This post deals with audio CD's and DVD video -- I have no BluRay media. And it's sort of a worst-case scenario, one where a person can't physically save their media. I've learned an immense amount because it looks like an straighforward project, but it's not. I've written another blog post covering most of the remainder of it.
I recently cleared-out my storage area. I hadn't been in there in 17 years. A few boxes of CD's and DVD's were part of the contents, and the memories of the times when I purchased the media surged in an instant. They welled-up so quickly and powerfully that I choked-up. How had I become so weak and disfigured? Uncomfortable insights.
In spite of such anguish, a person is unlikely to trash the media. It's natural for a person to honor their history. At some point, I want to light a cigar, take a keepsake media disc off a shelf, and put it in a player for a rewatch/relisten.
If no home with a shelf, then very small storage areas can still be rented for what we used to pay for large ones. Ie, rent for what used to be an 8x10', might now rent a 4x4'. A 4x4 can hold a few CD's and a few papers, etc. Maybe $45.
Minus a millionaire's floorplan or an affordable storage area, a person realizes their options are down to one: back the media up, probably in some mediocre fashion/bitrate without cover art, and probably without individual tracks or chapters. Is it even worth it at all? As in every other post on this blog, the answer is "it's entirely up to you", and "the info here pertains to Linux". I found it was worth it for some discs, and others I just threw away. It's about 20 mins per CD, and about X mins per DVD.
TLDR
audio: 20 mins per CD
- abcde to backup the files and to import online CDDB data for ID3 tags. Verify the fetching CDDB URL inside abcde.conf
- Easytag (or some ppl prefer picard) to clean up ID3 tags.
- Sometimes groups of MP3's will require renaming. Gprename can hasten this.
- Cover art files do not always download and will not be the complete inner notes. While abcde is working, scan the inner liner at 200-300 dpi as close as possible to a square (1:1) crop
- Later, can make thumbnails for the HTML oversight page.
$ abcde -d /dev/sr0 -o mp3:"-b 320" -c ~/.config/abcde.conf
dvd: handbrake to backup the files.
playback note
See my post, which suggests an HTML/M3U file solution. Some may wish to install LibreELEC and Kodi in an old laptop to work as a "CD player". All of the engineered "media manager" products are bullsh*t IMO -- more concerned with phoning home your collection than any other task. A simple HTML page works much faster, and more reliably and configurably.
If we formatted our external back-up drive with XFS, then it will not work with a Windows or Mac device. So we will at least need some device to access the files. And a 3.5mm jack in the device, to move the audio to a computer speaker system or some headphones. Alternatively, we could design some static web pages and make thumbnails of of the CD covers with built-in links to playlists or which open the folders and play the contents. The specs for m3u's are here.
Although I run icewm, my MIME settings are handled by xdg. So it will play an MP3 link, but download an M3U link.
$ xdg-settings get default-web-browser
chromium.desktop
Applications work as .desktop files. Not sure.
audio 90-100Mb per CD
Time and discwise, a CD like Fleetwood Mac's Rumours, about 40 mins long, takes about 12 min to rip and transcode at 320k. The 320K bitrate leads to a 90Mb folder of files, or roughly 2.3Mb per minute (at 320Kb). I find 90Mb per CD unproblematic at today's drive prices.
Quality wise, I prefer a full 320Kb MP3, esp if the music is nuanced with flanging or orchestration. IMO, 192K can be OK, but 128K is definitely not OK unless just a speech or some such.
Probably abcde is still the easiest attack on a stack of backup CD's. We need some way to save id3 information, and probably edit it (eg, easytag). Installing abcde will also pull-in python-eyed3. I installed glyph out of an abundance of caution for photo embedding.
$ yay -S abcde id3 python-eyed3
# pacman -S glyph id3v2
audio - abcde use and config
Abcde has an old school text conf file. Get the skeleton from /etc and copy it to wherever to edit it, eg
$ cp /etc/abcde.conf ~/.config/abcde.conf
$ nano ~/.config/abcde.org
CDDBURL="gnudb.gnudb.org/~cddb/cddb.cgi"
When executing, the config file is called with the "c" flag and path.
$ abcde -c "~/.config/abcde.conf"
Abcde does not rip directly to MP3; it rips the entire CD to WAV's (10 mins), then converts each WAV to whatever number of other formats we want. From rip to conversion to one other format is about 12 mins per CD. Native MP3 conversion bit rate is 128k. Of course, if a person simply wants the WAV's, then use the defaults, eg $ abcde -d /dev/sr0.
For this project I'm likely to to specify MP3's at 320kb, though a person can specify what they like (or leave off the ':"-b 320"' for 128k).
