True Peak way above 0dB even with limiter applied?

[plexuss]

Before True Peak was a thing, the goal was -0.3 DBFS not -0.1 and now in the era of streaming platform, the goal is -1 DBFS.
If you need an explanation look for SoundOnSound article ..

While this whole True Peak limiting thingie is nothing but a huge hoax to make you guys buy new stuff!!! Nothing but a theoretical issue nobody was ever able to prove audible for the slightest amount... This only proves the power of clever marketing ... like with all politics... stating something without any evidence and as soon as this statement was repeated enough, everybody jumps onto this lie and defend it religously... silly world!

This is the kind of false information that will get you banned on twitter. lol In fact, there has been studies done on this and it's one of the reasons why it exists in many IEEE and AES et al audio standards. The research shows that hard clipping causes wide-band intermodulation distoriton that is audible. And, that many up-stream processes can cause ISPs that will hard clip, such as codec processing and DAC conversion. This is where the "leave 1dB headroom if you compress to 128k streams" comes from. The research work and final reports are on the internet somewhere. I have them archived away. If you want I can dig them up or you can do your own research. But, it's real, it's audible, it sounds like shit and we should mitigate it in order to create good sounding audio and music. And stop spreading false information - do research before you post what you believe is a fact. Beliefs are not facts.

[The Noodlist]

Pre fader/post fader has overs during limiting for me in S1 3 Artist depending where the TP limiter is placed.

[Trancit]

Before True Peak was a thing, the goal was -0.3 DBFS not -0.1 and now in the era of streaming platform, the goal is -1 DBFS.
If you need an explanation look for SoundOnSound article ..

While this whole True Peak limiting thingie is nothing but a huge hoax to make you guys buy new stuff!!! Nothing but a theoretical issue nobody was ever able to prove audible for the slightest amount... This only proves the power of clever marketing ... like with all politics... stating something without any evidence and as soon as this statement was repeated enough, everybody jumps onto this lie and defend it religously... silly world!

This is the kind of false information that will get you banned on twitter. lol

And I love it!

If you want I can dig them up or you can do your own research. But, it's real, it's audible, it sounds like shit and we should mitigate it in order to create good sounding audio and music. And stop spreading false information - do research before you post what you believe is a fact. Beliefs are not facts.

I don´t mind but thank you anyway...

Nonetheless, that just means they didn´t do their codecs properly...

Noone can ever tell me that in nowdays music when listening as a wav file some intersample peaks will be audible at all against all of that distortion that was added on purpose to win the loudness war...
Everything else is (if there can be something audible via streaming) the job of the streaming platforms...
If the algos are too shitty that they need their crappy intersample peak limiting than it´s up to them to do it and not up to the artist/producer whatever...

[pixel85]

I would be happy to see some case study on a large group of casual listeners (thousands) listening through casual sound systems /
headphones, proving that they can hear inter-sample distortion and what's more important, it negatively affects their perception of music.
So far I could only read technical descriptions, regulations etc. - academic talk. But I couldn't find a study proving (or disproving) that inter-sample distortion has a real major impact on casual listener's listening experience.
I'm pretty sure that such a study would bring consensus that casual listeners don't give a flying f***k about it. They don't complain about music massacred with >10dB GR limiters, listened on iPhone/BT speaker or cheap headphones. It's hard to believe that microscopic level of inter-sample distortion would make them enjoy music listening less.

[kvaca]

Noone can ever tell me that in nowdays music when listening as a wav file some intersample peaks will be audible at all against all of that distortion that was added on purpose to win the loudness war...

[plexuss]

I would be happy to see some case study on a large group of casual listeners (thousands) listening through casual sound systems /
headphones, proving that they can hear inter-sample distortion and what's more important, it negatively affects their perception of music.
So far I could only read technical descriptions, regulations etc. - academic talk. But I couldn't find a study proving (or disproving) that inter-sample distortion has a real major impact on casual listener's listening experience.
I'm pretty sure that such a study would bring consensus that casual listeners don't give a flying f***k about it. They don't complain about music massacred with >10dB GR limiters, listened on iPhone/BT speaker or cheap headphones. It's hard to believe that microscopic level of inter-sample distortion would make them enjoy music listening less.

