New versions are up.K-Slash wrote:Thanks, please bump this topic when it's done .
ArtsAcoustic CL Series released
-
nick at artsacoustic nick at artsacoustic https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=65180
- KVRist
- Topic Starter
- 441 posts since 14 Apr, 2005
-
- KVRAF
- 1622 posts since 10 Jul, 2006 from Paris, France
Thanks, already installed the new demo .nick at artsacoustic wrote:New versions are up.K-Slash wrote:Thanks, please bump this topic when it's done .
By the way what about what Vitaly said ?
-
- KVRian
- 1115 posts since 2 Oct, 2001 from Berlin, Germany
Hm...Vitaly wrote:Hi,
I don't want to spoil a party, but apparently this compressor isn't 'opto' at all.
First, we all know that the main feature which makes a compressor opto is a dependence of it's release time on incoming signal length. It's not there.
Having done a couple of opto-based modeling projects myself (and built 'real' opto comps as well), I can only partly agree.
True, all LDRs have a certain system-immanent intertia. The memory effect is always present to some degree.
However, this effect is greatly dependent on the actual LDR. Especially, when you have types of longer attack & release times, the inertia can easily be a multiple of the 'regular' envelope times (whatever these are, as every manufacturer measures that differently). But when the system's response times are generally very low (there exist also LDRs with just a couple of msecs around), the inertia is of minor importance, simply because a developer has to choose traditional circuits such as RC networks in order to obtain variable attack & release settings.
I do not know the underlying circuit AA modelled here, but I just guess here they've chosen a fast device (perhaps for the sake of quick transient response) and modelled the rest in a plain vanilla-engineering style. Nothing wrong with that.
Again, this is also a great oversimplification. It's true when you arrange the LDR inside a voltage-divider network that attenuates the signal path (R = R1 / (R1 + R2)). You can think LA-2A here, this is how it works there (I've built similar stuff where my LDR resides in the negative feedback path of an OP amp. Such a thing works smooth as butter). As a voltage divider, an LDR/opto coupler/vactrol is always a soft-knee device. It's in the nature of things, as you can't achieve sudden or infinite steepness / slew rates with passive elements per se.Vitaly wrote: Second, Opto compressor has a soft knee. It's not there.
But:
More complex detector circuits as found with general-purpose or even mastering-grade compressors and limiters typically include level/threshold detectors (comparator circuits) and active rectifiers (remember through-zero errors of passive diode networks). Such circuits usually don't depend that much on 'system errors' such as the nonlinearity of an opto-coupling device. A clever developer will outweigh the benefit of such a component against the versatility of a fully-parametric design that can be built around such a device.
You may look at the schematics that Silonex post on their web pages, such as this one:
http://www1.silonex.com/audiohm/compressor.html
(note that the detection circuit in fig. 4 is a feed-back topology, as opposed to an LA-2A, afair. But the main difference is that D1 & Q1 do the level-sensing here, hence the hard-knee transfer curve. Feedback detection makes that a bit softer again, btw.).
Again, I have no clue what AA actually did, but I can imagine they wanted maximum versatility and 'a bit of' character involved.
This is also dependent on the actual detection circuit. You can add anti-log sensing pretty easily with just 1 OP and 1 diode as to compensate for the light-source's nonlinearity, and do just the same within the actual gain reduction element. Technology is our friendVitaly wrote: Third, it seems CL1 uses a linear compression curve instead of a traditional logarithmic. It might help you to save CPU usage but it is far from what real opto compressors have.
Sascha Eversmeier
drummer of The Board
software dev in the studio-speaker biz | former plugin creator [u-he, samplitude & digitalfishphones]
drummer of The Board
software dev in the studio-speaker biz | former plugin creator [u-he, samplitude & digitalfishphones]
-
nick at artsacoustic nick at artsacoustic https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=65180
- KVRist
- Topic Starter
- 441 posts since 14 Apr, 2005
Hi Vitaly,Vitaly wrote:Hi,
I don't want to spoil a party, but apparently this compressor isn't 'opto' at all.
First, we all know that the main feature which makes a compressor opto is a dependence of it's release time on incoming signal length. It's
not there.
Second, Opto compressor has a soft knee. It's not there.
Third, it seems CL1 uses a linear compression curve instead of a traditional logarithmic. It might help you to save CPU usage but it is far from
what real opto compressors have.
So, why did you call it 'opto'?
First off, you're right. Back in the time photoelectronic compressors/limiters did work like you describe.
