I have a question about padding in MPEG audio streams.

In this ISO-IEC 11172-3 9041091 document I read:
(1)
(2)
(3)

(With this 'calculation' being a 'number of slots' (N) (= 12 * bitrate / samplefreq for L.I, = 144 * bitrate / samplefreq. for L.II / L.III)).

Now, I guess it should be obvious, but can someone point out how to use this to verify padding in an existing MPEG audio file?

Thanks in advance :-)

I was hoping someone would care to shed some light on this, anyone?

Let's see whether I can elaborate on this myself a bit (I'm just writing this off the top of my head):

say Nc is this 'certain number of frames',
say Np is the number of frames which are padded, as a part of this 'certain number of frames' (with Np <= Nc ),
and Ns is the 'number-of-slots' from the suggested calculation, then

Nc x Ns + Np x 1 = something, with something ~ (1 / sample-frequency)..

Any ideas about 'something' ?
;-)

example: 128 kbps CBR mp3

• slots (i.e. bytes) per frame = 144*128/44.1 = 417.95918367
• unpadded frame: 417 bytes
• padded frame: 418 bytes
• 95.9% of all frames in a file should be padded.

So you can verify padding in an existing file like this: add the sizes (number of slots) of all frames (padded and unpadded) and divide that sum by the number of frames. The result is the actual average frame size in that file. Compare this to the theoretical average frame size (144*bitrate/samplefreq).

For the strict ISO verification as described in your quote you have to check for every frame of the file the condition mentioned: that the actual accumulated frame sizes (from beginning of file up to the current frame) are never larger and never smaller by more than one slot than the theoretical accumulated frame sizes (which may be fractional).

Thanks smack (the obbvious clearly escaped me here :-)