First, a single peak clip is not going to audible. the issue is high "clip density" where there are more clips in a period of time. If there is a lot of clipping, no double blind test is necessary because the level of wide-band intermod distortion will be plainly evident. Second, human hearing and perception are complex and difficult to quantify. Mostly we go by generalizations, gut feelings and hope. And, the knowledge that not everyone is going to agree on the quality of the sound. Third, measurable poor audio quality will be percieved by some as sounding good. This is related to my second point. Even audio professionals are prone to this, hence the concept of the loudness war and more focus on loudness management. This is why, for those that care, having a way to measure audio as a way to optimize quality is important. But, yes, some people are not going to care.

[foosnark]

First, a single peak clip is not going to audible.

Generally this should be true -- but some codecs do not handle it well. I once had two tracks professionally mastered, and they went to -0.5dB TP. There were very audible clicks when listening over Bluetooth headphones, which disappeared with wired headphones. (They also disappeared after encoding to 256kbps MP3, even with Bluetooth.)

Anyway... I haven't used Ozone, I usually use Elephant. On one recent track though it let peaks through that were well above what I asked for. I switched to Barricade and that did the job just fine.

[plexuss]

If ISPs are something you prefer to mitigate, look for a tool that gives you a sense of ISP density such as Youlean or the LVC meters. They show ISPs over time and help to identify areas of high ISP density (quantity of ISPs over time). Ideally you should not see any ISP warnings.

[exist01]

Have you tried it with the other IRC modes (esp. III and IV)? Also, I recall that the manual recommends setting your ceiling to -0.3 dB or lower, but I don't recall the reasoning for it. But, I've also noticed with Ozone in the past that setting the ceiling to -0.1 dB resulted in many "overs" and never really understood why. I guess I figured it had something to do with the speed and detectability of some transients.

I just tried the other modes, with the ceiling at -0.3 dB, and they are still producing about the same True Peaks, about +4.5 dB during the loudest parts.

it might also be just a single "fantom" peak/transient, that occurs in one single place in the whole song, and its so short in time it would appear only as a very thin line on the waveform, which can be somewhat hard to see i guess.
but if you render the wave out, and then normalize it; it would suggest its bringing that "fantom" peak down (which could be a slight click, pop, or something like that - it may not even be audible at all; but still "counts" as a peak/transient.)
if thats the case which is likely, i would suggest just put a clipper before the limiter, it will simply cut it out before hitting the limiter and you wouldnt have to first render then normalize.
a "fantom" peak can be caused by a bug in a plugin or similar, a hardware issue or just something went wrong somewhere.

if you look at the youlean, while playing the song - does it register the +3db Over at one single place in the song, or is it like, constantly hitting these numbers. because if it is, then its not just a fantom peak..
a screenshot of the full waveform would be cool.

The song plays at +2-5 dB in a lot of places, so it's not just a few individual peaks/transients. The numbers are quite consistently going much higher than 0 dB.

Here are some waveform screenshots: (ignore the the little orange thingies at the top, they are just some markers I've set in my DAW)

Post-Limiter:

SS_PostLimiter.jpg

Pre-Limiter: (this one doesn't peak over 0, just thought I'd show it in case it was relevant)

SS_PreLimiter.jpg

There do seem to be few bigger spikes in dB (which aren't all that audible), but the track as a whole is peaking way above 0 pretty much from the get-go, so I don't think those individual spikes are the sole issue.

Might be a stupid question, but did you hit the red X on youlean meter to reset the true peak meter?

Yes I made sure to reset it before/after each listen.

Are you seeing this in Cubase by any chance ?

It is in FL Studio.

