Rick Sewell is a Chicago engineering manager for Crawford Broadcasting. This article first appeared in the company's November newsletter.
In last month’s column I had proposed the idea that we needed an app for our smartphones that will let us know how well our encoding is making it through various listening environments. In my mind the PPM decoders attached directly to the output of an expensive monitor doesn’t tell the real story. Don’t get me wrong; it has its uses. It definitely will let you know if you have no PPM encoding on your audio so you can take steps to fix that. But it presents a best case scenario that just isn’t happening in the real world.
The idea that we can field test the encoding ourselves in various environments would allow us to know what’s really happening. Since my article went on to be published in Radio World as well as The Local Oscillator, I received some responses on the topic. One came from a fellow broadcast engineer who has been developing software for Windows that does exactly what I was recommending. I was sent a beta copy for me to evaluate.
The software will take recordings, assumedly with PPM encoding present, and yield a report of how many times it was able to decode the station’s encoding during that recording. It also shows a percentage which I assume was some sort of measurement of the actual levels of encoding present on the recording. The software also has a way to evaluate audio that is running live on your Windows computer. I haven’t tried to do too much with that but I would think it might be useful to evaluate encoding on streams.
Now it’s not an app on a smartphone, but it can still work in mobile situations that would simulate certain listening environments. All that was needed was to record the audio and then put that recording into the software for evaluation.
I chose to use a recording app on my Android Phone to record the audio through the microphone of the phone. This was in my mind the best approximation to a PPM meter. I then transferred the recordings to my Windows desktop to evaluate them in the software.
I decided to put the software to the test not only in different environments but with differences in encoding levels. My first tests were in an ideal situation. I had the phone set near my desktop HD Radio. This should be a fairly ideal listening environment having a clear signal with no interference. There was a little of bit of office noise in the background but not very much. I also recorded with my phone in my car, on the interstate with the windows open in the middle of a lot of traffic. One of the worst environments that you could have but still be a normal routine for PPM metering.
To add another variable to the tests I recorded in both listening environments from a station known to have watermark enhancement versus a station that did not have the enhancement. Both were playing similar music. Each recording was approximately three minutes. So I would not only be testing the software and encoding penetration but enhancement versus non-enhancement in two different environments.
The results that I obtained with my phone in my office next to my HD Radio were as follows. For the non-enhanced recording the software showed that it was able to decode the station’s encoding four times during the approximately three minute recording with 27 errors. Now for the other recording in the same environment with the watermark enhancement, the software showed that it was able to decode the encoding 35 times with five errors.
After that I decided to do the mobile test with the following results. For the enhanced watermark recording the software was able to decode the station encoding 12 times during the three-minute recording with 29 errors. For the recording without the enhanced watermark, the software was not able to decode the encoding at all with 25 errors.
I want to add this disclaimer on these results. I would say my tests at this point were not very scientific in the fact that I was using beta software with such a small sampling of recordings. To really make this a more scientific approach you would need more recordings using different recording devices in even more environments. That would be more than I have time to do.
Having said that, with even the small sample size that I conducted it does appear that our encoding is very challenged in less than ideal environments and that there is definitely something to be said for watermark enhancement.