I'm sure I'm not the only person who has started to rip CDs from their classical music collection and pretty soon got disheartened by the banal fashion with which it is treated by most music software. Most of the software available is based on a paradigm that says there is a Song by an Artist on an Album (although things are starting to change). It may allow for a Composer as well, but still assumes that all composers write only songs, and not works which contain a number of parts. There should (in theory) be at least two big benefits of ripping your collection: (1) the reduction in shelf space and (2) the ability to quickly locate the music you want. Most software achieves (1) but is not much help as regards (2). In addition, now that the physical CD case is not to hand when the music is playing, you would like to see some information about it. Frequently this is rudimentary. And of course, you don't want to put in too much work to achieve the required result. The information presented here is based on my approach to resolving these problems; an approach which has evolved over time and with which I am now very happy, although I’m sure it will evolve further. At the centre of this is the excellent "Muso" software which, I think uniquely, gives the classical music lover a "CD Insert" view of their albums coupled with powerful searching and linking facilities, plus an ability to play to either Logitech Media Server ("LMS" - formerly Squeezebox Server), iTunes, HQPlayer, Foobar or the embedded player. I have now coupled this with the power of MusicBrainz, which provides a rich and consistent source of "metadata" (data about your music).
WHAT'S THE PROBLEM?
There are five main hurdles to be faced in ripping and streaming classical CDs (some of which also apply to jazz albums) in a way that satisfies the music lover, rather than the casual listener.
- Poor availability of metadata, particularly when the album has a variety of ensembles, conductors, soloists, instruments etc.
- Lack of complete structure: failure to deal with the fact that much classical music has movements as parts of larger works which are not necessarily the same as physical albums.
- A focus on the Artist rather than the Composer (even assuming the latter is displayed/accessible).
- Player information screens limited to Song, Artist and Album.
- Lack of integration of sleevenotes into the experience.
This has meant that either the classical music lover has had to settle for a poor compromise or (in many cases, I suspect) just give up. Even very expensive network players frequently come with fairly rudimentary tools for getting value from your CD collection. Unless you have plenty of money and pay someone else to rip your collection for you, the investment of time required to do it, given the quality of the resulting experience, does not seem worth it. I decided to go the Logitech Squeezebox Touch route because at the time it seemed a good value DAC with a flexible software solution. Whilst in retrospect I still think this was a good choice; even with the addition of Custom Browse, I became frustrated and disappointed.
Since my early forays into this field, things have improved considerably, but the solutions are not well publicised. The improvement is partly because of the growth in the digital download market for classical music and improving metadata libraries. However, the first major development for me was the purchase of the "Muso" software which deals with several of the problems described. Following that, there have been further improvements in the LMS software which help enormously. Although development of LMS was effectively dropped by Logitech, along with its excellent Squeezebox hardware range (particularly the SB Touch, of which I have two), the development has been taken on by some of the original team and a vibrant user community and is open source. Also, you do not need specialised hardware players – an iPhone with iPeng works just as well (if not better) than a SB Touch and if you want a cheap server/player a Raspberry Pi is hard to beat.
The software heart of my system is therefore LMS + Muso.
My hardware is a Pi (server/player), a Windows laptop which runs Muso, 2 Touches and various iThings running iPeng.
To present the music information in the way I want, I have re-developed my approach to tagging the music to be focussed on MusicBrainz. This is an open music encyclopedia that collects music metadata and makes it available to the public. Furthermore, if it does not have metadata for your music, you can add it so that it does. This helps others as well as you (and means it is available again, should you ever need to re-rip or re-tag your music). Note that my tagging approach does not make assumptions about the player/library manager software and so could easily be used with other software (e.g. MusicBee) - not just LMS/Muso. MusicBrainz provides a free (open source) tagging application - Picard - but it is rather limited because it does not capture the rich classical music metadata available in the MusicBrainz database. To address this shortfall, I have written a custom plugin called "Classical Extras", which is also open source.
To make all this work, my approach in outline is:
- Rip the CDs to FLAC using the ripper of your choice. FLAC files are lossless, but compressed to save space, and provide for a limitless number and variety of tags for metadata. However, not all the metadata is added by the ripping software, so I either:
- Process the ripped files with SongKong. This uses metadata from MusicBrainz and Discogs; or
- Process them in Picard. This only uses metadata from MusicBrainz but incorporates my custom plugin "Classical Extras" which provides a much richer set of data and allows a great deal of customisation.
This still involves a bit of human review along the way, particularly to fix any shortcomings in the available metadata.
Take a look at the Screenshots to see what the result looks like.
If I later spot any errors in my tags, I correct these in MusicBrainz and Picard and re-import the data to Muso.
A simpler approach is just to tag with SongKong, import into Muso and fix any metadata issues directly in Muso. I shall call this the "basic" approach and my preferred method the "enhanced" approach. The advantages and disadvantages of these are discussed.
It is possible to "mix and match" the approach of course - starting with the basic method then enhancing subsequently as necessary i.e. tag in SongKong then import to Muso, but fix any metadata issues (possibly at a later date) with MusicBrainz/Picard, rather than in Muso directly. This may be better if you have a large number of files to tag. Note that SongKong writes MusicBrainz tags, which makes subsequent tagging in Picard simpler and quicker.
See the diagram for an overview of the process.