It’s been a busy day of analysis to bring this part of my Canberra trip to a posting, but I suppose, it’s probably good that I’ve done the analysis before the data got too stale. In this post, I will detail all of the radio-related data I managed to capture and note down during my time in Canberra.
As normal, I carried a regular whip antenna to use with my receivers, but that’s hardly an optimal solution for a first-floor room inside a brick and concrete building. Luckily enough, the hotel antenna had a standard PAL style socket, and it wasn’t any special sort of TV distribution system – it was mostly just an amplified version of what was received with an external aerial. This provided a superior signal for analysis. Touching the centre conductor to the BladeRF (since I didn’t have the right adapter to go from PAL to SMA), I was able to check the signal quality.
The DVB-T services can be seen with about 23dB SNR or so.
The same can be said of the other DVB-T and DAB services higher up. There is a periodic signal interfering, but that’s most likely to be an FM signal swamping out the SDR’s front-end as when the RTL-SDR was used, the interference was not visible.
The only tampering the hotel did was to analog modulate and inject some of their own in-house channels into the coax. It seems whatever modulator they were using wasn’t too careful about how it did it – it seems to not have bothered to filter out the unnecessary sideband and left it there to potentially interfere with an adjacent channel.
While digital radio has been long established in Sydney, it seems that in Canberra, we’ve entered a time warp where digital radio is still in a trial phase. It seems that this trial has been extended time and time again, rather than just simply entering revenue service. As a result, the scan of the airwaves was somewhat underwhelming. Scanning was done with my E4000 based RTL dongle with DABPlayer.
The service was broadcast on Channel 10B, with a single ensemble only named DAB+ CBR Trial. Reception SNR was only around 10dB using the hotel antenna which is “marginal” but at least, the received error rate was mostly zero. Using just a whip, it was impossible to maintain a lock on the ensemble.
Ensemble utilization is roughly 90% used for audio with no data services carried.
Some of the services carried are clearly of a test nature, whereas others are regular commercial services but with text that indicates that it is part of a trial. The trial seems to be to test coverage requirements based on transmission power, FEC rates, etc.
For those interested, the full technical data dump from DABPlayer made 28th November 2016 is here with a summary of the findings in the table below:
A mixture of services are carried including commercial and non-commercial stations. At the time of survey, ECC and FEC rate was identical for all stations, as was the mode (AAC+), 48khz sample rate, and a PTY of “none”. All services bar one claimed to be broadcasting slideshow data but this was found not to be true.
Instead, surveying the slideshow data proved to be really arduous as I’d wait multiple times on a station for periods of 10 or more minutes waiting for an image that never came. It seems only the SBS properties managed to send out any slideshows at all.
SBS Radio 1
SBS Radio 2
We move straight into digital TV where I used an R820T based dongle purely because of driver compatibility issues. Just like analysis of DVB-T in Sydney, I dumped more than 3 hours for each multiplex and analyzed the recordings once I got home.
It’s important to note that the analysis was performed on recordings made 29-30th November 2016, so they pre-date the launch of ABC HD and are consequently from older data than my recent posting about Freeview in Sydney despite the order of posting.
Just like in Sydney, the same five frequencies in VHF are used for the main multiplexes. The transmissions come from Telstra Tower on Black Mountain in vertical polarity.
The 177.500Mhz multiplex is supplied by SCA and carries mostly Nine Network programs. The 184.500Mhz multiplex is supplied by SBS NSW carrying SBS programs and carries an incorrect frequency of 543.500Mhz in the NIT. The 191.500Mhz multiplex is supplied by ABC Southern NSW carrying ABC programs and carries an incorrect frequency of 206.620Mhz in the NIT. The 219.500Mhz multiplex has a provider of MH Canberra and an incorrect frequency of 585.500Mhz in the NIT. It carries a mixture of WIN and Ten Network programs. Finally, the 226.500Mhz multiplex carries PRIME programs. All stations use the same standardized 64QAM 1/16 GI and 3/4 CR as we do in Sydney.
The channel lineup differs somewhat compared to the Sydney lineup I’m used to. However, this is relatively minor as it seems that their stations are in a bit of a time-warp as well. On the one hand, a number of HD services are in MPEG-2 whereas there are a number of other services which are in MPEG-4 including standard definition services with MPEG-4 audio. This somehow suggests that they’re both “making progress” and “falling behind” all at the same time. The channel variety is somewhat smaller, which should mean less bitrate pressures, but there’s also more shopping channels to contend with.