[astralprojection]

the post limiter wav seems pushed beyond 0dbfs indeed, but the pre one looks fine. that doesnt make much sense that a limiter in itself would push it above. just from the wave alone it looks like it was pushed above 0dbfs after limiting. and before it hits youlean. just triple check faders or any +gain i guess.

[pixel85]

I would be happy to see some case study on a large group of casual listeners (thousands) listening through casual sound systems /
headphones, proving that they can hear inter-sample distortion and what's more important, it negatively affects their perception of music.
So far I could only read technical descriptions, regulations etc. - academic talk. But I couldn't find a study proving (or disproving) that inter-sample distortion has a real major impact on casual listener's listening experience.
I'm pretty sure that such a study would bring consensus that casual listeners don't give a flying f***k about it. They don't complain about music massacred with >10dB GR limiters, listened on iPhone/BT speaker or cheap headphones. It's hard to believe that microscopic level of inter-sample distortion would make them enjoy music listening less.

First, a single peak clip is not going to audible. the issue is high "clip density" where there are more clips in a period of time. If there is a lot of clipping, no double blind test is necessary because the level of wide-band intermod distortion will be plainly evident. Second, human hearing and perception are complex and difficult to quantify. Mostly we go by generalizations, gut feelings and hope. And, the knowledge that not everyone is going to agree on the quality of the sound. Third, measurable poor audio quality will be percieved by some as sounding good. This is related to my second point. Even audio professionals are prone to this, hence the concept of the loudness war and more focus on loudness management. This is why, for those that care, having a way to measure audio as a way to optimize quality is important. But, yes, some people are not going to care.

In that case, yes absolutely. I've been referring to cases where production and mastering are good and intermod distortion is happening rarely (single peak clip here and there). If there's a lot of clipping then production should be moved a step back and reconsidered again because it means that mix is pushed too hard (or other kinds of issues).

[Kr3eM]

I have Ozone Maximizer on my track, with the ceiling set to -0.1 dB and with the True Peak setting enabled....

I haven't used the new versions of Ozone so I can't tell from the pics, but is "Learn Treshold" active by any chance?

[eneasheynen]

I know this may not pin down the cause, but it may solve your problem: i normally use 2 limiters in series, the first in regular sample peak mode, and the second one at the same ceiling just catching true peaks. I prefer the sound of sample peak limiters, and this keeps that sound, combined with no true peaks. My normal "true peak catcher" is the output section of TDR Limiter6.

[exist01]

I have Ozone Maximizer on my track, with the ceiling set to -0.1 dB and with the True Peak setting enabled....

I haven't used the new versions of Ozone so I can't tell from the pics, but is "Learn Treshold" active by any chance?

It is not.

I know this may not pin down the cause, but it may solve your problem: i normally use 2 limiters in series, the first in regular sample peak mode, and the second one at the same ceiling just catching true peaks. I prefer the sound of sample peak limiters, and this keeps that sound, combined with no true peaks. My normal "true peak catcher" is the output section of TDR Limiter6.

I tried putting two instances of Ozone, one for regular limiting and another for true peak catching, but it didn't seem to do anything.


I am still experiencing this problem, now with a different track that is peaking above 5.0 dBTP. The only way to push the TP down is by setting Ozone's ceiling extremely low, like at -6.0 dB. This will cause the TP to be reduced by the same amount, but it absolutely destroys transients as well as any possibility of getting a decently loud track. So yeah I have no idea what's going on and what to do.

All I can do atm is simply ignore the TP reading and hope that the sound translates when it is played on other systems, because by far the best sounding version in my DAW is the one that peaks above +5.0 dBTP.

[exist01]

Wait wait wait wait wait wait I just figured it out. I had an EQ after the limiter that was only set to remove frequencies below 20hz and above 18khz. I moved it before the limiter and am now getting the TP I have my limiter set to. That makes no sense though, how does an EQ that only removes frequencies cause the TP to go nuts as if there were no limiter? I am so confused yet so relieved that the issue is solved. What the heck!?