from excel file 12-bit binary signed numbers to decimal values into matlab

Hello sir, I have 12-bit signed binary numbers in excel sheet ( two columns, inphase and quadrature phase components). How do I read the data into matlab and convert into Decimal values? I tried by using xlsread command but those binary numbers reading as decimal values ( it is giving in xEy format). Please help me.....

Thank you.

[Guillaume]

I have 12-bit signed binary numbers in excel sheet

Excel does not have any concept of 12-bit, or signed, or binary number. So that's certainly not what you have in excel.

What exactly do you have in the spreadsheet? Giving us an example spreadsheet (a file, not a screenshot!) would help greatly.

[samala rakesh]

Hi Guillaume,

Thank you so much for the response.

I attached .csv sample file. The data in the file is 12-signed binary number(first row is for inphase and second row is for quadrature phase). Now, I need to convert those numbers into decimal values.

[Guillaume]

Ok, the numbers are actually encoded as text (0 and 1 digits).

Which convention is used to encode negative numbers?

• sign-magnitude? i.e. -1 is: 100000000001
• one's complement? i.e. -1 is: 111111111110
• two's complement? i.e. -1 is: 111111111111
• something else?

[samala rakesh]

encoding is done by sign-mangnitude.

Answer 1

[Guillaume]

It's easier if you force matlab to read these numbers as text:

fid = fopen('data.csv');
text = textscan(fid, '%s%s', 'Delimiter', ',');   %read as text
fclose(fid);
text = [text{:}];  %concatenate as a Nx2 cell array
digits = cellfun(@(t) t-'0', text, 'UniformOutput', false);  %convert char vector of '0' and '1' into double vector of 0 and 1
numbers = cellfun(@(d) (-1)^d(1) * polyval(d(2:end), 2), digits);  %convert binary digits into actual number

A more efficient algorithm would concatenate the text as a 3D char array and permute the dimensions at the end instead of processing cell arrays.

[samala rakesh]

Great sir. Its's working fine. Thank you so much.

[samala rakesh]

If it is encoded with 2's complement, can we do with 'cellfun()'?

[Guillaume]

numbers = cellfun(@(d) polyval(d, 2) - d(1)*4096, digits)

is one way to do it for 2's complement. Probably not the most efficient, since it's processing the sign bit twice.

In any case, digits is a cell array of 1x12 integers representing the bits of each number. From there, it's easy to decode them according to whichever encoding you use.

Important note: The code in my answer never checks that the numbers are indeed encoded as 12 bits. The sign-magnitude version will give incorrect results if the leading 0 bits are omitted, while the 2's complement version will error. Adding the required check is left as an exercise to the reader...

[samala rakesh]

Ok, I understand. Thank you so much.