Make MATLAB output in strict ASCII

MATLAB console output used high-ASCII sometimes, e.g. for its prompt ( » ). In R2025A it may output UTF8, e.g., for zpk() output:

z =
0×1 empty double column vector
p =
9.900494692205900e-01
k =
9.950657811787363e-03

This 'feature' is a nuisance when interchanging text with other programs. Is there a built-in function to convert these occasional UTF non-ASCII characters to a reasonable approximation in standard ASCII? ( like 'x' or '>>') ?

Or is this behavior to be considered a bug that will be fixed (there appear to be quite a few nuisance bugs in R2025a).

-marcel

[Walter Roberson]

z = zeros(0,1)

z =

  0×1 empty double column vector

The character used between the 0 and the 1 is the '×' character, U+00D7, decimal 215. This character only takes a single byte to represent in UTF-8. This character is present in most ISO-8896-* character sets, and most of the common Windows code pages (for example it is missing in Windows 1255 - Hebrew)

[marcel hendrix]

> z = zeros(0,1)

One recognizes an expert by his concise examples :-) I actually used it just now to verify the problem.

Matlab R2023b did not use the 'small x'. (I have gone back to R2023b as I encountered a host of [what I consider to be] breaking changes in R2025a).

[Walter Roberson]

I was wrong, by the way, in saying that 0xD7 required only one byte in UTF-8.

Answer 1

[Walter Roberson]

Is there a built-in function to convert these occasional UTF non-ASCII characters to a reasonable approximation in standard ASCII? ( like 'x' or '>>') ?

No, there is no built-in function for that purpose.

Any such hypothetical built-in function would have to deal with the problem of contention about what the replacements should be. For example should × be replaced with x or with * or with by or with times ? The answer is likely to depend on context -- in some places × should be replaced by x but in some places it should be replaced by *

[Walter Roberson]

MATLAB console output used high-ASCII sometimes, e.g. for its prompt ( » ).

I have never ever observed MATLAB using » for its >> prompt.

What system Language setting are you using?

[marcel hendrix]

I see that prompt when I call MATLAB from C interactively using engOpen and engOutputBuffer and display the text in a standard Windows console (ASCII, obviously not the UTF16 variant).

I could of course translate the output in that call, but then it would be nice to have at least a documented list of what high-ASCII MATLAB is using (or a way to stop it altogether).

[Stephen23]

It sounds like Windows applies some kind of "smart quotes" transformation to its console. In which case this would depend mostly on your OS locale and language settings.

[Walter Roberson]

As an experiment, you should have the C program write the text to a file, and later examine the bytes of the file. This will allow you to distinguish between "output by the program" compared to "as displayed in Windows console"

[marcel hendrix]

The prompt ( » ) is $C2 $BB, so MATLAB appears to emit UTF-16. I think I'll special case my engOutputBuffer parser for this sequence and for decimal 215.

[Image Analyst]

Can you post a screenshot?

[Walter Roberson]

The prompt ( » ) is $C2 $BB, so MATLAB appears to emit UTF-16

That is UTF-8 rather than UTF-16