Strange results with NicAc3Source

[Nikos]

Source:
test.wav, mono, 24 bit, 48 khz, 20 min, peek= -5.8 db, RMS = -30 db
The test file is part from RIO Bravo HD movie.
Rio Bravo.eac3 --> eac3to (Default Settings) --> Rio Bravo.wav

Group 1
aften -b 256 -dnorm 31 -dynrng 5 -readtoeof 1 test.wav 1.ac3

1.ac3 --> NicAc3Source DRC=0 --> 1a.wav (peak = -5.8 db, RMS = -30 db)
1.ac3 --> NicAc3Source DRC=1 --> 1b.wav (peak = -5.8 db, RMS = -30 db)


Group 2
aften -b 256 -dnorm 31 -dynrng 1 -readtoeof 1 test.wav 2.ac3

2.ac3 --> NicAc3Source DRC=0 --> 2a.wav (peak = -5.8 db, RMS = -30 db)
2.ac3 --> NicAc3Source DRC=1 --> 2b.wav (peak = -7.2 db RMS, = -31 db)


Group 3
aften -b 256 -dnorm 27 -dynrng 5 -readtoeof 1 test.wav 3.ac3

3.ac3 --> NicAc3Source DRC=0 --> 3a.wav (peak = -5.8 db, RMS = -30 db)
3.ac3 --> NicAc3Source DRC=1 --> 3b.wav (peak = -5.8 db, RMS = -30 db)


Group 4
aften -b 256 -dnorm 27 -dynrng 1 -readtoeof 1 test.wav 4.ac3

4.ac3 --> NicAc3Source DRC=0 --> 4a.wav (peak = -5.8 db, RMS = -30 db)
4.ac3 --> NicAc3Source DRC=1 --> 4b.wav (peak = -6.6 db, RMS = -29.7 db)


From Group 1 and Group 3 (with -dynrng 5) i conclude clearly that NicAc3Source ignore the Dialog Normalization value.
From Group 2 and 4 (with -dynrng 1) NicAc3Source with DRC=0 ignore the Dialog Normalization value,
but with DRC=1 the results are strange. Any explain please?
How NicAc3Source treat the Dialog Normalization value and DRC value inside the ac3 stream?

I used NicAudio_20070821.

[tebasuna51]

From Group 1 and Group 3 (with -dynrng 5) i conclude clear that NicAc3Source ignore the Dialog Normalization value.
From Group 2 and 4 (with -dynrng 1) NicAc3Source with DRC=0 ignore the Dialog Normalization value,
but with DRC=1 the results are strange. Any explain please?
How NicAc3Source treat the Dialog Norm. Value and DRC value inside the ac3 stream?

1) NicAc3Source ignore DialNorm always.

2) NicAc3Source apply the DRC values inside the ac3 stream when DRC=1.

3) Aften use the DialNorm to construct the DRC values inserted in the ac3 stream. Then Aften is the culprit of this difference.

I have found differences between the DRC values calculated with Aften and others encoders. I think the Aften DRC is still in beta stage and must be improved.

[Nikos]

Thanks tebasuna51, always you are very explanatory.
I will do some tests with another encoder.