apostolof-ece-auth-gr
/
authMultimediaAssignment2018
1
0
Fork 0
Code Releases Activity
Browse Source

Add two extra frames with zero padding on the end, Minor fix

master
Apostolos Fanakis 6 years ago
parent
cc350b8f39
commit
1beb80d214
11 changed files with 116 additions and 34 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Unified View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 23
      Level_1/AACoder1.m
  2. BIN
      Level_1/decoded1.wav
  3. 4
      Level_1/demoAAC1.m
  4. 23
      Level_2/AACoder2.m
  5. BIN
      Level_2/decoded2.wav
  6. 6
      Level_2/demoAAC2.m
  7. 76
      Level_3/AACoder3.m
  8. 2
      Level_3/AACquantizer.m
  9. BIN
      Level_3/decoded3.wav
  10. 16
      Level_3/demoAAC3.m
  11. BIN
      Level_3/encoded.mat

23
Level_1/AACoder1.m
View File

@ -18,12 +18,20 @@ function AACSeq1 = AACoder1(fNameIn)
[originalAudioData, ~] = audioread(fNameIn); [originalAudioData, ~] = audioread(fNameIn);
% Splits the audio in frames and determines the type of each frame % Splits the audio in frames and determines the type of each frame
frameTypes{fix((length(originalAudioData) - 1025) / 1024), 1} = 'OLS'; frameTypes{ceil(length(originalAudioData) / 1024), 1} = 'OLS';
frameTypes{1} = 'OLS'; frameTypes{1} = 'OLS';
for i = 1:length(frameTypes) - 2 for i = 1:length(frameTypes) - 2
nextFrameStart = (i + 1) * 1024 + 1; nextFrameStart = (i + 1) * 1024 + 1;
nextFrameStop = nextFrameStart + 2047; nextFrameStop = nextFrameStart + 2047;
frameTypes{i+1} = SSC(1, originalAudioData(nextFrameStart:nextFrameStop, :), frameTypes{i}); if nextFrameStop > length(originalAudioData)
tmpZeroPadded = originalAudioData(nextFrameStart:length(originalAudioData), :);
tmpZeroPadded = padarray(tmpZeroPadded, [(2048 - length(tmpZeroPadded)) 0], 'post');
frameTypes{i+1} = SSC(1, tmpZeroPadded, frameTypes{i});
clearvars tmpZeroPadded
else
frameTypes{i+1} = SSC(1, originalAudioData(nextFrameStart:nextFrameStop, :), frameTypes{i});
end
end end
% Assigns a type to the last frame % Assigns a type to the last frame
@ -40,7 +48,16 @@ function AACSeq1 = AACoder1(fNameIn)
for i = 0:length(frameTypes) - 1 for i = 0:length(frameTypes) - 1
currFrameStart = i * 1024 + 1; currFrameStart = i * 1024 + 1;
currFrameStop = currFrameStart + 2047; currFrameStop = currFrameStart + 2047;
frameF = filterbank(originalAudioData(currFrameStart:currFrameStop, :), frameTypes{i+1}, WINDOW_TYPE);
if currFrameStop > length(originalAudioData)
tmpZeroPadded = originalAudioData(nextFrameStart:length(originalAudioData), :);
tmpZeroPadded = padarray(tmpZeroPadded, [(2048 - length(tmpZeroPadded)) 0], 'post');
frameF = filterbank(tmpZeroPadded, frameTypes{i+1}, WINDOW_TYPE);
clearvars tmpZeroPadded
else
frameF = filterbank(originalAudioData(currFrameStart:currFrameStop, :), frameTypes{i+1}, WINDOW_TYPE);
end
AACSeq1(i + 1).frameType = frameTypes(i + 1); AACSeq1(i + 1).frameType = frameTypes(i + 1);
AACSeq1(i + 1).winType = WINDOW_TYPE; AACSeq1(i + 1).winType = WINDOW_TYPE;

BIN
Level_1/decoded2.wav → Level_1/decoded1.wav
View File

Binary file not shown.