$ abcde -d /dev/sr0 -o mp3:"-b 320"
audio - ID3 info
ID3 is no longer reliably being developed and there are different versions which might conflict. Of the versions, I consider ID3v2 compatible with nearly any player and therefor a reliable ID3 format for my MP3 files. ID3v2 is also supposedly the ID3 version EasyTag apparently uses. So after first tagging some MP3 files with EasyTag, I verified the ID3 version by installing id3v2. Then I ran its utility in a terminal, $ id3v2 -l [name].mp3, to verify my ID3 tags were version 2, ie, ID3v2.
$ id3v2 -l 02.Dreams.mp3
id3v2 tag info for 02.Dreams.mp3:
TIT2 (Title/songname/content description): Dreams
TPE1 (Lead performer(s)/Soloist(s)): Fleetwood Mac
TALB (Album/Movie/Show title): Rumours
TYER (Year): 1977
TRCK (Track number/Position in set): 02/11
TCON (Content type): Rock (17)
02.Dreams.mp3: No ID3v1 tag
Even with a functional CDDB URL in our ~/.config/abcde.conf , we often need to do the ID3 and art manually.
# pacman -S easytag
$ yay -S gprename
audio - ID3 art
It appears EasyTag only recognizes photos if gdk-pixbuf2 is installed.
# pacman -S gdk-pixbuf2
I arrange the scanner (xsane) for 300dpi, and preview the image close to a square. The resulting scan I then then scale. Get the WxH to about 1400x1400. This makes about a 500Kb file to add to each music file in that folder, so if 10 files in the CD, it adds 5MB to the entire folder. The 1K x 1K pic will display pretty sharp if running the file in VLC. Not sure other players. If I have an internal booklet in the CD, I scan it all and make a PDF to put in the folder.
Adding the art: put the image in the directory with the MP3's, open EasyTag in that directory, selecting all the files I want to receive that image (typically cover image). Then I go to the right and select "images" tab, add the pic with the "+" icon, and saving it to the files. A person will figure it out fast.
Removing art is fickle. I can do it in EasyTag, but it seems to stay in the VLC cache so I can never be sure. A sure way is to 1) Delete the directory for VLC album art: ~/.cache/vlc/ and, 2) go nuclear via "cd"ing into the music files directory and then...
$ id3v2 -r "APIC" *.mp3
audio - manual tasks
If I find old directories full of WAV's, like CD's I made back when, or whatever, I can batch convert the WAV's into 320k cbr MP3's.
$ for f in *.wav ; do lame -b 320 $f ; done
If I find a track that I want to slow it down (I don't mind pitch) and put a 0.8 speed version in the CD directory, perhaps to use with mixes, then...
$ sox foo.mp3 -C 320 foo8.mp3 speed 0.8
If I want to combine all the tracks in the CD into a blob, so I can play easily when driving or whatever...
$ sox foo1.mp3 foo2.mp3 foo3.mp3 output.mp3
There's a lot more about sox here.
Then I can get the art for it online or design something. Either way, put it in the folder with the files and add the art and info via EasyTag.
dvd 1.3G each
dvd - art
Use at least 300x300 ppi and some similar dimension. Give it the same name as the film file, but a JPG extension. When company arrives, they can easily peruse a collection with a file manager, or by browsing the photos in something like Geeqie.
dvd - content
Most movies (1.7 hrs) are 1.1GB (480P) and about 20 mins to back-up, assuming a person continues their other computer tasks throughout the encode. Episodes of TV shows are typically. If I have a higher-res recent DVD, I use the Fast(not Super-Fast)720p setting in Handbrake. It's 2 passes. Otherwise I used 480p Fast. run Fast The HandBrake software install (see picture below) is about 80Mb.
The rip is two parts, extraction and transcoding, and DVD structure is described here.
First, don't forget the obvious...
# pacman -S libdvdcss libdvdread libdvdnav
...since literally *none* of the players will do anything but spawn errors without explanation if these are lacking.
One fairly simple Arch solution is HandBrake.
# pacman -S handbrake
To execute, ghb which, along with icon, a person can find with ye olde...
$ pacman -Ql handbrake
...and there are flags of course, not sure where to locate.
I start HandBrake from a terminal (again, "ghb") even though it's a GUI. This is because the CSS keys take a minute or two to propagate and I can't see when they're finished within the application. Once I see the keys have propagated, I can return to GUI selection options.
dvd - storage
Probably need to think about file structure. At least like "movies, bbc, adultswim, sitcoms, detective". Or some such. Then include the scan of the cover art with the file
No comments:
Post a Comment