I find myself replying to so many comments about DRM that I felt making this page would be easier for everyone.
DRM is often considered the future of shortwave radio and people tend to push for broadcasters to move to broadcasting in DRM, myself included, and to me it doesn’t make sense. DRM is a great technology and it has amazing uses for sending data such as webpages, images and more over shortwave but it is not an ideal format for audio to be sent at all and I’ll explain why below.
Unlike with video, when data is discarded from audio it is very noticeable. These noticeable changes are often referred to as artefacts. To have minimal artefacts in audio one needs to use at minimum about 128kbps and to eliminate artefacts something in the range of 300-600kbps is needed. This data quickly becomes very large and sending it is spectrally inefficient compared to sending raw audio.
The digitisation of video was generally a good thing because lots of tricks can be performed to make digital video drastically smaller than analogue video while retaining similar perceptible quality. The same is not true for the digitisation of audio. The UK is a great example of this with their DAB system. A lot of stations broadcast in 64kbps mono over DAB and some are now as low as 32kbps stereo. The artefacts from these low bitrates are very noticeable, especially to younger ears, and sound drastically worse than a decent FM signal.
DRM broadcasts audio encoded between about 10 and 30kbps depending on the amount of error correction and bandwidth used. As you could start to imagine, this is where it’s main problem lies. There are simply not enough bits to send artefact free audio data and a significant portion of those bits are set aside for error correction.
With modern SDR techniques, such as those used in the SDR# software (not sponsored, I’m just a massive fan), it is possible to lower the noise floor of an AM signal and eliminate the majority of fading related issues.
A lot of supporters of DRM claim that DRM achieves “FM quality sound” which is all but impossible using the bitrates that DRM provides in any known codec. Text is only so good at explaining sound quality so here are a few examples of real-world DRM transmissions:
To be fair to DRM I’m including some using the most modern codecs DRM has to offer in these samples as well as the most commonly used ones. These are all “off air” real world demonstrations.
While not bad these are fatiguing for my ears to listen to for extended periods of time.
Now we shall compare that to AM at similar SNR ranges. Now, to be fair to AM I will match the audio SNR to -6dB to the data SNR of the DRM. It is worth noting that every time that the bandwidth of audio doubles the signal looses 3dB so while DRMs signal strength might seem low, if it was the same transmission in AM the signal strength would appear higher since the entirety of the bandwidth is not used in speech or most music.
AM & “Digitally Processed AM” Examples:
R. Joystick (weak signal)
Voice of Korea(very weak)
CRI (medium signal)
A comparison of AM and DRM’s current potential:
Spotify has adopted the use of xHE-AAC for their low data usage option using 24kbps. This allows us to find a song on shortwave and compare it to the realistic maximum potential of that same song over DRM technology. Now I currently have a sample of Mariya Takeuchi’s song Eki from both AM shortwave in CombStereo and also Spotify’s xHE-AAC. Here is a short sample of both. You’ll notice especially the vocal quality is much better on AM even if you can hear the lightning (this signal is not a great one):
To add: receiver compatibility, transmitter compatibility, ease of reception, reliability and fading management etc…
Shortwave has a decent outlook for the future of it as a medium for hobbyists and music shows… but in what form? I hope that the following become commonplace in the future:
- Noise Reduction Technology
- AMSync / DSB modes with reliable carrier locks
- Compatible AM stereo encoding e.g. CQUAM or CombStereo