Transparent, but affordable limiter - does such a beast exist?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

thermionic

Well-known member
Joined
Jun 3, 2004
Messages
1,671
Hi,

Some friends of mine have a club / venue installation. Instead of using a PA system they have have an overgrown hi-fi, using hi-fi carts on the turntables and Levinson amps etc. Rather than risk any visiting music selectors blowing the K-horns, they have a Drawmer limiter prior to the x-over; transparent, it ain't. In an ideal world they'd buy a GML 8900, but that's a lot of money, and it's way over-specced in terms of features considering they purely want to use it as a limiter. What would you fit? Is there a DIY project that's uber-transparent?

TIA

Justin

edit - you're going to suggest the PICO limiter, aren't you, right?

edit again - I now see PICO PCBs are unobtainium... Hmm...
 
thermionic said:
edit - you're going to suggest the PICO limiter, aren't you, right?

edit again - I now see PICO PCBs are unobtainium... Hmm...

Uhh, huh. And uhh huh again.  They are being sold commercially as the FCS P3S. http://www.mercenary.com/fcs-p3s.html
 
Hi,


  you might find a pair of NTP 179-400 limiters. I used to lend them to  a chum who used them on a club system. he was blown away. I use mine in the studio daily, and they are very transparent, and still quite cheep! They will probably need recappping. they've got those horrible Erie(?) red plasticcy electrolytics.


      Kindest regards,

    AndyP
 
Why not design your own.

The PICO is designed around a chip the does it all, the THAT 4301; detector and vca and three utility opamps in one chip. You just need the surrounding balancing I/O circuitry which is just a THAT 1246 on the input and THAT 1646 on the output, control circuitry (very basic) and you're done. You can pull everything directly from the 4301 demo datasheet.

http://www.thatcorp.com/datashts/4301demo.pdf
http://www.thatcorp.com/datashts/4301data.pdf

Why people think the PICO is so magically delicious I don't really understand.  ???

Mark
 
I guess maybe you haven't compared a Pico to all the other 4301 variations on the market.   I haven't heard another with so much smooth control.  I'd be happy with no dynamics but 12 of them.  If anyone cares to buy my pile of vintage tube limiters at market prices, you can have them.   You can buy my Distressors too. 
 
emrr said:
I guess maybe you haven't compared a Pico to all the other 4301 variations on the market.   I haven't heard another with so much smooth control.  I'd be happy with no dynamics but 12 of them.  If anyone cares to buy my pile of vintage tube limiters at market prices, you can have them.   You can buy my Distressors too. 

Quite an impressive statement...luckily people also use compressors for other reasons as well. There are some impressive compressors in existence that is not up for commercial grabs...ask me how I know this.
 
Biasrocks said:
Why people think the PICO is so magically delicious I don't really understand.  ???

Because it sounds really good. So does the DAOC. Both are as good as the best commercial units. Both are in the "clean" camp. The DAOC has cool tubes and transformers but in a blind test I could probably pick out the type of gain reduction element faster than the amplifier type. I like clean.
 
So really, it's the THAT 4301 analog engine that is responsible for the sound, since the PICO is  culled from the data sheet implementation with the addition of THAT IN/OUT balanced line receivers/drivers.

Mark
 
It's pretty clear to the ear that it's the surrounding design values and parts selection that control the effectiveness of the chip.  If it wasn't, then anyone could hack out the sample circuit with equal success. 

I thinks it's silly to think that ANY sample circuit provided by a chip manufacturer is the final word on implementation. 
 
There are two contributory factors to transparency of a dynamics processor. First the linearity of the path while operating at a constant gain, and secondly how you vary the gain, which is what dynamics processors do.

The latest generation THAT corp VCAs while not perfect, are IMO good enough to not degrade transparency, and one must ask compared to what? The alternate gain elements are generally not more linear, and if preferred for some specific coloration they add, that is not consistent with a pursuit of transparency.

Take as a given that we are manipulating the gain, in response to signal level changes. Any gain changes that are not very very slow, will multiply the primary audio signal and create modulation products in the audible frequency range.  In very simple terms, if you want a limiter that doesn't cause audible artifacts, you must keep the gain changes small and slow, or somehow mask them so they are concealed behind the primary audio signal. Even if you can tolerate a simple slow attack, a very slow release can also be audible, when the primary signal drops off suddenly, and you can then hear the noise floor slowly ramping.

This is a catch 22 for dynamic design, you want to alter the gain quickly enough to prevent overload, and audible pumping, but without distorting the signal or being apparent when working. 

There are multiple approaches in how we treat the side chain rectification, and smoothing of the level measurement for gain calculations each with their own pros and cons; half wave, full wave, peak, average, RMS, attack, release, hold, adaptive, soft knee, and dare I say, techniques not yet tried, while many have over the decades.

