Resampling Issue

Here’s one I haven’t seen before:

I have a track mastered at 48kHz with an output ceiling of -.30 dB on the WL Master Level as per the AOM Invisible G2 Limiter.

As I was about to bounce the final mix of the track, I engaged the built in Wavelab ReSampler at 44.1 kHz and my Master Level meters shot clear into the red! Not all the time, but enough that the track’s previous output ceiling of -.30 dB is comprimised and I cannot use the resampled (and now distorted) track.

What’s even more baffling, is that when I limit with the FabFIlter Pro L in the exact same way, the Resampling issue is totally non existent and the WL Master Level meters remain at -.30 dB. So what is happening between these two plugins that is causing WL 9.5 to smoothly process one but add gain to the other?

Any advice? I would hate to have to give up the AOM Limiter because WL is choking on it!

Thanks - DL

The resampler is going to create higher peaks, but you might find the difference between the 2 limiters output if you remove the resampler and render each to a temporary file (unnamed file), then run global analysis on the renders to see the true peaks and digital peaks in the 48k temporary renders.

Thanks Bob, but what’s so baffling is that the FabFilter Pro L produces files which keeps the level EXACTLY the same when resampling…always has… but the AOM Invisible G2 produces files which don’t respond well at all to being resampled. That’s the mystery to me. They “do” the same thing, but something is getting baked into the AOM’s file which causes major issues with Resampling

Why?

That might be evident in the Analysis. The FabFilter 48k temp render might analyze a bit lower than -0.3 before resample. The Invisible Limiter render might analyze at -0.3. (If you get what I get, doing a quick defaults kind of test with -0.3 ceiling on both here). Try it and look at the 48k temp render digital and true peaks before resample.

I have the Pro L (although not used much) and it is possible to produce true peak limited outcomes: if you have selected oversampling and/or ISP metering.

I am not sure if the AOM does the same … meaning that it is likely that the resulting files will have true peaks over the threshold selected.

The problems lies in the invisble G2limiter. To be honest… it sucks when it comes to stabilty and other things (try bypassing it, it still processes!). One of the biggest bugs is that even though oversampling is on and everything is set the way it should for preventing ISP’s, it still produces a large amount of ISP’s.
A waste of money if you ask me…

Yes, different limiters and their settings (oversampling, ISP detection, etc) will produce different results after resampling. It’s normal.

It’s also dependant on the material and how loud/hard it’s pushed.

I would say in most cases, -0.3 ceiling might not be enough to survive a resample without clipping to some degree, not to mention the eventual lossy encoding the master WAV will probably encounter down the line.

I’d pick a safer ceiling level such as -0.5 or even -1dBFS.

It will also depend if you are analyzing true peaks or digital peaks. True peaks will read slightly higher as well.

I use a variety of limiters for different projects but I almost always follow the preferred limiter with Limiter 6 and only use the ISP module (the rest are turned off and hidden) as an extra safety net against intersample peaks. It’s a very good true peak limiter, works well in WaveLab, and it has a delta mode so you can hear only the peaks that it’s catching.