Audio quality and latency – the first two things that came to mind when considering Cleanfeed. I tested both in a couple of long-distance situations.
First was an actual interview I did with Hills a few months ago via Cleanfeed, he located in London and me in Cincinnati. We spoke on a Friday evening while the seemingly whole of London Netflix was in full force, yet over the course of our 90-minute discussion, I noticed just two very small audio “blips” – impressive.
This was followed by connecting (again from Cincinnati) with my friend and fellow engineer, Martin Garneret, who was mixing a live show in Boston. I sent him an invite via email through Cleanfeed – this invite has the option of never expiring.
Connecting to both his audio from the console and running the TeamViewer/Yamaha Studio Manager combo, I was able to hear and control his QL5 remotely. We were both were quite impressed with the audio quality, only noticing an occasional speed-up for just a moment along with some harmless blips.
Although Cleanfeed’s encoding method is dynamic and the frequency response varies a bit (very high-frequency low-pass filter), I can feed the remote site pink noise and analyze the return using a transfer function to see any frequency discrepancies and adjust my monitoring accordingly. (By the way, I’ve worked up a detailed workflow that you can check out on my SoundNerdsUnite.org website.)
Functions & Choices
Cleenfeed has optimized its process in-step with the Opus codec community (to whom Hills tips his hat) in order to ensure exceptional audio quality with a very low bit rate, which adds up to impressive performance. Hills adds that Cleanfeed goes to great lengths to ensure no processing beyond the Opus codec is made to the audio. There’s a choice of three different encoding bit rates:
1) Speech optimized. The recommended choice for the highest quality speech, with or without light processing such as dynamics compression; mono, approximately 64 kilobytes per second (kbps).
2) Music optimized. The recommended choice for music and production, or other wideband audio, including heavily processed audio such as where multiband compression has been applied; stereo, approximately 172 kbps.
3) Talkback. Where the requirement is clear speech with the minimum technical requirements; mono, approximately 24 kbps.
Depending on network conditions, latency is often lower then a phone call. You can check out the latency specs by clicking the “toolbar” within the Cleanfeed window. I can attest that the latency was not a distraction in my tests. The metering utilizes PPM (peak program meter) ballistics. The meter will turn red if digital clip is reached, so just keep your audio in the green.
Where It’s Going
In addition, Cleanfeed recently added, in beta, the ability to record “sessions,” which I took advantage of during the interview. A few options are available here – a simple “mix” of the session, or each participant can be recorded separately (a great solution for podcasting and remote journalists).
Keep in mind that at the moment, the recording is “buffered” and not streamed to the hard drive. If the browser crashes, you lose the recording. Hills notes that Cleanfeed is hard at work addressing this.
Finally, the audio data is end-to-end encrypted; the company takes security seriously and uses industry standard protocols in this regard.
My take is that Cleanfeed is a smaller, bootstrapped enterprise focused on the professional broadcasting world for revenue while offering very competitive options for us “little guys.” I would also be remiss not to point out other options in this landscape, including SourceConnectNow, Zencastr, and ipDTL.
The bottom line is that in addition to the huge benefits these platforms offer to podcasters and broadcasters, I also see a bright future for true quality audio done remotely for a range of live sound applications.
For my fellow sound nerds in need, the benefits in terms of just teaching and troubleshooting is quite impressive.