A look at the video bitrates gives an interesting result. The red and green services are high definition, with red services in MPEG-2 (better compatibility) and green services in MPEG-4 (better compression efficiency). Because of the use of MPEG-2, those HD stations in red consume much more bitrate to maintain accepable quality. As this is pre-launch of ABC HD, ABC News24 remains a 720p50 service. Both SBS HD and 7mate Canberra use horizontally compressed 1440x1080i frame formats, whereas the H.264 based HD services use the full 1920x1080i. Interestingly, from an audio perspective, no channels offered 5.1-channel audio at the time of survey. Some SD channels are using reduced frame formats as well of 704x576i or 640x576i to try and save bitrate.
The purple services are standard definition services using MPEG-4 encoding. WIN NETWORK has just the above still frame broadcast constantly, with a wasteful 384kbit/s audio stream attached. The “To Be Advised” channel is just a black screen with silence. Aside from that, GOLD, ishoptv and Aspire are all shopping channels which people don’t really want to watch, so it’s good to see them taking only about 1Mbit/s each in video. However, since this might affect the audiences which can tune in, TVSN is still in MPEG-2 with a more generous 1.8Mbit/s. Sydney’s shopping channels seem to consume a bit more bitrate.
As with what is seen in Sydney, RACING.COM is in MPEG-4 as well, with 1.6Mbit/s which is much less than the 2.5Mbit/s we have in Sydney. However, unlike in Sydney, 9Life is running regular service in MPEG-4 using just 1.5Mbit/s which is half of what we use in Sydney to bring it in MPEG-2. It’s also celebrating its first birthday.
When it comes to MPEG-2 SD services, there’s a clear gap between the services with “preference” and those without. Note how NITV and SBS VICELAND are in TVSN shopping territory, “making do” with just 1.6-1.7Mbit/s. In Sydney, the respective bitrates are about 2.5-2.9Mbit/s, so it seems they probably do have different “weights” on bitrates in different markets.
However, all of this means that regular MPEG-2 SD bitrates sit in several tiers – 4.3-4.8Mbit/s tier for the stations they really care about, 3.6-3.7Mbit/s for the stations they sort-of care about, 3.4Mbit/s for the stations they have, 3.0Mbit/s for the stations they carry and 1.6-1.8Mbit/s for the stations they don’t care about.
As a result, the bitrate trend they have in Canberra is more spread, but also means that some SD services are still quite higher than the 3-ish Mbit/s average around Sydney. Part of the reason is because of less choice, and another part is due to the adoption of MPEG-4 in some multiplexes, but retaining MPEG-2 HD is a factor that works against this.
In case you care for the per-PID rates, they are summarized in the table above. On the whole, all but the last two muxes show good utilization – the last two could probably spread another 1Mbit/s into their services without major issue.
I’ll also take a look at FM broadcast, although it seems that AM is also used relatively heavily in this area. Unfortunately, I didn’t have any suitable antennas for AM, so I only looked at FM.
My station log is as follows:
98.3Mhz 2XX FM (No RDS) 101.5Mhz Triple J (No RDS) 102.3Mhz ABC Classic FM (No RDS) 103.9Mhz ABC NewsRadio (No RDS) 104.7Mhz Hit104.7 (RDS) 105.5Mhz SBS Radio (No RDS) 106.3Mhz Mix106.3 (RDS)
As you can see, there’s not many stations heard from the hotel room at all. It just points to a relatively “quiet” radio market down here. Only two of the stations were seen carrying RDS – this also seems like a time-warp.
RDS Summary for Hit104.7 PI = 1047 (Outstanding) PS = Hit104.7 TP = 0 TA = 0 M/S = 1 ECC = F0 LIC = DI = 1 PIN = 0. 00:00 PTY = Pop Music / Country (10) PTYN = Local Time = 2016/11/28 (Mon) - 14:59 UTC Time = 2016/11/28 (Mon) - 03:59 0A = 49.9 % (1 of 2.0) 1A = 3.6 % (1 of 28) 2A = 17.9 % (1 of 5.6) 3A = 3.5 % (1 of 28) 4A = 0.1 % (1 of 716) ODA 1: AID = CD46 Proportion = 3.6 % (1/28) App. Group = 8A RDS Summary for Mix106.3 PI = 171B (Outstanding) PS = Mix106.3 TP = 1 TA = 0 M/S = 1 ECC = LIC = DI = 1 PIN = PTY = Other Music / Classical (15) PTYN = Local Time = 2016/11/28 (Mon) - 15:09 UTC Time = 2016/11/28 (Mon) - 04:09 0A = 56.1 % (1 of 1.8) 2A = 19.3 % (1 of 5.2) 3A = 3.5 % (1 of 29) 4A = 0.2 % (1 of 635) ODA 1: AID = CD46 Proportion = 3.5 % (1/28) App. Group = 8A
Both stations that carry RDS carry CD46 data TMC data.