IMO there is no single simple answer and it depends on your personal subjective criteria, and program demands.

The THAT corp VCA and rectifier chip is a very respectable starting point. Use the data sheet as a beginning, but don't expect it to reveal the best possible final solution. While you can probably get 80-90% there without too much heavy lifting.

Since there is no perfect solution, pick your poison, based on your willingness to invest time to study the prior art, and maybe invent a little something new. It is also useful to have a rudimentary understanding of how our hearing adapts to level shifts, as there are some time windows that can be used advantageously to mask gain changes surrounding signal level shifts.   

JR
 
I think I already mentioned researching the prior art.  In addition to THAT app notes, there are several old AES papers (pre-prints)  written by DBX engineers over the decades, that are worth looking at.

There have been discussions about the PICO design over on Wayne's forum. Roger has become more defensive about sharing after turning the PICO into a commercial product. That is completely understandable, it's bid'ness. 

If you want a PICO buy a PICO, if you want to design your own "transparent" limiter,,, do some more research and melt some solder. You should gain far more satisfaction from tweaking out your own version, than copying someone else's work effort.

JR
 
DBX could build something as good as the Pico in the sub-$1K range, but they haven't.  Yes, in theory, anyone willing to spend the time might theoretically do so.  I know Roger has gone through countless revisions over the years.  The latest commercial versions differ in many ways from the final DIY offerings.  Which ties another nice little bow on this conversation; the Pico can't be had as inexpensive DIY anymore, so it's a moot point.  I apologize for those of us who mention it wistfully.    It does handle high amounts of gain reduction with very little change in sound.  I haven't used anything else that can do so much with so little negative impact, anywhere near it's commercial price.  Color limiting is a separate discussion, and off topic.    I've hammered acoustic folk music into a stereo Pico when I needed a loud live CD that the client could walk away with immediately, and you really can't hear the limiting in action.  It is more 'mastering grade' than many tools marketed at that niche, and we have multiple mastering houses present here at PP who use it. 

So back on topic, what else is an option for Thermionic?
 
emrr said:
So back on topic, what else is an option for Thermionic?

I hardly think designing something even directly from app notes would work out to be less money than a GML 8900 if he bills for his time. I think the suggestion of the NTP's is a very good one. They come up all the time on ebay for not very much money.
 
emrr said:
DBX could build something as good as the Pico in the sub-$1K range, but they haven't.

emrr said:
I know Roger has gone through countless revisions over the years.  The latest commercial versions differ in many ways from the final DIY offerings.

emrr said:
the Pico can't be had as inexpensive DIY anymore, so it's a moot point.

emrr said:
It does handle high amounts of gain reduction with very little change in sound.

emrr said:
I haven't used anything else that can do so much with so little negative impact, anywhere near it's commercial price.

emrr said:
I've hammered acoustic folk music into a stereo Pico when I needed a loud live CD that the client could walk away with immediately, and you really can't hear the limiting in action.

emrr said:
and we have multiple mastering houses present here at PP who use it.

emrr said:
I guess maybe you haven't compared a Pico to all the other 4301 variations on the market.  I haven't heard another with so much smooth control.  I'd be happy with no dynamics but 12 of them.  If anyone cares to buy my pile of vintage tube limiters at market prices, you can have them.  You can buy my Distressors too.

The most annoying thing about the "pico" is the frequent viral marketing hype on this, and other forums by the same three people.

The bottom line is, the 4301 is the easiest (and cheapest) way to DIY a limiter. Just do it. You can tailor it to your specific needs by pre-setting it to what you need. 10:1 pre-set ratio... Threshold trimmer instead of a front panel control would keep people with itchy trigger fingers away from resetting it. quick attack and release, etc... and now you have what you need for less than 40 bucks and a chassis.
 
drpat said:
The most annoying thing about the "pico" is the frequent viral marketing hype on this, and other forums by the same three people.

The bottom line is, the 4301 is the easiest (and cheapest) way to DIY a limiter. Just do it. You can tailor it to your specific needs by pre-setting it to what you need. 10:1 pre-set ratio... Threshold trimmer instead of a front panel control would keep people with itchy trigger fingers away from resetting it. quick attack and release, etc... and now you have what you need for less than 40 bucks and a chassis.

I would classify EMRR as a very enthusiastic customer, not a viral marketer. While I doubt Roger would discourage him from posting his accolades.

I suspect the results are replicable. The technology around dynamics processing are pretty well known and modern VCAs are good enough to be a non-issue (that wasn't always the case). The THAt corp chip set and app notes will get you most of the way there. I recall having to roll my own log conversions and other heavy lifting to make a simple comp, back when.

So +1 to just do it...  unless you don't want to..

JR

 
Rodger got a nice start here with people supporting him at first ,
there's always the FMR RNC compressor
find a used one , pretty decent nevermind the good price
 

Latest posts

Back
Top