"Look Ahead" analog limiter?

I was wondering if anyone knows of any designs for a limiter similar to the one included in the Safe Sound Audio P1.  I was thinking of an idea for a headphone amp monitor mix device and this would be perfect to make sure that the tracking input doesn't clip.  Kinda like a clip saver for tracking with a nice headphone monitor section.

Morning_Star said:

I was wondering if anyone knows of any designs for a limiter similar to the one included in the Safe Sound Audio P1.  I was thinking of an idea for a headphone amp monitor mix device and this would be perfect to make sure that the tracking input doesn't clip.  Kinda like a clip saver for tracking with a nice headphone monitor section.

Click to expand...

??  A look ahead "analog" limiter. How do you propose the time travel in analog domain to see ahead?  Tape loop with two heads, record player with two needles?

JR

"Audio Design F601-RS" has optional delay line in signal chain.

JohnRoberts said:

How do you propose the time travel in analog domain to see ahead?

Click to expand...

I was wondering about that myself. According to reviews of the Safe Sound Audio P1 they vary the threshold/ratio depending on the slew rate of the signal when it exceeds a certain level (-6dB?) below the set threshold.

JDB.
[which, I'm sure, trades one bag of artefacts for another one]

JohnRoberts said:

??  A look ahead "analog" limiter. How do you propose the time travel in analog domain to see ahead?  Tape loop with two heads, record player with two needles?

JR

Click to expand...

That's why I put it in quotes.  I know that this is not possible in the analog world.  But the Safe Sound P1 is apparently the closest you can get with the design it has.

>>>> How do you propose the time travel in analog domain to see ahead?

One word: FLUX CAPACITORTM. Period.

This is no magical way to anticipate the future from the present.

I am aware of at least one professional dynamics processor (RANE noise gate) that uses a delay line to buy some look ahead time. In recording there is no real need to be real time so I don't see why you couldn't use digital recording software to generate a duplicate track offset in time.

Send the prelayed track to the side chain ahead of the normal track you want to limit, and presto you have a faster than real time limiter...

Another old studio trick is to play the track backwards before processing so the fast attacks are now all in the tails, and the attacks (old tails) are much more friendly.


JR

just don't track too hot, it's digital, your mixes will sound better if you peak at -6

nicholas

Man, is everyone really not reading my post?  I don't mean that it is a real look ahead limiter.  I was just wondering if anyone knew of any designs that function like the Safe Sound P1 limiter section.  I wanna use it for a clip protector for tracking.  

A Feedforward Controlled Delay-Line Limiter

The attack time in a conventional limiter must always be a compromise between distortion and overshoot on transient signals. We describe the design of an inexpensive and compact limiter which overcomes these problems by introducing an analogue delay line in the signal path. A novel feed-forward control system has been adopted to provide accurate reciprocal control at all signal levels.
Authors: Evans, C. J.; Dawson, J.
Affiliation: Amplification and Recording, Cambridge, England
AES Convention:55 (October 1976) Paper Number:1162

>>>>>> Another old studio trick is to play the track backwards before processing so the fast attacks are now all in the tails, and the attacks (old tails) are much more friendly.

old trick for new dogs (daw jockeys) is to limit reversed (backw.) track to fake variable attack itb (your limiter/comp must be on auto-release setting).

But, somebody could come up with a DIY-vibe trick - i.e. how to ab-use the soundcard/converter latency to give you some "lookahead" time.

Why not?

The Safe Sound P1 doesn't create a delay.  That's what I'm interested in.  Why is there so much talk about everything but that limiter design? 

I understand how it would function with a delay.  But I'm looking for a design like the one in the P1.

Perhaps because the problem is pretty well understood.

A hard limiter close to clipping has to use fast attack because you don't have much room for overshoot. There has been research into the duration of transient distortions that are tolerable before they become audible (as distortion), so clamping overshoot for some modest time, and actually releasing quicker at first will be less objectionable than truly fast attack and slow release..

You need to throw it on the bench for the type of music you want to process and see what works least badly.  Different genres may have subtle differences.

JR

I think there are both GE (mid '50's) and Langevin (late '40's) tube broadcast limiters with microscopically tiny analog delay lines built it to achieve semi-look-ahead timing.    But that doesn't address your question either. 

The P1 quote says true 'look ahead' limiting but with absolutely no latency.    All that says is that the delay is short enough to be imperceptible as latency.  There has to be delay, if it's 'look ahead'. 

I must say the sales blurb does not make much sense (will control overshoot to 0.5dB...).
They may actually use a clipper which takes care of the initial attack; as soon as it happens, this is used to trigger a very fast side-chain. Behringer used to do that in one of their units (I don't know who they stole it from...).

jdbakker said:

According to reviews of the Safe Sound Audio P1 they vary the threshold/ratio depending on the slew rate of the signal when it exceeds a certain level (-6dB?) below the set threshold.

Click to expand...

That would be a way to predict peaks without needing a delay line.

Full article: http://www.soundonsound.com/sos/oct03/articles/safesound.htm

JD 'WOM?' B.

Come to think of it, I ran into this exact problem for a power amp design I did back in the '80s.

I designed a modest 35W continuous power amp with a short term boosted tracking power supply, that allowed me to put out 2x the voltage swing, or 4x power- 140W, on transients (AMR PMA 70+) .  The difficulty was only opening up the boosted rail when the signal was close to clipping for good efficiency, but getting the boosted rail open fast enough for fast transients without clipping the leading edge.

My solution was adding some HF lead in the side chain, so it would respond quicker or actually over respond to HF... I suspect you could do the same thing in your limiter by boosting HF in side chain... You will get too much limiting for pure HF tones, but they don't occur very often in the real world.


JR
 

Do you mean something like a frequency modulated sidechain? HF bias in tape recorders?

No, a few dB of simple HF boost....  Season to taste. I don't recall exactly what curve I used since that was 20 years ago..

JR

http://home.btconnect.com/ssa/whitepaper/whitepaper.htm