The quest for inter sample peak limiting

[stinkyfingers]

Elephant

edit: if you use the massey limiter @ 192k then convert to 44.1k, you will get ISPs...

[sat159p1]

I found iZotope resampler to be very good one. And about ISPs, the Barricade limiter never let any peak above the ceiling and I use it frequently.

[sanddigger1]

VladG free Limiter n°6' last stage is very effective at suppressing ISPs. I'm using it often after AD clipping or Ozone 5 limiter. In my opinion, it changes the sound a lot less than using Ozone in oversampling mode.

[WBM]

Elephant or ozone 4 or 5 good intersamples catchers

[jimdrake]

Originally Posted by stinkyfingers

edit: if you use the massey limiter @ 192k then convert to 44.1k, you will get ISPs...

true, but i was guessing more along the lines of fabfilter (approx 0.5dB) rather than the massey on its own at 1x sample rates (more like 3dB).

all the other suggestions do't work on pro tools, except ozone, haven't looked at that in a while.....

[Alexey Lukin]

Originally Posted by sat159p1

And about ISPs, the Barricade limiter never let any peak above the ceiling and I use it frequently.

Barricade is not reliable for ISPs in my experience: it creates ISPs of +0.83 dB in ISP mode on a sine sweep when it passes ~15 kHz.

[Babaluma]

Again, Elephant with 4 x OS is GREAT at catching 99.9% of all ISPs.

[bendermastering]

Also check out Elixir from Flux::. It was just updated today/yesterday, finally at 64bit plus some other performance improvements.

[Joe_caithness]

I used Elephant in OS x4 and never have had an intersample peak.

[huejahfink]

TC Brickwall Limiter set to -0.1, upsampling mode. Really tight on ISP!

AFAIK upsampling is done on the sidechain for the detection circuit, and not on the audio itself, making this a pretty transparent way to remove ISP

[neilwilkes]

This does the trick for me

NuGen Audio ISL Inter sample True Peak Limiter

[Laurend]

From a EBU paper about R-128, here are le potential errors regarding the over-sampling used for ISP detection:

ISPOver-sampling ratio
4
8
10
12
14
16
32

Under-read (dB) max fnorm = 0.45
0.554
0.136
0.087
0.060
0.044
0.034
0.008

Under-read (dB) max fnorm = 0.5
0.688
0.169
0.108
0.075
0.055
0.042
0.010