Hit104.7 carries data for HERE CBR.
Mix106.3 carries data for SUNA TMC. I didn’t see any subcarrier audio services.
I did an obligatory scan for different radio services – for example, the Canberra ATIS. The results are shown below, but noting that there is likely a frequency offset as I was using an uncorrected RTL-SDR dongle.
127.450Mhz Canberra ATIS 151.670Mhz Vietnamese Ethnic Radio Service 164.875Mhz Channel Marker - carrier periodic 165.200Mhz DMR Trunking Channel 165.350Mhz Data (Taxi Pager?) 165.500Mhz NFM Voice Trunk Channel (LTR?) 165.910Mhz Trunking Control/Paging/Data 167.645Mhz Data (Taxi Pager?) 416.09375Mhz Data (Taxi Pager?) 416.30625Mhz Data (Taxi Pager?) 416.33125Mhz P25 Voice/Control 416.39375Mhz Bursty Data 416.43125Mhz Data (Taxi Pager?) Weak 416.45625Mhz P25 Voice/Control Strong 416.55625Mhz NFM Voice Channel (Only Carrier) 416.69375Mhz P25 Voice/Control 416.70625Mhz P25 Voice 416.83125Mhz P25 Voice 416.95625Mhz P25 Voice 417.10625Mhz P25 Voice/Control 417.20625Mhz P25 Voice 417.33125Mhz P25 Voice 417.35625Mhz P25 Voice 418.60000Mhz P25 Voice 418.81875Mhz P25 Control 418.83125Mhz P25 Voice 419.08125Mhz P25 Control 419.10625Mhz P25 Voice 419.56875Mhz P25 Control 450.01875Mhz NFM Voice (210.7Hz CTCSS) 450.39375Mhz NFM Voice (229.1Hz CTCSS) 450.56875Mhz NFM Voice (107.2Hz CTCSS) 460.49375Mhz 22khz wide data 460.59375Mhz 22khz wide data 461.14375Mhz Narrow Continuous Data 461.41875Mhz Narrow Continuous Data 461.61875Mhz Narrow Continuous Data 461.71875Mhz Narrow Continuous Data 461.81875Mhz Narrow Continuous Data 461.91875Mhz Narrow Continuous Data 416.16875Mhz Bursty Data System 1 486.86875Mhz Bursty Data System 1 469.19375Mhz Bursty Data System 1 (Strong) 469.34375Mhz Bursty Data System 1 467.86875Mhz Bursty Data System 2 467.89375Mhz Bursty Data System 2 467.94375Mhz Bursty Data System 2 (Strong) 468.14375Mhz DMR Voice Linked 1 468.55625Mhz DMR Voice Linked 1 469.39375Mhz DMR Voice Linked 1 467.86875Mhz DMR Voice Linked 2 467.89375Mhz DMR Voice Linked 2 467.94375Mhz DMR Voice Linked 2 (Strong) 490.86875Mhz DMR Voice 491.56875Mhz Data with pips 492.49375Mhz Digital Voice (TETRA?) 493.26875Mhz Digital Voice (TETRA?
There were a few interesting different data type radios on the air, but as expected, there was a lot of stuff in the harmonized 410-430Mhz government radio spectrum. Other than that, it was very much piecemeal … and lots quieter than in Sydney.
RIP: Telstra 2G GSM Network
While I was in Canberra, it was also D-day for Telstra’s 2G GSM network which was slated to be switched off on the 1st December 2016. I watched as the network scan the night before at 8:37pm had the network still available …
… and the scan made the morning after did not. Farewell Telstra 2G – you have served us well. Unfortunately, for some, this means their old phones will no longer be useful and for others with dual-SIM phones where the secondary SIM radio supports only 2G, they will not be able to use that with Telstra (or likely Vodafone).
This post summarized a lot of analysis of the radio received in Canberra of various sorts, and also brings the planned series of posts on my Canberra conference trip to a close. It was interesting to see the time-warp that is Canberra’s broadcast stations, as well as photograph random things, trace out their mesh Wi-Fi networks, visit ANU and listen to speakers at the conference. But since time was limited, I did as much as I could with my limited time. However, that being said, while this is the end of the series … it might not really be the end just yet …