4
Level_1/demoAAC1.m
View File

@ -15,7 +15,5 @@ function SNR = demoAAC1(fNameIn, fNameOut)
[audioData, ~] = audioread(fNameIn); [audioData, ~] = audioread(fNameIn);
snr(audioData(1025:length(decodedAudio)-1024, 1), audioData(1025:length(decodedAudio)-1024, 1) - decodedAudio(1025:end-1024, 1)) SNR = snr(audioData(1025:end-1378, :), audioData(1025:end-1378, :) - decodedAudio(1025:length(audioData)-1378, :));
snr(audioData(1025:length(decodedAudio)-1024, 2), audioData(1025:length(decodedAudio)-1024, 2) - decodedAudio(1025:end-1024, 2))
SNR = snr(audioData(1025:length(decodedAudio)-1024, :), audioData(1025:length(decodedAudio)-1024, :) - decodedAudio(1025:end-1024, :));
end end

23
Level_2/AACoder2.m
View File

@ -21,12 +21,20 @@ function AACSeq2 = AACoder2(fNameIn)
[originalAudioData, ~] = audioread(fNameIn); [originalAudioData, ~] = audioread(fNameIn);
% Splits the audio in frames and determines the type of each frame % Splits the audio in frames and determines the type of each frame
frameTypes{fix((length(originalAudioData) - 1025) / 1024), 1} = 'OLS'; frameTypes{ceil(length(originalAudioData) / 1024), 1} = 'OLS';
frameTypes{1} = 'OLS'; frameTypes{1} = 'OLS';
for i = 1:length(frameTypes) - 2 for i = 1:length(frameTypes) - 2
nextFrameStart = (i + 1) * 1024 + 1; nextFrameStart = (i + 1) * 1024 + 1;
nextFrameStop = nextFrameStart + 2047; nextFrameStop = nextFrameStart + 2047;
frameTypes{i+1} = SSC(1, originalAudioData(nextFrameStart:nextFrameStop, :), frameTypes{i}); if nextFrameStop > length(originalAudioData)
tmpZeroPadded = originalAudioData(nextFrameStart:length(originalAudioData), :);
tmpZeroPadded = padarray(tmpZeroPadded, [(2048 - length(tmpZeroPadded)) 0], 'post');
frameTypes{i+1} = SSC(1, tmpZeroPadded, frameTypes{i});
clearvars tmpZeroPadded
else
frameTypes{i+1} = SSC(1, originalAudioData(nextFrameStart:nextFrameStop, :), frameTypes{i});
end
end end
% Assigns a type to the last frame % Assigns a type to the last frame
@ -43,7 +51,16 @@ function AACSeq2 = AACoder2(fNameIn)
for i = 0:length(frameTypes) - 1 for i = 0:length(frameTypes) - 1
currFrameStart = i * 1024 + 1; currFrameStart = i * 1024 + 1;
currFrameStop = currFrameStart + 2047; currFrameStop = currFrameStart + 2047;
frameF = filterbank(originalAudioData(currFrameStart:currFrameStop, :), frameTypes{i+1}, WINDOW_TYPE); if currFrameStop > length(originalAudioData)
tmpZeroPadded = originalAudioData(nextFrameStart:length(originalAudioData), :);
tmpZeroPadded = padarray(tmpZeroPadded, [(2048 - length(tmpZeroPadded)) 0], 'post');
frameF = filterbank(tmpZeroPadded, frameTypes{i+1}, WINDOW_TYPE);
clearvars tmpZeroPadded
else
frameF = filterbank(originalAudioData(currFrameStart:currFrameStop, :), frameTypes{i+1}, WINDOW_TYPE);
end
[frameF(:, 1), TNScoeffsL] = TNS(frameF(:, 1), frameTypes{i+1}); [frameF(:, 1), TNScoeffsL] = TNS(frameF(:, 1), frameTypes{i+1});
[frameF(:, 2), TNScoeffsR] = TNS(frameF(:, 2), frameTypes{i+1}); [frameF(:, 2), TNScoeffsR] = TNS(frameF(:, 2), frameTypes{i+1});

BIN
Level_2/decoded2.wav
View File

Binary file not shown.

