The Philips CD-i was capable of playing various file formats: Audio CD's, Photo CD's, Video CD's, enhanced CD+Graphics were supported as well. But not many people know the compatibility of CD-i with the CD BGM format as well. While every CD-i player is capable of playing CD BGM discs as well. What was it about? CD-BGM or CD-BackGround Music is a type of CD defined by Philips, Sanyo and Shinano-Kenshi in the mid 80s. Sometimes the system is being referred to as BMS (Background Music System). CD-BGM is used to store up to 10 hours of audio for the purpose of playing background music in stores, shopping malls, etc. CD-BGM uses ADPCM level B mono audio to accomplish this. The music is stored in 8 tracks, all of which are divided in titles. Usually there are about 15 titles per track, resulting in about 120 songs per disc (about 8 hours). Although dedicated professional CD-BGM players were made available to play the discs (of which the Philips BMS 3000 was the most well-known), every CD-BGM disc also needs to include a CD-i application to allow for playback on a CD-i player.
Jac Goudsmit worked at Codim during the CD-i era and he was responsible of developing these BGM discs. The BGM format was ahead of its time and I wish it would have been easier to produce BGM discs: I would have made many of them for myself. Nowadays you can play MP3 CD-ROMs on almost all DVD players of course but in those days it was the only way to get more than 74 minutes of music on one CD and play it back on something other than a computer.
Jac: "I worked for Codim for about 2 years (starting in 1994) and for most of that time, I was the person who did the actual work on the BGM discs we produced.
Actually it wasn't a lot of work. I didn't have to stay with the computer the whole time, but the music was transferred in real time and the processing was slow, so it took (pretty much) an entire day to make one disc-image, and there were usually about 5 discs per month to make, each with a different genre of music -- a week's work to get them all done. The only way to do that was to have one (expensive) PC, a DAT tape recorder and some other hardware dedicated to making BGM discs.
We got the music for each disc from BMS on two DAT tapes, and a playlist as text file on a floppy disk. The music was always mono, so the left and right channels of the DAT tape contained different music. Most discs had 8 hours of music in total but some only had 4 hours (including the VERY few discs with "light" music on them, i.e. pop music).
We had a Compaq Deskpro PC (I think a 386 at 40MHz but I could be wrong) with a special Sony audio card that had DSP's on board to record audio to harddisk in the exact same ADPCM format as CD-i uses. All the programs we used for the task were programmed in (I think) Turbo C for MS-DOS, as far as I know they were all programmed by one of my predecessors at Codim.
The first task was to record the tapes to harddisk, in ADPCM mode C (or was it B), stereo, resulting in two large files (one for each tape). 2 tapes, 2 hours each tape, so this took 4 hours and required a little bit of interaction to change the tape and restart the program. Sometimes I would forget to stop the recording on time and the harddisk would be too full to do the actual processing so I had to start over...
Then a program would read all the 8 mono files and determine where each tune started, so that the time codes could be added to the track list. This took very long, many hours.
Then once again some interaction was required and this was the hardest - or at least the most annoying - part of the job. We had a program that we could use to play back each of the tracks (according to the detected time codes in the track list) to some headphones so we could verify that the time codes were okay. But sometimes especially with classical music where there are silent gaps in the music, the program would think it was the end of the track and would get all the following tracks wrong. In other cases, the source tapes were so noisy that we had to do several retries on the process to make it actually find the gaps between the tracks. I remember spending many hours editing time codes manually by trial-and-error. Remember this was all done in DOS, no graphics on the screen to point to silences or anything.
After this was all done, we used our Sparc-based Microware CD-i mastering tools to make a disc image, effectively re-multiplexing all the music -- I always thought it would be a lot easier if we could just leave the music multiplexed and don't do all that demultiplexing and re-multiplexing, but I never had time to fix the software. Also the process would have been a lot faster if it wouldn't have been necessary to transfer all the 600+ MB of information over the 10mbit coax network 3 or 4 times.
Next, we would play the image file back on one of our CD-i 605 players via an emulator, and if everything was working correctly, we would write an 8mm Exabyte backup tape with the image file, and/or we would burn a CD-R from the image. I don't know where the master tape went for further processing, but we also got the pressed discs back from the mastering plant, so we could inspect them (and sometimes keep one, shhh don't tell anyone) before we sent them back to BMS along with the DAT tapes. The BMG discs were always unlabeled (i.e. silver on both sides) and packed in little paper envelopes, so BMS (and probably Codim) would save money on printing costs and jewel cases.
Some BGM Discs even have the CD-Interactive logo on it |
BGM discs are basically CD-i formatted discs with a few extra limitations: the audio always has to be ADPCM-C (or was it B), the multiplexed audio file had to start at a certain sector of the disc, and the track list (which was in a certain text format) also had to be in a specific sector location. This was so that the relatively dumb BMG players could play the discs: they didn't actually run the CD-i application (they didn't have video so they didn't need to), they would just read the track list from a specific sector location, and they would play the compressed audio from a specific location too. The BGM disc format specified that CD-i players would also have to be able to play the music in some way, and the format left plenty of space to put a CD-i application on the disc. By the way we had a BMG3000 player, it looked like a CD player with a caddy instead of a tray, and with a couple of extra buttons so you could not only change to the next and previous tune but also to different tracks of the same disc.
I think Philips BMS was competed out of the market by Muzak. Muzak may even have bought BMS from Philips, I don't remember.
When I was hired, the merger was in progress. We were working from a villa in the Parklaan, one of the most beautiful streets in the city of Eindhoven. The merger fell through around the time that my probation period was over, and I was given the choice to continue working for either of the companies. I decided to stay with Codim because my best friend who I had known for years, was already working there before the merger plans began.
When the merger fell through, we moved to another office on the Stratumsedijk, where another merger operation began with a company whose name I forgot. That merger fell through as well.
I have many fond memories of the time I worked for the company: we had a great team of fun people, and we did lots of stuff together. I remember going out for lunch with the other guys to the place with the girl with the pretty eyes; I remember us going to the bar on the corner on Friday afternoons for a drink and a snack, paid for by Codim; I remember I installed radio/TV cable in the building so all the offices could listen to the radio and watch TV; I remember lots of days when I would be there until late in the evening, working (or not) and watching TV.
I also remember not getting along with the boss. One time I worked until 5am on a particularly long and boring Vapro disc. I finally finished it, burned a disc and called a taxicab to take it to the Vapro office, because they HAD to POSITIVELY have it that day, they said. When I came in at 11am or so the same day, I got yelled at by the boss for coming in so late and looking so scruffy, and when I called Vapro the next week, they said they hadn't even had a look at the disc yet. I was so pissed off that I never worked that late again.
All in all, they were good times. Another friend that my other friend and I had known for years, joined the company later. All three of us then worked in one room of the office, which was a lot of fun. He emigrated to the USA a couple of years before I did, and he got me a job at Be Inc. which, just like Codim, is also a company that I'm proud to have worked at. But that's a whole different story."
[Thanks, Jac Goudsmit. Pictures by Andy AVAV Systems]