开发者

How to write a binary file using Bash?

My problem is that I need to create a file with this exact bytes: 48, 00, 49, 00.

I cannot use C, perl, other sc开发者_Go百科ripting language (the target is an embedded device). I tried this using awk, and in desktop it does work:

# awk  'BEGIN{ printf "%c%c%c%c", 48, 00, 49, 00 }' | hexdump
0000000 0030 0031                              
0000004

However the target platform is running busybox v1.13.2 and this code does not work there. The awk version there does not output ascii "0" (all other values are ok).

What are your recommendations?


you can use the following command:

echo -n -e \\x48\\x00\\x49\\x00 > myfile


you could try echo, that also allows arbitrary ascii chars (those numbers are octal numbers).

echo -n -e \\0060\\0000\\0061\\0000  | hexdump


POSIX AWK standard says that passing a 0 to AWK's printf with %c format can result in unspecified behaviour. However... POSIX echo also is very limited, and though octal and hexadecimal specifiers (and -n) will work on GNU echo and BASH built-in... They may not work everywhere. To maximize the chance that you get consistent behaviour on all POSIX systems, it is better to use the shell command line's printf than either of these.

$ printf '\060\000\061\000' | od -An -tx1
 30 00 31 00

This looks odd to me though... You may be wanting to output 0x48, 0x00, 0x49, 0x00 -- which looks like a pretty pilot number in octal:

$ printf '\110\000\111\000' | od -An -tx1
 48 00 49 00


I needed to write binary files from hex using busybox within an old Android shell. This printf with a redirect worked in my use case.

Write the binary data in hex:

# busybox printf '\x74\x65\x73\x74' > /sdcard/test.txt

Display the result in hex:

# busybox hexdump -C /sdcard/test.txt
00000000  74 65 73 74                                       |test|
00000004

Display the result in ascii:

# cat /sdcard/test.txt
test


A couple of more general functions to output integers:

le16 () { # little endian 16 bit binary output 1st param: integer to 2nd param: file 
  v=`awk -v n=$1 'BEGIN{printf "%04X", n;}'`
  echo -n -e "\\x${v:2:2}\\x${v:0:2}" >> $2
}

le32 () { # 32 bit version
  v=`awk -v n=$1 'BEGIN{printf "%08X", n;}'`
  echo -n -e "\\x${v:6:2}\\x${v:4:2}\\x${v:2:2}\\x${v:0:2}" >> $2
}

Use to make an audio WAV file header for iio data stream:

channels=2
bits_per_sample=16
let "block_align = channels * bits_per_sample / 8"    

wave_header () { # pass file name and data size as parameters; rest are constants set elsewhere
  data_size=$2
  let "RIFFsize = data_size + 44 - 8"
  let "bytes_per_sec = sampleHz * block_align"

  echo -e -n "RIFF" > $1
    le32 $RIFFsize $1
  echo -e -n "WAVEfmt " >> $1
    le32 16 $1  # format size
    le16 1 $1   #format tag: 1 = PCM
    le16 $channels $1
    le32 $sampleHz $1
    le32 $bytes_per_sec $1
    le16 $block_align $1
    le16 $bits_per_sample $1   # bits per sample
  echo -e -n "data" >> $1
    le32 $data_size $1
}

Linux iio ADC data capture to WAV file:

sampleHz=8000
milliseconds=15  # capture length 

cd /sys/bus/iio/devices/iio:device0

cat /sys/bus/iio/devices/trigger0/name > trigger/current_trigger

echo 0 > buffer/enable

echo 0 > scan_elements/in_voltage0_en  #
echo 1 > scan_elements/in_voltage1_en  #  
echo 1 > scan_elements/in_voltage2_en  # 
echo 0 > scan_elements/in_voltage3_en  #

echo $sampleHz > sampling_frequency
sampleHz=`cat sampling_frequency`  # read back actual sample rate

let "buffer_length = block_align * sampleHz * milliseconds / 1000"
echo $buffer_length > buffer/length

cd $HOME

echo 1 > /sys/bus/iio/devices/iio:device0/buffer/enable
wave_header data.wav $buffer_length
head -c $buffer_length /dev/iio:device0 >> data.wav  # LE16 data
echo 0 > /sys/bus/iio/devices/iio:device0/buffer/enable


I don't know what's in busybox, but this might work because printf is smaller than awk.

$ printf "%c%c%c%c" 48 0 49 0 | hexdump

This is the output:

$ printf "%c" 1 | hexdump
0000000 0031                    
0

上一篇:

下一篇:

精彩评论

暂无评论...
验证码 换一张
取 消

最新问答

问答排行榜