Digital Developments Stateside, Dream and KiwiSDRs
Kevin Ryan highlights HD Radio developments in the USA, hunts for DRM signals using KiwiSDR Software-Defined Receivers
Kevin Ryan highlights HD Radio developments in the USA, hunts for DRM signals using KiwiSDR Software-Defined Receivers and provides a short tutorial on using the Dream DRM software.
I have occasionally mentioned the digital radio system invented by iBiquity Digital Corporation and trademarked as HD Radio. It was developed for the USA but is used in Mexico and now in Canada too.
The latter is currently struggling to adopt a digital radio standard.
The technology is also known as IBOC (In-Band-On-Channel) because the digital information is added as sidebands around the analogue signal in the same allocated channel space. IBOC is specified for both AM (medium wave) and FM.
Radio stations usually start using IBOC in the hybrid mode, simulcasting the analogue audio service in digital, and possibly adding one or two additional digital audio channels, before switching to all-digital mode.
This is the roadmap foreseen by the designers but not all radio station owners or listeners welcome the technology.
In the USA, AM stations can use channels 30kHz wide (in theory) to broadcast audio with a 15kHz bandwidth, although the majority of stations use nothing like that.
When IBOC is enabled, the analogue signal bandwidth is constrained to 5kHz, and up to three digital signals are inserted into the remaining channel space.
The primary digital signal (this is usually a copy of the analogue audio) is placed the furthest away from the analogue; the secondary programme is in the space in between them, and the tertiary programme shares the same bandwidth as the analogue signal.
This is still an amplitude-modulated signal and there are sidebands above and below the carrier frequency.
Given that stations are planned on a 10kHz grid, the potential for interference from sky waves at night, which are generated by the 50kW transmitters, has made the deployment on medium wave difficult and subject to a lot of criticism.
An unwanted digital signal will decrease the signal-to-noise ratio of an analogue station. By contrast, when both the wanted and unwanted stations are using digital modulation, the effect may not be that noticeable.
On the 16th of July, WWFD on 820kHz in Maryland USA changed from hybrid mode IBOC to the full or all-digital mode (Fig. 1). There were such tests in the past but WWFD has permission from the FCC to operate like this for a year.
WWFD broadcasts with just 4.3kW during the day and carries the Gamut music format. There is conflicting information on the internet that suggests that it uses a Newstalk format.
The night-time power is 430W, so it is unlikely to affect other stations.
There are shades of ‘DRM-thinking’ here. Dave Kolesar, the station’s engineer and programme director, believes that something needs to be done to give AM broadcasting a future. He has decided to take the lead to give AM a chance to compete with FM HD Radio, Sirius XM and streaming radio in cars. The conversion to digital will provide basic, text-based, information on the screen, such as details of the songs being played.
There is the video of a listener tuning into the all-digital signal from WWFD at this URL.
There is another video claiming to show the effects of HD Radio carriers on the AM band. The spectrum display shown in the background lets you see how the digital carriers are packed into a crowded AM band.
I must confess that the speaker lost my attention now and I do not agree with all his observations.
See what you think about his conclusions, by visiting this website:
www.youtube.com/watch?v=15chNbj-GK0
The US FM Band
Like other countries, the USA is debating the future of Band II, in which FM broadcasts are located.
However, unlike European countries (where Band III has been allocated for DAB) the USA still has several TV services operating in this band.
Nautel, a transmitter manufacturer thought about this and came up with a plan to maximize the use of the FM band, if and when the existing stations adopt the all-digital mode in IBOC.
I think they are hoping to hasten the move to all-digital broadcasting.
In the IBOC FM hybrid modes, the analogue audio occupies the spectrum ± 100KHz around the carrier; the digital signals are in the space between ±100 – 200KHz, as shown in Fig. 2.
The same signal is in both the upper and lower sidebands.
In these modes, the HD receiver has to first tune to the FM signal, prior to moving over, as seamlessly as possible, to the digital equivalent.
In the all-digital mode, the space occupied by the FM signal was expected to be filled with more digital carriers providing secondary services. Here, the receiver mutes the non-existent FM signal, before locking onto the digital one.
Today’s receivers haven’t implemented secondary carrier support, and Nautel wants to re-use the vacant spaces where the FM signals used to be.
Nautel has come up with several configurations using multiplexes from 400 to 1200 kHz wide, which can accommodate from nine to 35 audio services, by slotting in 100kHz wide digital blocks, either with a copy of an existing block or at increased power for single blocks to improve robustness.
Of course, no matter how much they adapt the transmission end, the receivers will also have to be changed, in order to be able to cope with the proposed changes to the error correction algorithms.
Like many of these ideas, it needs to be proven ‘in the field’, as it were, and Nautel suggests using vacant (or vacating) spectrum for this.
First, Nautel wishes to use 87.5 to 87.7MHz for a national network or – more correctly – spectrum that could be used for a number of regional networks, as many HD Radios tune down this far in what Nautel calls the ‘European’ mode.
