Check These Out
The FLAC audio must be encoded at 16000Hz sampling rate (SoX is your friend).
Outputs a short JSON string, the actual speech is in the hypotheses->utterance, the accuracy is stored in hypotheses->confidence (ranging from 0 to 1).
Google also accepts audio in some special speex format (audio/x-speex-with-header-byte), which is much smaller in comparison with losless FLAC, but I haven't been able to encode such a sample.
Surround the first letter of what you are grepping with square brackets and you won't have to spawn a second instance of grep -v. You could also use an alias like this (albeit with sed):
alias psgrep='ps aux | grep $(echo $1 | sed "s/^\(.\)/[\1]/g")'
Trace python statement execution and syscalls invoked during that simultaneously
This is very helpful to place in a shell startup file and will make grep use those options all the time. This example is nice as it won't show those warning messages, skips devices like fifos and pipes, and ignores case by default.
Convert all jpegs in the current directory into ~1024*768 pixels and ~ 150 KBytes jpegs
The ssh command alone will execute the sudo command remotely, but the password will be visible in the terminal as you type it. The two stty commands disable the terminal from echoing the password back to you, which makes the remote sudo act as it does locally.
Download google video with wget. Or, if you wish, pass video URL to ie mplayer to view as stream.
1. VURL: replace with url. I.e. http://video.google.com/videoplay?docid=12312312312312313#
2. OUPUT_FILE : optionally change to a more suited name. This is the downloaded file. I.e. foo.flv
# Improvements greatly appreciated. (close to my first linux command after ls -A :) )
Breakedown pipe by pipe:
1. wget: html from google, pass to stdout
2. grep: get the video url until thumbnailUrl (not needed)
3. grep: Strip off everything before http://
4. sed: urldecode
5. echo: hex escapes
6. sed: stipr of tailing before thumbnailUrl
7. wget: download. Here one could use i.e. mplayer or other...