Multiply tall array with array

Question

Ervin 2024-4-25

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2111791-multiply-tall-array-with-array

回答： Edric Ellis 2024-4-29

I have a tall array, that contains audio samples created from a dataset like this:

dataS = tall(fileDatastore(fullfile(appData.OutputDirectoryPath, 'tmp_*.mat'), ...
    'ReadFcn', @(fname) getfield(load(fname), 'resampled'), ...
    'UniformRead', true));

This data is large, but its element count is known. I need to multiply this data with a signal:

t = ((0:N-1)'/fsn);
carrier = exp(1j*2*pi*carrierFreq*t);
ccy = dataS.*carrier;

The last line fails, with the following error:

Error using  .* 
Incompatible non-scalar tall array arguments. Each of the tall arrays must be the same size in the first dimension, must be derived from a single tall array, and must not have been indexed differently in the first
dimension (indexing operations include functions such as VERTCAT, SPLITAPPLY, SORT, CELL2MAT, SYNCHRONIZE, RETIME and so on).

I assume the issue is that it cannot guarantee that the arrays are compatible, but I can guarantee it. How can I do it?

2 个评论
显示无隐藏无

the cyclist 2024-4-25

在 MATLAB Online 中打开

Have you verified that the two arrays are the same size? Try putting this in your code, prior to that calculation:

size(dataS)
size(carrier)

Ervin 2024-4-25

在 MATLAB Online 中打开

Yes, I have, but I actually need to use:

gather(size(dataS))

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Edric Ellis 2024-4-29

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2111791-multiply-tall-array-with-array#answer_1449721

在 MATLAB Online 中打开

You need to use a bit of a trick to construct the colon vector. (This trick is alluded to on this doc page). Basically, you need to use find together with an element-wise expression derived from dataS such that the value of the expression is always true. This causes find to return 1:N in such a form that it can be combined with dataS. Like this:

dataS = tall(rand(1000,1));
% Use a trick to generate 1:(height(dataS))
cvec = find(true | isnan(dataS));
% Now, continue to compute t
fsn = 0.1;
t = (cvec-1) / fsn;
carrierFreq = 1000;
carrier = exp(1j*2*pi*carrierFreq*t);
gather(head(dataS .* carrier))
Evaluating tall expression using the Local MATLAB Session:
- Pass 1 of 2: 0% complete
- Pass 1 of 2: Completed in 0.26 sec
- Pass 2 of 2: 0% complete
- Pass 2 of 2: Completed in 0.18 sec
Evaluation completed in 0.85 sec
ans = 
   0.7841 + 0.0000i
   0.0410 - 0.0000i
   0.9299 - 0.0000i
   0.8883 - 0.0000i
   0.5102 - 0.0000i
   0.1375 - 0.0000i
   0.5657 - 0.0000i
   0.9033 - 0.0000i