Questions and Help

[syong39]

Hello.

While reading the code, I came across some questions. I'm having difficulty understanding

1. "Bin" in Ebu128LoudnessMeter.h and .cpp. Could you explain each of the variables for me? I'm not sure what "Bin" means.

2. Regarding the if statement in the STEP3 section of Ebu128LoudnessMeter.cpp's processBlock, I'm not quite sure what the condition means.

like this

if (numberOfSamplesInTheCurrentBin + bufferForMeasurement.getNumSamples() < numberOfSamplesPerBin)

Thank you for reading this and have a nice day.

[klangfreund]

1. The comment for bin explains it in detail:

LUFSMeter/Ebu128LoudnessMeter.h

Lines 143 to 166 in 783a59d

	/**
	After the samples are filtered and squared, they need to be
	accumulated.
	For the measurement of the short-term loudness, the previous
	3 seconds of samples need to be accumulated. For the other
	measurements shorter windows are used.
	This task could be solved using a ring buffer capable of
	holding 3 seconds of (multi-channel) audio and accumulate the
	samples every time the GUI wants to display the measurement.
	But it's easier on the CPU and needs far less memory if not
	all samples are stored but only the summation of all the
	samples in a 1/(GUI refresh rate) s - e.g. 1/20 s - time window.
	If we then need to determine the sum of all samples of the previous
	3s, we only have to accumulate 60 values.
	The downside of this method is that the displayed measurement
	has an accuracy jitter of max. e.g. 1/20 s. I guess this isn't
	an issue since GUI elements are refreshed asynchron anyway.
	Just to be clear, the accuracy of the measurement isn't effected.
	But if you ask for a measurement at time t, it will be the
	accurate measurement at time t - dt, where dt \in e.g. [0, 1/20s].
	*/
	vector<vector<double>> bin;

Also take a look at the visual representation of the algorithm at

https://github.com/klangfreund/LUFSMeter/blob/master/docs/developmentNotes/141017_LUFS_Meter_IntegratedLoudness.png

The bins are the sums (accumulators) on the top row.

2. The condition is explained in the comment directly above the condition. Once you understand your first question, you'll probably get this too.

Cheers