Extracting values from a char variable

I am running FFmpeg through the system() function and the output is being stored within a char variable cmdout. I have added minimal reproducible code.

cmdout = "Guessed Channel Layout for Input Stream #0.0 : monoInput #0, wav, from 'testwav.wav':  Duration: 00:00:02.92, bitrate: 256 kb/s  Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 16000 Hz, mono, s16, 256 kb/sStream mapping:  Stream #0:0 -> #0:0 (pcm_s16le (native) -> opus (libopus))Press [q] to stop, [?] for helpOutput #0, opus, to 'test.opus':  Metadata:    encoder         : Lavf59.27.100  Stream #0:0: Audio: opus, 16000 Hz, mono, s16, 8 kb/s    Metadata:      encoder         : Lavc59.37.100 libopussize=       0kB time=00:00:00.12 bitrate=   8.8kbits/s speed=N/A    size=      13kB time=00:00:02.92 bitrate=  35.8kbits/s speed= 241x    video:0kB audio:11kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 12.397972%bench: utime=0.000s stime=0.016s rtime=0.012sbench: maxrss=14764kB";

This is how it displayed on the command window (without the %).

     %Guessed Channel Layout for Input Stream #0.0 : mono
     %Input #0, wav, from 'testwav.wav':
     %  Duration: 00:00:02.92, bitrate: 256 kb/s
     %  Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 16000 Hz, mono, s16, 256 kb/s
     %Stream mapping:
     %  Stream #0:0 -> #0:0 (pcm_s16le (native) -> opus (libopus))
     %Press [q] to stop, [?] for help
     %Output #0, opus, to 'test.opus':
     %  Metadata:
     %    encoder         : Lavf59.27.100
     %  Stream #0:0: Audio: opus, 16000 Hz, mono, s16, 8 kb/s
     %    Metadata:
     %      encoder         : Lavc59.37.100 libopus
     %size=       0kB time=00:00:00.12 bitrate=   8.8kbits/s speed=N/A    
     %size=      13kB time=00:00:02.92 bitrate=  35.8kbits/s speed= 241x    
     %video:0kB audio:11kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 12.397972%
     %bench: utime=0.000s stime=0.016s rtime=0.012s
     %bench: maxrss=14764kB

I am trying to extract out the values of the following

• 'Duration: 00:00:02.92'
• 'bitrate: 256'
• 'Stream #0:0: Audio: opus, 16000 Hz, mono, s16, 8 kb/s'
• 'encoder : Lavc59.37.100 libopus'
• 'utime=0.000'
• 'stime=0.016'
• 'rtime=0.012'

So far I have I am currently using the regexp() function to extract out the values into a struct.

Values  = regexp(cmdout,'(?<utime>utime=.{5})','names'); % Only one working so far

for k = 1:length(Values)
   disp(Values(k))
end

    utime: "utime=0.000"

I have been unable to add multiple regex statements without it failing to find anything. I was also not sure if it was possible to have the struct just store the actual value e.g. utime: '0.000' rather than how it currently also contains the text.

[Stephen23]

Please include sample data which also includes the newline characters.

[Adam]

I have added it below

Guessed Channel Layout for Input Stream #0.0 : mono
Input #0, wav, from 'testwav.wav':
  Duration: 00:00:02.92, bitrate: 256 kb/s
  Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 16000 Hz, mono, s16, 256 kb/s
Stream mapping:
  Stream #0:0 -> #0:0 (pcm_s16le (native) -> opus (libopus))
Press [q] to stop, [?] for help
Output #0, opus, to 'test.opus':
  Metadata:
    encoder         : Lavf59.27.100
  Stream #0:0: Audio: opus, 16000 Hz, mono, s16, 8 kb/s
    Metadata:
      encoder         : Lavc59.37.100 libopus
size=       0kB time=00:00:00.12 bitrate=   8.8kbits/s speed=N/A    
size=      13kB time=00:00:02.92 bitrate=  35.8kbits/s speed= 241x    
video:0kB audio:11kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 12.397972%
bench: utime=0.000s stime=0.016s rtime=0.012s
bench: maxrss=14764kB

[Stephen23]

The text contains multiple lines which start "Stream #0:0: ": what is the distinguishing characteristic of the one line that you want? How should our code know which ones to ignore and which one to detect?

[Adam]

The one that contains "Audio: opus" after it.

[Stephen23]

And how to distinguish between the lines containing "encoder :" ?

Answer 1

[Stephen23]

str = fileread('text.txt')

str = 
    'Guessed Channel Layout for Input Stream #0.0 : mono
     Input #0, wav, from 'testwav.wav':
       Duration: 00:00:02.92, bitrate: 256 kb/s
       Stream #0:0: Audio: pcm_s16le ([1][0][0][0] / 0x0001), 16000 Hz, mono, s16, 256 kb/s
     Stream mapping:
       Stream #0:0 -> #0:0 (pcm_s16le (native) -> opus (libopus))
     Press [q] to stop, [?] for help
     Output #0, opus, to 'test.opus':
       Metadata:
         encoder         : Lavf59.27.100
       Stream #0:0: Audio: opus, 16000 Hz, mono, s16, 8 kb/s
         Metadata:
           encoder         : Lavc59.37.100 libopus
     size=       0kB time=00:00:00.12 bitrate=   8.8kbits/s speed=N/A    
     size=      13kB time=00:00:02.92 bitrate=  35.8kbits/s speed= 241x    
     video:0kB audio:11kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 12.397972%
     bench: utime=0.000s stime=0.016s rtime=0.012s
     bench: maxrss=14764kB'

In three separate REGEXP calls (combining may be possible, but only with significant effort and complexity):

rx1 = "(Duration|bitrate|encoder)\s*:\s+([^,\n\r]+)";
tk1 = regexp(str,rx1,'tokens');
tk1 = vertcat(tk1{:})

tk1 = 4×2 cell array
    {'Duration'}    {'00:00:02.92'          }
    {'bitrate' }    {'256 kb/s'             }
    {'encoder' }    {'Lavf59.27.100'        }
    {'encoder' }    {'Lavc59.37.100 libopus'}

rx2 = "(Stream\s+\S+\s+Audio:)\s+(opus[^\n\r]+)";
tk2 = regexp(str,rx2,'tokens','once')

tk2 = 1×2 cell array
    {'Stream #0:0: Audio:'}    {'opus, 16000 Hz, mono, s16, 8 kb/s'}

rx3 = "([a-z]time)=(\d\.?\d*)";
tk3 = regexp(str,rx3,'tokens');
tk3 = vertcat(tk3{:})

tk3 = 3×2 cell array
    {'utime'}    {'0.000'}
    {'stime'}    {'0.016'}
    {'rtime'}    {'0.012'}

[Adam]

Thanks for the help. Is there any interactive or alternative sources for understanding the regexp commands. Something like Regex 101 but for MATLAB

[Stephen23]

"Something like Regex 101 but for MATLAB"

There are several interactive tools on MATLAB FEX:

https://www.mathworks.com/matlabcentral/fileexchange?q=regular+expressions

For example, my one is intended to be the simplest interactive tool on FEX:

https://www.mathworks.com/matlabcentral/fileexchange/48930-interactive-regular-expression-tool

Not as fancy as Regex 101, but it might help you enough.

"...alternative sources for understanding the regexp commands"

Repeated reading of the documentation. And when I write "repeated", I don't mean twice.

https://www.mathworks.com/help/matlab/matlab_prog/regular-expressions.html