6
Level_2/demoAAC2.m
View File

@ -14,8 +14,6 @@ function SNR = demoAAC2(fNameIn, fNameOut)
decodedAudio = iAACoder2(AACSeq2, fNameOut); decodedAudio = iAACoder2(AACSeq2, fNameOut);
[audioData, ~] = audioread(fNameIn); [audioData, ~] = audioread(fNameIn);
snr(audioData(1025:length(decodedAudio)-1024, 1), audioData(1025:length(decodedAudio)-1024, 1) - decodedAudio(1025:end-1024, 1)) SNR = snr(audioData(1025:end-1378, :), audioData(1025:end-1378, :) - decodedAudio(1025:length(audioData)-1378, :));
snr(audioData(1025:length(decodedAudio)-1024, 2), audioData(1025:length(decodedAudio)-1024, 2) - decodedAudio(1025:end-1024, 2))
SNR = snr(audioData(1025:length(decodedAudio)-1024, :), audioData(1025:length(decodedAudio)-1024, :) - decodedAudio(1025:end-1024, :));
end end

76
Level_3/AACoder3.m
View File

@ -53,12 +53,20 @@ function AACSeq3 = AACoder3(fNameIn, fnameAACoded)
[originalAudioData, ~] = audioread(fNameIn); [originalAudioData, ~] = audioread(fNameIn);
% Splits the audio in frames and determines the type of each frame % Splits the audio in frames and determines the type of each frame
frameTypes{fix((length(originalAudioData) - 1025) / 1024), 1} = 'OLS'; frameTypes{ceil(length(originalAudioData) / 1024), 1} = 'OLS';
frameTypes{1} = 'OLS'; frameTypes{1} = 'OLS';
for i = 1:length(frameTypes) - 2 for i = 1:length(frameTypes) - 2
nextFrameStart = (i + 1) * 1024 + 1; nextFrameStart = (i + 1) * 1024 + 1;
nextFrameStop = nextFrameStart + 2047; nextFrameStop = nextFrameStart + 2047;
frameTypes{i+1} = SSC(1, originalAudioData(nextFrameStart:nextFrameStop, :), frameTypes{i}); if nextFrameStop > length(originalAudioData)
tmpZeroPadded = originalAudioData(nextFrameStart:length(originalAudioData), :);
tmpZeroPadded = padarray(tmpZeroPadded, [(2048 - length(tmpZeroPadded)) 0], 'post');
frameTypes{i+1} = SSC(1, tmpZeroPadded, frameTypes{i});
clearvars tmpZeroPadded
else
frameTypes{i+1} = SSC(1, originalAudioData(nextFrameStart:nextFrameStop, :), frameTypes{i});
end
end end
% Assigns a type to the last frame % Assigns a type to the last frame
@ -76,7 +84,15 @@ function AACSeq3 = AACoder3(fNameIn, fnameAACoded)
for i = 0:length(frameTypes) - 1 for i = 0:length(frameTypes) - 1
currFrameStart = i * 1024 + 1; currFrameStart = i * 1024 + 1;
currFrameStop = currFrameStart + 2047; currFrameStop = currFrameStart + 2047;
frameF = filterbank(originalAudioData(currFrameStart:currFrameStop, :), frameTypes{i+1}, WINDOW_TYPE); if currFrameStop > length(originalAudioData)
tmpZeroPadded = originalAudioData(nextFrameStart:length(originalAudioData), :);
tmpZeroPadded = padarray(tmpZeroPadded, [(2048 - length(tmpZeroPadded)) 0], 'post');
frameF = filterbank(tmpZeroPadded, frameTypes{i+1}, WINDOW_TYPE);
clearvars tmpZeroPadded
else
frameF = filterbank(originalAudioData(currFrameStart:currFrameStop, :), frameTypes{i+1}, WINDOW_TYPE);
end