In fact, there are only two FM stations in operation on 87.9MHz on the West Coast, so a bit more space could be allocated in other parts of the USA.
Second, Nautel wants to extend into the 76-88MHz, band currently used for TV channels 5 and 6, and the company quotes industry support for using this spectrum for a digital-only radio band.
I found this idea very interesting. I suppose that the solution has to use existing and familiar technology so that station owners and regulators will adopt it.
Seven (or possibly eight) DAB multiplexes would fit in the proposed new digital-radio band but the FCC won’t go for this technology – just as Europe won’t use HD Radio.
There is a lot of information buried on the Nautel website and here is my suggested reading list if you want to know a bit more:
https://tinyurl.com/ya4ko7km
DRM in China
I have recently received some reports from monitors stating that China’s domestic DRM broadcasts had restarted. I thought that I would like to hear the signal myself, by means of using a remote SDR.
I accessed remote receivers from the Global Tuners group in the past but only for AM reception. I tried that route without success.
It was then that I discovered the KiwiSDR system of web-accessible SDR receivers.
KiwiSDR
The KiwiSDR is a software-defined radio covering 10kHz to 30MHz. It consists of a custom-built circuit board (cape) you connect to one of the BeagleBone Green (BBG) single board computers. You can build one yourself by adding an antenna, power supply and network connection or you can buy them ready built in two versions: the KiwiSDR assembled board or the more complete KiwiSDR kit version including BBG, enclosure and GPS antenna.
Both versions include software supplied on a micro-SD card. It is not a cheap SDR, and shipping will add considerably to the overall cost.
If you are interested, you can find plenty of information on the Seeed [sic] website.
http://wiki.seeedstudio.com/BeagleBone_Green/
www.seeedstudio.com/KiwiSDR-Board-p-2725.html
You can find public SDRs by opening this URL:
You will need an HTML5-capable browser (Chrome and Firefox are the recommended ones) in order to listen to one of the 100 or so public KiwiSDRs anywhere in the world. Up to four people can listen simultaneously to one radio, and each listener tunes independently. The usage, such as 2/4, indicated on the list of public SDRs is not really accurate and you may find that your chosen SDR is, at times, not available.
Choose one site of interest and a screen opens. It displays, from top to bottom, the location of the SDR, a bar with possible station identifications (this is very out of date), the frequency tuning scale with a yellow bandwidth symbol (showing where the SDR is tuned) and a large waterfall display.
At the bottom-left is a welcome text box and at the bottom-right the radio interface (Fig. 3).
Here are some familiar controls; users can enter a numeric frequency, select a broadcast band, utility bands or amateur band, choose reception modes, waterfall zoom, colour controls and an S-meter.
There is some more complete documentation at this URL.
http://kiwisdr.com/ks/using_Kiwi.html
It is a good idea to learn how to use the SDR on AM. I tuned to Radio Thailand in English via a KiwiSDR in Hanoi, Vietnam. I quickly decided that I like this public network, as I have missed many DRM tests in the past by being unaware of its existence.
Decoding DRM from China
China National Radio started DRM testing again on 6030kHz from a transmitter listed as Beijing. The DRM power is 30kW, beamed to northern China. The A18 High Frequency Coordination Conference (HFCC) publication lists the transmission period as 20.25 – 18.05 UTC or 04.25 to 02.05 Beijing time.
The tests began on 24th July, and CNR is testing various modes and bit rates.
Finding a KiwiSDR near China was easy enough and it didn’t take long to learn how to drive the interface using the on-screen help facility. Keep an eye on the tuned frequency, as it seems too easy to get it to move to a random one; perhaps, I am just a bit clumsy.
I tried an SDR at Novosibirsk in central Russia and I could hear the familiar DRM rhythm on 6030kHz. However, when I piped the audio through to the Dream software, using a Virtual Audio Cable, the application didn’t display the signal. It was late in the evening and I couldn’t quite figure this out.
However, by the next morning, I realized that Dream expects to receive an IQ signal; just piping the audio from an AM or LSB/USB demodulator doesn’t work.
The first step was to set the correct output on the KiwiSDR. You can do this by using the IQ mode, the last option on the list of demodulators.
The second step was to use the Virtual Audio Cable software (or its equivalent) to connect this signal to Dream. On my Windows 7 PC, I changed the ‘default playback device’ to ‘cable input’ (V-B Audio Virtual Cable) from the internal speakers.
In standard mode, Dream has no setting that will work with the signal from the KiwiSDR, and I had to launch Dream using a command line prompt with a couple of command line arguments at the end.
I put the following command into a shortcut on my desktop:
“C:\Dream\dream.exe” –c 5
This must be written exactly as shown, in bold type including the quotation marks.
You may have Dream installed in another folder, so you will need to adjust this for your setup.
For example, on another PC where there are a few users with their own accounts, the shortcut is "C:\Users\Kevin\Documents\DREAM DRM\dream.exe" –c 5.
The ‘c 5’ argument tells Dream to select channel 5 for an I/Q signal.
