Limiter with "Zero" attach time?
Hi Experts,
I am encountering a serious problem in my sigma DSP project.
Can I get zero attach time limiter in sigma studio for ADAU1701 ?
There are 2 type limiters in sigma studio, one is RMS limiter the other is peak compressor
I found it is very hard to limiter a signal without heavy distortion even I set the attach time to max 10000.
I try to use below peak compressor, but it can't limit the output signal to a fixed threshold level, that is say,when input level increase,the output will follow increase as well. And it will stop to work when input signal above +6dB.
Sir,
How can I do to get a limiter with attach time less than 2ms ?
Thanks a lot.
Audio Signal Processors
SigmaDSP Audio Processors
adau1701
Dear Sir,
Any suggestion?
Thank you
Moving discussion to the SigmaDSP forum for follow-up.
-Joe
Hello,
One way to achieve a fast limiter is through a feed-forward approach. This works by measuring the incoming peaks to instantly calculate (instead of settling upon via feedback) the required gain reduction. An example is shown below:
The absolute value and AB blocks select the higher peak level of the two channels to insure that neither goes over the limit. The clipper forces a normal gain of 1 when below threshold. Then the peak envelope block measures this peak level. Feedback around the external-decay version of this block provides an exponential (constant dB rate) decay -- this technique is described at Workaround: Creating constant-dB envelope decay using an existing linear decay algorithm
The reciprocal of the resulting level value is applied to slew volume controls to limit the audio level. A short look-ahead delay is shown, but it may not be needed.
With a rising musical input, we see the gain (top graph) dive below 1.0 as needed to keep the level (bottom graph) below 1.0 peak at all times. Unlike limiters that close a feedback loop around the output volume, this one doesn't overshoot its correction. Thus it doesn't sound as bad on music as most others. I tried two other designs that bombed terribly before this one -- it wasn't easy.
Best regards,
Bob
Hi @KJBob,
I have the exactly the same problem but on the 1452. I have tried this algorithm, but have difficulty on the holding time tuning as I have used the same values as the one above. Can you give a working sample for the 1452 version?
Thanks is advance!
Dan
Hello Daniel,
Try the attached project, making adjustments as needed for your application:
Limiter Recovery Rate -- the rate at which gain returns to normal after a loud sound is gone
Control Attack-Decay at the Slew Volume -- set it very fast to get the desired "zero attack time" --
for setting time constants, see https://ez.analog.com/message/165526#165526
Look-Ahead -- Increase to eliminate any overshoot on sudden inputs. SigmaStudio seems to have trouble saving this setting with the project.
Headphone Volume -- This is just to reduce headphone level for my tests.
Limiter Level Adj -- Increasing this gain beyond unity reduces the level at which the limiter cuts in.
Best regards,
Bob
Dear Bob,
A million thanks to you!
Dan
Please log in to reply.







