree the threshold drops by 10dB. However, I think your definition of ratio is wrong. It is the ratio of the change in input signal to the change in output signal. This is defined almost entirely by the slope of the gain control element. It is independent of the gain before or after it in the signal chain or gain in the side chain. You may think this is a surpising result but it is true as measuememts will verify - which is how I doscovered it in the first place.

Cheers

Ian

 

[etheory]

Hi Ian,

Thanks for the discussion, it's good to nut this stuff out.

However, I think your definition of ratio is wrong. It is the ratio of the change in input signal to the change in output signal.

That doesn't change anything I said, it still applies equally whether measured on a relative or absolute basis.

This is defined almost entirely by the slope of the gain control element.

Not according to THAT Corp in any of their application notes it's not.

It is independent of the gain before or after it in the signal chain or gain in the side chain.

Again, look at how a compressor is implemented in any of the THAT Corp application notes. Ratio is the application of gain in the side-chain. If it's not, how do you change the ratio then?

You may think this is a surprising result but it is true as measurements will verify - which is how I discovered it in the first place.

I've recently written an audio plug-in based on measurements of compressors I've made from spare components I had lying around. And I've also measured the phenomenon I'm talking about whilst working on the e274 diode comp a while back.

Ratio definitely changes with side-chain gain. Yes, the VCA element response plays a part, but that's a fixed quantity. So to be able to change the ratio with a fixed VCA-element, it's the only option you have. Multiple classic circuits and info you can find online also backs this up.

For more info here's a good dissection from THAT Corp: http://www.thatcorp.com/datashts/dn107.pdf - note the way the ratio is defined. This equally applies with feedback compression, albeit with a different way to compute that ratio.