In Dream, you now need to select the correct device by choosing this path:
Settings -> Sound Card -> Signal Input -> Device -> Cable Output (V-B Audio Virtual Cable).
Finally, I had to select the I/Q option within Dream by following this path:
Settings -> Sound Card -> Signal Input -> Channel -> I/Q Pos Split
This will centre the IQ signal correctly for the software. Select the setting View -> Evaluation dialogue and the Power Input Spectrum and you should see the square-topped, digital, signal.
Success
I managed to decode the signal but only after switching to a KiwiSDR at Khabarovsk in Far Eastern Russia, rather than at my initial location of choice, in Novosibirsk.
When I checked on a map, I found that Khabarovsk was directly in line with the beam from the Beijing transmitter.
Fig. 4 shows the Dream receiver screen and part of the evaluation screen. I am really pleased that I can now decode DRM directly from stations worldwide, rather than relying on the reporting of others.
DRM in India
I then discovered a KiwiSDR in New Delhi that has two DRM broadcasts on 828 and 1368 kHz. I have spent ages trying to decode these using Dream.
Others, it would seem, are having the same problem, and the discussion thread is on the DRMRX Forum.
www.drmrx.org/forum/showthread.php?t=2808
A number of contributors have added various theories about why this is happening. The KiwiSDR works on short wave, and I decode the BBC World Service on 15620kHz at 0800 UTC.
I cannot decide whether this issue may be caused by reception conditions, receiver-overloading on medium wave or a form of conditional access to restrict reception to radios made in India (Fig. 5).
AIR is back in DRM on 7550kHz from 1820-2230 to Europe and from 2245 to 0045 UTC to Japan.
Voice of Nigeria
VON returned to broadcasting in DRM on their 15120kHz frequency, starting at or just before 18.00 UTC in late July. I think that the audio sounded a bit better. The station continues to use a very low bit rate in Mode C that has fewer carriers than other modes, and the resultant audio is just 3kHz wide.
I decided to contact them to let them know that their 60-minute news programme was coming in loud and clear and found that their e-mail address did not work. I also found that all contact e-mail addresses for all their language services bounced back.
I sent them a message via Facebook’s Messenger app but received no reply.
Perhaps they don’t want to know if we are listening to them or not!
The Dream DRM Software
I constantly refer to this software and I think most users are self-taught. I did have the original DRM Software Radio installed on my PC but have lost the installer file and kept the licence key file. The Dream software is the easiest way to decode a DRM signal.
The DRM software is located in the SourceForge repository, and the best way to get it up and running on Windows is to download the zip file.
https://sourceforge.net/projects/drm/files/dream/2.1.1
Open the Dream.exe application from whichever folder you installed it in. The window will display Scanning until you tune it to a DRM signal and pipe it through to the decoder.
The menu has four options: File, View (only the Evaluation Dialogue works now), Settings and Help.
The Dream Evaluation Screen
The 2.1.1 version splits the screen into five windows and a cluster of settings down the bottom. The first window shows error rates and frequency offsets but also includes a measure of signal-to-noise ratio (SNR).
As you use the software, you will learn what minimum-value SNR is required for the successful decoding of the various Main Service Channel (MSC) bit rates.
The top-middle box shows the status of decoding; all six LEDs need to be Green, nothing else will do. The MSC, Service Description Channel (SDC) and Fast-Access Channel (FAC) are on the main Dream screen.
The top-right window shows the parameters of the signal and is for information only.
The middle section is a display and a chart selector that controls what is shown there.
Across the bottom, there are the settings that you can leave at their default values.
The Channel Estimation settings are either Wiener or Linear and the statistics they gather are used to create a noise reduction filter. Linear reduces the load on the CPU but is less effective than Wiener.
I always use the latter and I leave the Time Sync Tracking at Guard Energy.
In the middle section (Fig. 6) you can log the quality of the DRM signal to file or save the audio as a .wav file. You should not need to ‘flip the input spectrum’ when using today’s SDRs and virtual cables.
The final setting is MLC: Number of Iterations: Choosing 1, and increasing the number, might improve the decoding rate. However, each improvement in the bit error rate is small.
You can always try it and see if it makes a difference.
The last window (Fig. 6) is the Interferer Rejection. It is either Bandpass Filter or Modified Metrics. You can also apply both. The BF algorithm tries to limit adjacent channel interference, provided it is close in to the DRM signal. On many occasions, this will have no effect. Modified Metrics uses statistics from the individual carriers and attenuates those that are not contributing to the decoding of the audio but may degrade the signal further if a lot of fading is present.
Small-Scale DAB
Ofcom wants to find out how much take-up there might be for a new round of licensing and has asked for expressions of interest to be with them by late September. Ofcom plans to publish a summary of the consultation later in the year.
Summary
DAB has been static over the summer, but HD Radio seems to be waking up to new possibilities. I am glad to note that there has been a good deal of activity on DRM, especially in India, Nigeria and New Zealand. Meanwhile, WINB is fixing problems with its transmitters and will be getting back on the air.
This article was featured in the October 2018 issue of Radio User