Splitting an mpeg4 file into multiple parts with ffmpeg

I have a 5.3GB mpeg4 file (of the just released ‘From Bedrooms to Billions: Amiga Years‘ documentary about the Commodore Amiga – I haven’t watched it yet but may post a review later), and the size of this single file is giving me some issues 🙂

It’s too large for a FAT32 formatted flash drive, since the largest file you can store on that file system is 4GB. It’s also too large to burn to a 4.7GB single-sided DVD. So I could have reformatted a flashdrive to exFAT (I don’t think the PS4 reads NTFS drives, and my Mac won’t write to NTFS, at least not without other drivers), but the easiest option is probably to split the file to smaller parts.

Now, if you’ve done any video editing before, rendering 1080p video and any amount of transcoding to different formats takes a ton of time. I didn’t want to recode the video or change the format in any way, I just wanted to split the file.

Turns out ffmpeg will do exactly this, and to split into 2x 1hr15m smaller files only took a few minutes on my i7 Mac. I took suggestions from this question, and ended up with these two commands:

ffmpeg -i inputfile.mp4 -ss 00:00:00.0 -c copy -t 01:15:00.0 outputfile-pt1.mp4

ffmpeg -i inputfile.mp4 -ss 01:15:00.0 -c copy outputfile-pt2.mp4

The first one takes input from the start and for 1hr15mins writes to the output file, and the second starts from 1hr15min into the input file and writes the remainder from that point onwards to the second output file.

Inline text replacement with sed

Replacing values in files is incredibly easy with sed. Here’s some examples:


sed 's/match/replace/g' file.txt

Find match, replace with replace, globally (all matches), in file.txt


sed 's/match/replace/g' file.txt > file2.txt

Same as before, but write results to new file, file2.txt


sed -n 's/match/replace/p' file.txt

-n suppresses output of the results, but /p prints out just the matching patterns that are replaced.


sed -i.old 's/match/replace/g' file.txt

Replace matches in the input file with inline replace (-i), renaming original file file.txt.old and writing inline replace results to file.txt


Node.js and Mongoose: connection hanging on find()

Following this tutorial to build a simple REST api on Node.js with Express, attempting to retrieve all docs from my MongoDB gave these errors:

2016-05-19T20:10:47.311-0700 [initandlisten] connection accepted from #4 (2 connections now open)

2016-05-19T20:10:47.325-0700 [conn4]  authenticate db: nodetest { authenticate: 1, user: "", nonce: "xxx", key: "xxx" }

2016-05-19T20:10:47.336-0700 [conn4] Failed to authenticate @nodetest with mechanism MONGODB-CR: ProtocolError field missing/wrong type in received authenticate command

Some quick Googling turned up the same error in this post that just suggested authentication was failing because no user was passed. So created a new used with:

db.createUser({user : 'nodetest', pwd: 'passwordhere', roles: ['readWrite', 'nodetest']})

… and problem solved.

Update: onscreen keyboard for Raspberry Pi 7″ touchscreen

Quick followup from last week’s post on setting up an onscreen keyboard. I couldn’t find a way of getting an onscreen keyboard, either matchbox-keyboard or florence, to respond on Raspbian’s logon screen. Rather than spend more time on investigating this, I just enabled the logon to desktop in raspi-config, and now logged on, either of the keyboards work as expected, perfectly well.

Florence seems to be the better of the two I tried. Once opened from the menu you can drag it around, or minimize it to a keyboard icon, so you can pop-it open as needed.

If I could get it to work from the logon screen too that would be awesome, but avoiding that issue and once you get to the desktop, it works great.