Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[mwheelerk]

I received great help and explanation of the Dynamic Range values in the other thread. Now I would like to understand Peak Level (R128) and (Sample). What are the differences between these two measurements?  I am assuming that -0.0dbtp or -0.0db are theoretical maximum peaks and that any on the + side of the scale (+0.2dbtp) would incdi ate break walling or clipping. Am I on the right path with my thoughts or have I fallen completely off the track?

As usual any shared knowledge and insight is appreciated.

[Hendrik]

Both values represent the same, its just a different way to measure. The EBU-R 128 standard defines a method to measure "True Peak" values, which is what the R128 column represents, and the "Sample" column is an older less accurate way to calculate them.
Peaks can go slightly above 0, typically in the range of max +0-2 dBTP, which is caused by lossy encoding and so-called inter-sample peaks. Its generally not a problem, as volume leveling typically reduces the playback volume a bit, and makes room for the peaks without clipping them.

[6233638]

Peak Level (Sample) measures the highest point in the audio data. With lossless files, this should never exceed 0 dBFS.
With lossy files that use floating-point, it is possible for this to exceed 0 dBFS.
 
Peak Level (R128) is an indicator of True Peak Level, and will indicate whether a file has inter-sample clipping.
In the lossless data, a sample can never exceed 0 dBFS.
However, we do not listen to samples, we listen to analog waveforms.
 
When you have samples at or near 0 dBFS, the resulting analog waveform may be clipped.
 

 
The True Peak Level is calculated by upsampling the audio during analysis, which gives you a better idea of what the resulting waveform will be.
This is still only an estimate, so an additional 1dB of headroom is reserved.
 
So if a track has a Peak Level (R128) of -0.5 dBTP, it will be played back at -0.5 dB. (so that there is 1dB of headroom to account for errors)
If the Peak Level (R128) is above 0dB, say 1.5 dBTP, then it will be played back at -2.5 dB to prevent inter-sample clipping.

[mwheelerk]

Very good. Thanks so much for the precise explanations.

So if I am seeing a Peak Level (R128) reading for a file of +0.6dBTP; +0.6 Left; +0.1 Right or +1.0dBTP; +0.8 Left; +1.0 Right on a lossless AIFF file am I possibly seeing the results of clipping from mastering levels?

[6233638]

It's not necessarily clipped. There are tools which can simply push the volume up so that the loudest sample in a reasonably mastered file is 0 dBFS.
What a lot of people don't realize though is that doing this will often result in inter-sample peaks on playback, even if the file itself is not clipped.
But if you reduce the volume before playback (as Media Center does) so that the inter-sample peaks are below 0 dBFS then you won't have any clipping.
 
 
Of course it's entirely possible that the reason you have samples at 0 dBFS, or inter-sample peaks above 0 dB is because the waveform is clipped.
Without inspecting it, there's no way to know for sure.