[frameF(:, 1), TNScoeffsL] = TNS(frameF(:, 1), frameTypes{i+1}); [frameF(:, 1), TNScoeffsL] = TNS(frameF(:, 1), frameTypes{i+1});
[frameF(:, 2), TNScoeffsR] = TNS(frameF(:, 2), frameTypes{i+1}); [frameF(:, 2), TNScoeffsR] = TNS(frameF(:, 2), frameTypes{i+1});
@ -105,16 +121,50 @@ function AACSeq3 = AACoder3(fNameIn, fnameAACoded)
prev1FrameStop = prev1FrameStart + 2047; prev1FrameStop = prev1FrameStart + 2047;
prev2FrameStart = (i - 2) * 1024 + 1; prev2FrameStart = (i - 2) * 1024 + 1;
prev2FrameStop = prev2FrameStart + 2047; prev2FrameStop = prev2FrameStart + 2047;
SMRL = psycho(...
originalAudioData(currFrameStart:currFrameStop, 1), ... if prev1FrameStop > length(originalAudioData)
frameTypes{i+1}, ... tmpZeroPaddedPrev = originalAudioData(prev1FrameStart:length(originalAudioData), :);
originalAudioData(prev1FrameStart:prev1FrameStop, 1), ... tmpZeroPaddedPrev = padarray(tmpZeroPaddedPrev, [(2048 - length(tmpZeroPaddedPrev)) 0], 'post');
originalAudioData(prev2FrameStart:prev2FrameStop, 1)); tmpZeroPadded = originalAudioData(currFrameStart:length(originalAudioData), :);
SMRR = psycho(... tmpZeroPadded = padarray(tmpZeroPadded, [(2048 - length(tmpZeroPadded)) 0], 'post');
originalAudioData(currFrameStart:currFrameStop, 2), ... SMRL = psycho(...
frameTypes{i+1}, ... tmpZeroPadded(:, 1), ...
originalAudioData(prev1FrameStart:prev1FrameStop, 2), ... frameTypes{i+1}, ...
originalAudioData(prev2FrameStart:prev2FrameStop, 2)); tmpZeroPaddedPrev(:, 1), ...
originalAudioData(prev2FrameStart:prev2FrameStop, 2));
SMRR = psycho(...
tmpZeroPadded(:, 2), ...
frameTypes{i+1}, ...
tmpZeroPaddedPrev(:, 2), ...
originalAudioData(prev2FrameStart:prev2FrameStop, 2));
clearvars tmpZeroPadded tmpZeroPaddedPrev
elseif currFrameStop > length(originalAudioData)
tmpZeroPadded = originalAudioData(currFrameStart:length(originalAudioData), :);
tmpZeroPadded = padarray(tmpZeroPadded, [(2048 - length(tmpZeroPadded)) 0], 'post');
SMRL = psycho(...
tmpZeroPadded(:, 1), ...
frameTypes{i+1}, ...
originalAudioData(prev1FrameStart:prev1FrameStop, 1), ...
originalAudioData(prev2FrameStart:prev2FrameStop, 2));
SMRR = psycho(...
tmpZeroPadded(:, 2), ...
frameTypes{i+1}, ...
originalAudioData(prev1FrameStart:prev1FrameStop, 2), ...
originalAudioData(prev2FrameStart:prev2FrameStop, 2));
clearvars tmpZeroPadded
else
SMRL = psycho(...
originalAudioData(currFrameStart:currFrameStop, 1), ...
frameTypes{i+1}, ...
originalAudioData(prev1FrameStart:prev1FrameStop, 1), ...
originalAudioData(prev2FrameStart:prev2FrameStop, 1));
SMRR = psycho(...
originalAudioData(currFrameStart:currFrameStop, 2), ...
frameTypes{i+1}, ...
originalAudioData(prev1FrameStart:prev1FrameStop, 2), ...
originalAudioData(prev2FrameStart:prev2FrameStop, 2));
end
end end
[SL, sfcL, GL] = AACquantizer(frameF(:, 1), frameTypes{i+1}, SMRL); [SL, sfcL, GL] = AACquantizer(frameF(:, 1), frameTypes{i+1}, SMRL);

