Why "Approximate llr" input of the comm.LDPCDecoder has 4 different values when QAM16 is used?

Question

Paul Kubric 2020-10-7

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/607286-why-approximate-llr-input-of-the-comm-ldpcdecoder-has-4-different-values-when-qam16-is-used

编辑： Paul Kubric 2020-10-7

Hello! I'm trying to understand the work of the comm.LDPCDecoder module. When I implement QAM16 modulation on the encoded bits (using comm.LDPCEncoder with a standard parity-check matrix 32400x64800) and then demodulate these qam modulated symbols I receive 64800 values (qamDemod, check the snippet below) which are obtained using approximate log-likelihood algorithm. These demodulated values consist of 4 different values: -1.6, -0.4, 0.4 and 1.6. I thought that the output of the QAM demodulater (or the input to the LDPC Decoder) should take two different values as 1s and 0s have to be denoted using only two values (I guess I don't fully understand the LDPC decoder). I cannot understand why these values can take 4 different values. If I use QAM64 then I have 8 different values from the QAM demodulator. Any help will be appreciated!

Cheers,

Paul

P.S. I'll provide my snippet:

Mod_ind = 2^4; % QAM modulation index
H_matrix = dvbs2ldpc(1/2); % Buiilt-in DVBS2 PCM with a code 
%rate 1/2
ldpcEncoder = comm.LDPCEncoder(H_matrix);
ldpcDecoder = comm.LDPCDecoder(H_matrix);
data = (randi([0 1],32400,1)); % Generating data bits
encData = ldpcEncoder(data); % LDPC encoding
qamMod = qammod(encData,Mod_ind,'Gray','UnitAveragePower',1,...
    'InputType','Bit'); % QAM modulation
qamDemod = qamdemod((qamMod),...
    Mod_ind, 'Gray','UnitAveragePower', 1, 'OutputType', 'approxllr');
rxBits = ldpcDecoder(qamDemod); % QAM demodulation