Why is a mastered track peaking over 0db?

[noceur]

Greets all,

Scratching my head trying to understand how it's possible that a mastered track, limited to 0 dBFs, can still peak over 0 dB on the channel and master fader. This is happening both in Ableton and Logic, both with all plugins disabled and without any. Also, track is not warped. Any ideas?

[Shift Gorden]

I think that's because tracks should be limited to about -0.3db, as opposed to 0.0db. I have a feeling that's what it is, mate.

[noceur]

tbh, i don't know what it's limited to. these are commercial releases i am looking at, with some peaking over 1-2 db. I think the answer is use of a limiter that is not a true peak (intersample peak) limiter.

[Tarekith]

Do you have warping turned on for that clip in Live? Any plug ins on those tracks, even disabled? Are these mp3s?

[timday]

/edit never mind you already mentioned the intersample peak issue

[noceur]

No warping, tried both with plugin-free signal path as well as disabled plugins. Yes, mp3s.

[Tarekith]

The decoding of the mp3s is the issue. Often times that leads to peaks that weren't in the original signal, the joys of compressed audio files.

[Stromkraft]

The decoding of the mp3s is the issue. Often times that leads to peaks that weren't in the original signal, the joys of compressed audio files.

Isn't this a mastering error? And a reason some mastering houses deliver also the MP3 as well as lossless?

[Danny_DJ]

1. as already mentioned, there are the physics of intersample peaks. many limiters only block the highest samples from being over 0dB. But since the actual analogue signal is a curve that can rise above two sampling points, there might be higher peaks.

2. since lossy conversions definitely alter the actual stored data (you might not be able to hear it, these codecs have become very good even in medium high settings and are often indistinguishable even for professionals on professional systems), there might be changes to the highest peaks as well. so you better leave a headroom of 1dB from your master file. that should be fine.

[Tarekith]

The decoding of the mp3s is the issue. Often times that leads to peaks that weren't in the original signal, the joys of compressed audio files.

Isn't this a mastering error? And a reason some mastering houses deliver also the MP3 as well as lossless?

No, because the mastering engineer usually has no control over what MP3 algorithm will be used to make the compressed file. Even if we make it ourselves, you might need to lower the output 2-3dB to prevent clipping. Another reason I like AAC over MP3, with the mastered for iTunes tools you can make sure you're not clipping anywhere later on.

Though I will admit, unless you're going to be doing more production with the MP3 later on (and why would you not use the wav?), I personally wouldn't stress too much about this post conversion clipping. Pretty much all MP3s have it, and no one is up in arms about it to date.

[papawise]

It's not a matter of MP3 only.
I have plenty of mastered tracks in lossless FLAC peaking over 2db in the master channel without any artifacts in the audio and no plugins of course.

The reason? I don't know it, but they are gaining 2db's that's a lot and perhaps a good kept secret.

[jestermgee]

I think most people left this party a good 2.5 years ago there mate.

[H20nly]

[fishmonkey]

It's not a matter of MP3 only.
I have plenty of mastered tracks in lossless FLAC peaking over 2db in the master channel without any artifacts in the audio and no plugins of course.

The reason? I don't know it, but they are gaining 2db's that's a lot and perhaps a good kept secret.

https://hydrogenaud.io/index.php/topic,98753.0.html