However, i think it's time to clarify some stuff here:
We did not emulate a photoelectronic compressor in summary, but we based the crucial element, the envelope detector, on a potocell basis. The heart of an optoelectronic design back in the time was on one hand, to realize a discret signal flow which is not suffering from noise added from the components, but the other hand was way more important: with a photocell you're able to get a very smooth and unique sounding envelope detection. This to our ear is the real key to the great sound of an optoelectronic compressor. Not to be confused, i'm speaking of the colour that exactly the envelope detection does, to our ears was the key factor. The negative side effect like memory effect of a photocell and due to that the uncontrollable program dependency, as well as the complete absence of fast attack and release times was common to the function of photoelectronic compressors, but are not really the positive aspect of an optoelectronic layout in our opinion.
An example: you have a hard overshooting transient, which is just a fraction of time related to the signal length. In a photoelectronic compressor you'd end up with a release that ducks the most important part of the signal after the spike, due to the memory effect of the photo cell. Negative side effect, in our opinion, as it is not really controllable.
Also, just because it's implemented in most optoelectronic compressors that way, it doesn't mean, that this design is only capable of a soft knee. A photo cell, when driven within it's specified range, is able to deliver a precise knee, very similar to a hard knee. We've implemented a semi hard knee, which is close to a hard knee.
So in summary: For us the crucial factor in an optoelectronic design was the sound of the envelope detector. To our ears that's the hidden feature, why some optoelectronic compressors sounded so good, as opposed to the public opinion that the programdepedancy or the slow attack times are the key factor of the unique sound. In the digital world in this case it's rather easy to overcome the real disadvantages of such cirquits, so we've modelled only the crucial parts from that design, not the negative side effects.
This also belongs to the rest of the components in our compressors. Only what does benefit to the sound was closeley modelled, if there's negative side effects, we did correct it to what we think it should do, or we left them out completeley.
Btw, our compression curve is non-linear.
Edit:
Sasha has beaten me by lenght with a way more detailled explanation. Thanks sascha!!
-
- KVRist
- 201 posts since 2 Jan, 2003
Thanks Sascha and Nick!
Nick, please consider 2 inputs (without ex. side-chain input) version to run in Wavelab.
Cheers,
Vitaly.
Nick, please consider 2 inputs (without ex. side-chain input) version to run in Wavelab.
Cheers,
Vitaly.
-
nick at artsacoustic nick at artsacoustic https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=65180
- KVRist
- Topic Starter
- 441 posts since 14 Apr, 2005
-
- KVRist
- 201 posts since 2 Jan, 2003
-
nick at artsacoustic nick at artsacoustic https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=65180
- KVRist
- Topic Starter
- 441 posts since 14 Apr, 2005
-
- KVRian
- 1290 posts since 13 Mar, 2007
nick at artsacoustic wrote:
Btw, our compression curve is non-linear.
You're refering to above the knee/threshold ; correct .?
2-21-2010 My correction.
NO , you where talking about the timing curves of the attack and release;how the volume envelopes, I.E. gain reduction and gain restoration are implemented!!!
the equivalent of the fader on a channel strip or even a poentiometer on a guitar !! : wheter he resistance in linear or exponential in relation to the mechanical travel.( or virtual , temporal function in a comps level control)
2-21-10 I got crossed up by terminilogy and was thinking about Transfer Curve here:
That's one of those "holy grail " properties that is celebrated in some mythical old hardware ( which was actually a limitation of the state of the art components in analog electronics back in the day ).
The ability to use faster attack times but still let some of the higher signal level through ( transients) without squashing the crap out of the very first transient part of the signal .
Yes?? NO
I'll be putting these through more indepth testing tommorow. love the design , controls and Concept !!
Cheers
Last edited by contrary on Fri Jan 22, 2010 4:06 am, edited 1 time in total.
Financial solvency and KVR Mix as well as oil and water.
-
nick at artsacoustic nick at artsacoustic https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=65180
- KVRist
- Topic Starter
- 441 posts since 14 Apr, 2005
-
- KVRist
- 297 posts since 15 Apr, 2004
When trying the CLMS1 demo(1.06) in Reaper 3.15 i get a crash as soon as i select the "smack me" preset.
I also get a error when just scanning the plugin the first time.
//L
I also get a error when just scanning the plugin the first time.
//L
-
- KVRist
- 194 posts since 27 Apr, 2005 from ATL
I'm not getting any crashes or errors with clms1 (full version 1.06). Btw, I'm using Reaper 3.161.lazerkind wrote:When trying the CLMS1 demo(1.06) in Reaper 3.15 i get a crash as soon as i select the "smack me" preset.
I also get a error when just scanning the plugin the first time.
//L
Intel CoreI7, ASUS mobo, 32GB RAM, RME Babyface, Win10 64-bit