You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
124 lines
5.0 KiB
124 lines
5.0 KiB
6 years ago
|
function AACSeq3 = AACoder3(fNameIn, fnameAACoded)
|
||
|
%Implementation of AAC encoder
|
||
|
% Usage AACSeq3 = AACoder3(fNameIn, fnameAACoded), where:
|
||
|
% Inputs
|
||
|
% - fNameIn is the filename and path of the file to encode
|
||
|
% - frameAACoded is the filename and path of the mat file that will
|
||
|
% be written after encoding
|
||
|
%
|
||
|
% Output
|
||
|
% - AACSeq3 is an array of structs containing K structs, where K is
|
||
|
% the number of computed frames. Every struct of the array consists
|
||
|
% of:
|
||
|
% * a frameType,
|
||
|
% * a winType,
|
||
|
% * chl.TNScoeffs which are the quantized TNS coefficients of
|
||
|
% this frame's left channel,
|
||
|
% * chr.TNScoeffs which are the quantized TNS coefficients of
|
||
|
% this frame's right channel,
|
||
|
% * chl.T which are the psychoacoustic thresholds of this frame's
|
||
|
% left channel,
|
||
|
% * chr.T which are the psychoacoustic thresholds of this frame's
|
||
|
% right channel,
|
||
|
% * chl.G which are the quantized global gains of this frame's
|
||
|
% left channel,
|
||
|
% * chr.G which are the quantized global gains of this frame's
|
||
|
% right channel,
|
||
|
% * chl.sfc which is the Huffman encoded sfc sequence of this
|
||
|
% frame's left channel,
|
||
|
% * chr.sfc which is the Huffman encoded sfc sequence of this
|
||
|
% frame's right channel,
|
||
|
% * chl.stream which is the Huffman encoded quantized MDCT
|
||
|
% sequence of this frame's left channel,
|
||
|
% * chr.stream which is the Huffman encoded quantized MDCT
|
||
|
% sequence of this frame's right channel,
|
||
|
% * chl.codebook which is the Huffman codebook used for this
|
||
|
% frame's left channel
|
||
|
% * chr.codebook which is the Huffman codebook used for this
|
||
|
% frame's right channel
|
||
|
|
||
|
% Declares constant window type
|
||
|
WINDOW_TYPE = 'SIN';
|
||
|
|
||
|
% Reads the audio file
|
||
|
[originalAudioData, ~] = audioread(fNameIn);
|
||
|
|
||
|
% Splits the audio in frames and determines the type of each frame
|
||
|
frameTypes{fix((length(originalAudioData) - 1025) / 1024)} = 'OLS';
|
||
|
frameTypes{1} = 'OLS';
|
||
|
for i = 1:length(frameTypes) - 2
|
||
|
nextFrameStart = (i + 1) * 1024 + 1;
|
||
|
nextFrameStop = nextFrameStart + 2047;
|
||
|
frameTypes{i+1} = SSC(1, originalAudioData(nextFrameStart:nextFrameStop, :), frameTypes{i});
|
||
|
end
|
||
|
|
||
|
% Assigns a type to the last frame
|
||
|
if strcmp(frameTypes{length(frameTypes) - 1}, 'LSS')
|
||
|
frameTypes{length(frameTypes)} = 'ESH';
|
||
|
elseif strcmp(frameTypes{length(frameTypes) - 1}, 'ESH')
|
||
|
frameTypes{length(frameTypes)} = 'ESH';
|
||
|
else
|
||
|
frameTypes{length(frameTypes)} = 'OLS';
|
||
|
end
|
||
|
|
||
|
% Encodes audio file
|
||
|
huffLUT = loadLUT();
|
||
|
AACSeq3(length(frameTypes)) = struct;
|
||
|
for i = 0:length(frameTypes) - 1
|
||
|
currFrameStart = i * 1024 + 1;
|
||
|
currFrameStop = currFrameStart + 2047;
|
||
|
frameF = filterbank(originalAudioData(currFrameStart:currFrameStop, :), frameTypes{i+1}, WINDOW_TYPE);
|
||
|
[frameF(:, 1), TNScoeffsL] = TNS(frameF(:, 1), frameTypes{i+1});
|
||
|
[frameF(:, 2), TNScoeffsR] = TNS(frameF(:, 2), frameTypes{i+1});
|
||
|
|
||
|
if i < 2
|
||
|
% TODO: what happens on the first two frames?
|
||
|
else
|
||
|
prev1FrameStart = (i - 1) * 1024 + 1;
|
||
|
prev1FrameStop = prev1FrameStart + 2047;
|
||
|
prev2FrameStart = (i - 2) * 1024 + 1;
|
||
|
prev2FrameStop = prev1FrameStart + 2047;
|
||
|
SMR = psycho(...
|
||
|
originalAudioData(currFrameStart:currFrameStop, :), ...
|
||
|
frameTypes{i+1}, ...
|
||
|
originalAudioData(prev1FrameStart:prev1FrameStop, :), ...
|
||
|
originalAudioData(prev2FrameStart:prev2FrameStop, :));
|
||
|
[SL, sfcL, GL] = AACquantizer(frameF, frameTypes{i+1}, SMR);
|
||
|
[SR, sfcR, GR] = AACquantizer(frameF, frameTypes{i+1}, SMR);
|
||
|
end
|
||
|
|
||
|
[streamL, huffcodebookL] = encodeHuff(SL, huffLUT);
|
||
|
[streamR, huffcodebookR] = encodeHuff(SR, huffLUT);
|
||
|
[sfcL, ~] = encodeHuff(sfcL, huffLUT, 12);
|
||
|
[sfcR, ~] = encodeHuff(sfcR, huffLUT, 12);
|
||
|
|
||
|
AACSeq3(i + 1).frameType = frameTypes(i + 1);
|
||
|
AACSeq3(i + 1).winType = WINDOW_TYPE;
|
||
|
AACSeq3(i + 1).chl.TNScoeffs = TNScoeffsL;
|
||
|
AACSeq3(i + 1).chr.TNScoeffs = TNScoeffsR;
|
||
|
AACSeq3(i + 1).chl.T = what; % TODO: find dis shit
|
||
|
AACSeq3(i + 1).chr.T = what;
|
||
|
AACSeq3(i + 1).chl.G = GL;
|
||
|
AACSeq3(i + 1).chr.G = GR;
|
||
|
AACSeq3(i + 1).chl.sfc = sfcL;
|
||
|
AACSeq3(i + 1).chr.sfc = sfcR;
|
||
|
AACSeq3(i + 1).chl.stream = streamL;
|
||
|
AACSeq3(i + 1).chr.stream = streamR;
|
||
|
AACSeq3(i + 1).chl.codebook = huffcodebookL;
|
||
|
AACSeq3(i + 1).chr.codebook = huffcodebookR;
|
||
|
end
|
||
|
|
||
|
save(fnameAACoded,AACSeq3);
|
||
|
|
||
|
if false
|
||
|
[idx,label] = grp2idx(sort(frameTypes));
|
||
|
hist(idx,unique(idx));
|
||
|
set(gca,'xTickLabel',label)
|
||
|
|
||
|
sum(idx(:) == 1)
|
||
|
sum(idx(:) == 2)
|
||
|
sum(idx(:) == 3)
|
||
|
sum(idx(:) == 4)
|
||
|
end
|
||
|
end
|