2
Level_3/AACquantizer.m
View File

@ -73,6 +73,7 @@ function [S, sfc, G] = AACquantizer(frameF, frameType, SMR)
quantErr = sumsqr(subFrameF - frameFDequant); quantErr = sumsqr(subFrameF - frameFDequant);
end end
% Undoes the last incrementation
if quantCoeff(band) ~= initialQuantCoeff if quantCoeff(band) ~= initialQuantCoeff
quantCoeff(band) = quantCoeff(band) - 1; quantCoeff(band) = quantCoeff(band) - 1;
S(frameWlow:frameWhigh) = sign(subFrameF) .* round(( ... S(frameWlow:frameWhigh) = sign(subFrameF) .* round(( ...
@ -136,6 +137,7 @@ function [S, sfc, G] = AACquantizer(frameF, frameType, SMR)
quantErr = sumsqr(subFrameF - frameFDequant); quantErr = sumsqr(subFrameF - frameFDequant);
end end
% Undoes the last incrementation
if quantCoeff(band) ~= initialQuantCoeff if quantCoeff(band) ~= initialQuantCoeff
quantCoeff(band) = quantCoeff(band) - 1; quantCoeff(band) = quantCoeff(band) - 1;
S(currFrameStart + frameWlow:currFrameStart + frameWhigh) = sign(subFrameF) .* round(( ... S(currFrameStart + frameWlow:currFrameStart + frameWhigh) = sign(subFrameF) .* round(( ...

BIN
Level_3/decoded3.wav
View File

Binary file not shown.

16
Level_3/demoAAC3.m
View File

@ -17,25 +17,25 @@ function [SNR, bitrate, compression] = demoAAC3(fNameIn, fNameOut, frameAACoded)
AACSeq3 = AACoder3(fNameIn, frameAACoded); AACSeq3 = AACoder3(fNameIn, frameAACoded);
totalSize = 0; totalSize = 0; % In bits
for frameIndex = 1:length(AACSeq3) for frameIndex = 1:length(AACSeq3)
totalSize = totalSize + 2; % frameType totalSize = totalSize + 2; % frameType
totalSize = totalSize + 2 * 4; % TNS coefficients totalSize = totalSize + 4 * 4; % TNS coefficients
totalSize = totalSize + length(AACSeq3(frameIndex).chl.G) * 4; totalSize = totalSize + length(AACSeq3(frameIndex).chl.G) * 16;
totalSize = totalSize + length(AACSeq3(frameIndex).chr.G) * 4; totalSize = totalSize + length(AACSeq3(frameIndex).chr.G) * 16;
totalSize = totalSize + length(AACSeq3(frameIndex).chl.sfc); totalSize = totalSize + length(AACSeq3(frameIndex).chl.sfc);
totalSize = totalSize + length(AACSeq3(frameIndex).chr.sfc); totalSize = totalSize + length(AACSeq3(frameIndex).chr.sfc);
totalSize = totalSize + length(AACSeq3(frameIndex).chl.stream); totalSize = totalSize + length(AACSeq3(frameIndex).chl.stream);
totalSize = totalSize + length(AACSeq3(frameIndex).chr.stream); totalSize = totalSize + length(AACSeq3(frameIndex).chr.stream);
totalSize = totalSize + 4; % Codebook of left channel
totalSize = totalSize + 4; % Codebook of right channel
end end
decodedAudio = iAACoder3(AACSeq3, fNameOut); decodedAudio = iAACoder3(AACSeq3, fNameOut);
bitrate = totalSize / (length(decodedAudio) / 48000); bitrate = totalSize / (length(decodedAudio) / 48000);
[audioData, ~] = audioread(fNameIn); [audioData, ~] = audioread(fNameIn);
compression = (length(audioData) * 64) / totalSize; compression = (2 * length(audioData) * 16) / (length(audioData) / 48000) / bitrate;
snr(audioData(1025:length(decodedAudio)-1024, 1), audioData(1025:length(decodedAudio)-1024, 1) - decodedAudio(1025:end-1024, 1)) SNR = snr(audioData(1025:end-1378, :), audioData(1025:end-1378, :) - decodedAudio(1025:length(audioData)-1378, :));
snr(audioData(1025:length(decodedAudio)-1024, 2), audioData(1025:length(decodedAudio)-1024, 2) - decodedAudio(1025:end-1024, 2))
SNR = snr(audioData(1025:length(decodedAudio)-1024, :), audioData(1025:length(decodedAudio)-1024, :) - decodedAudio(1025:end-1024, :));
end end

BIN
Level_3/encoded.mat
View File

Binary file not shown.
Write Preview
Loading…
Cancel
Save