In the last part, I decided to design and build myself an audio induction loop receiver. Part of this was just out of pure curiosity as to what sort of signals exist in the electromagnetic spectrum at audio frequencies, but the other part was a desire to see what is broadcast on commercial induction loops. The most regular induction loop systems I have contact with are those of the Sydney Trains M-set (Millenium) and A-set (Warratah) trains. Unfortunately, I haven’t had much contact with the H-Set (Oscars) which also feature audio frequency induction loops. In the case of M and H-sets, the induction loops are only available upstairs, whereas on the A-sets, they are available throughout the carriage.
In theory, getting a signal should be as simple as getting into the carriage, and sitting your butt down in the designated area. For the most part, this holds true for the A-sets, but I haven’t had so much luck with the M-sets. It seems that certain consists have a rather low field emission from their loop systems, and it’s not surprising since some M-sets have strange PA system flaws as well. Even then, I have found that strategically sitting in certain seats or areas, which vary from train to train, results in up to 12dB of difference in signal, with potentially even more in signal to noise ratio because some seats are closer to electrical equipment that makes emissions which interfere with hearing coils. In general, the audio loop transmissions mirror what is heard on the regular PA system, although many trains do have a slight power-up delay of the audio loop which results in clipping of the first syllable or alert tone. It was a jackpot finding for me, because I have been listening to their DVA systems over time, which make changes from time to time, but never really seem to get documented. This was my chance to get a (semi) clean recording. Unfortunately, such matters are always complicated – you can’t control the passengers who will whip out a GSM phone and ruin your day, nor can you control the whines, buzzes and clicks from track circuits, pantograph arcing and other electrical devices on the train. It’s at this time you might be wondering why the T-set (Tangara), C-set, S-set, K-set and V-sets don’t have audio frequency induction loops. Part of this comes down to cost, but another one is interference. I took a ride in a C-set with my receiver and it was a glorious cacophony of PWM
In order to keep the peace and prevent any embarrassment, I’ve decided that all manual announcements will not be uploaded. Some of them can be rather entertaining, others can be somewhat sad, but I think the announcers would appreciate that I don’t make them famous for it. However, I have no qualms about exhibiting the work of, what is essentially, a piece of code. Before every special voice announcement on an A-set, you get this three-tone jingle. This is notably absent on the M-set. Before most automated voice announcements, you get this two-tone jingle, however, it seems many guards have somehow disabled this. The DVA system seems to be versatile, and while most automated messages are in a female voice, there are information phrases which are provided by a male voice, some better-sounding than others. The ones I have collected include the following:
- Emergency help points are for emergency use only
- Help us prevent delays – if you are feeling unwell
- You could be fined for placing your feet on the seats
- If you have an Opal Card
- Please keep your personal belongings with you at all times
- Surveillance cameras continually monitor
- This train terminates here
- From the 1st of September
- Please help to keep your trains clean
It should be noted that many of these announcements are seasonal/fashionable, and are subject to change based on situation and timing. For example, there used to be a “Thank you for travelling with CityRail” phrase, which was later replaced with “Thank you for travelling with us”, but this phrase is rarely scheduled by guards for a playback.
Station Names and Announcements
Given my limited time, I haven’t been able to ride around the whole network quite just yet. That being said, the preferable A-sets travel around the suburban network, so here’s a list of all the stations and (where linked), I have a sample for the station names under the various pre-set phrasing scenarios (i.e. next stop x, this train will stop at x, this station is x). In some cases, I have managed to extract the station names from a stopping pattern message (like this one) but didn’t catch the train through to get the individual phrases. As they are template-synthesized, it should be possible to build these messages manually. Unfortunately, as I only had the chance to catch an M-set down the Bankstown line, the audio for stations on the Bankstown line is notably poorer than the others (where I caught exclusively A-sets). I may, in the future, add more audio to this as the recordings are processed (a time consuming process). If you’re lazy, here’s a file that has all the station names appended into a single file.
These messages seem to be mostly played, but sometimes aborted due to their length and replaced by a manual guard’s announcement. These messages are for specific stations and services which inform customers to change for certain other services or landmarks.
- Lidcombe, and another Lidcombe
- Macquarie University
- Town Hall
- Wynyard, and another Wynyard
If you don’t have an audio induction receiver, then your recordings will sound like this. Not as nice, eh. But that doesn’t mean I don’t like regular audio … for example, this one at Town Hall makes me feel like I’m there. What about the wonder of Digital Signal Processing (DSP)? Well, the noisy Opal special announcement filtered through DSP gains a “swimminess” to it, but it otherwise much quieter.
Interestingly, while audio frequency induction loops are normally claimed to have marginal sound quality, in many cases with the right positioning, the sound quality was excellent. It’s probably the first concerted effort to document the Sydney Trains M-set/A-set DVA system (which shares common phrases and voices). Train announcements have come a long way from the crackling, unintelligible, guard-specific announcements given in older silver-tin-can trains. Even then, with the automated DVA system, it seems that each guard has their own personality when it comes to scheduling or producing certain types of announcements, and some still prefer to give manual announcements from time to time. It’s been a big effort to get this one online, with a total of over 300 files produced and catalogued. I hope you enjoy it all. All of the recordings are made by me personally, and I reserve the rights to them. If you do have any particular use for these samples, please let me know and get in contact.