Zero-Latency Headphone Mixes for Remote Guests with RƎLAY
Remote guests have become a daily fixture on Australian radio, from Triple J interviews recorded over the NBN to ABC Local drive-time crosses that hop between Adelaide studios and a host working from a kitchen in Brisbane. The format is simple: the talent sits in front of a laptop, hears the program feed through their headphones, and talks back without the awkward half-second delay that turns a chat into a game of tennis. Achieving that seamless experience requires more than a good microphone and a stable uplink; it depends on how RƎLAY, the browser-based virtual radio environment, is wired into the monitoring path.
The RƎLAY family of tools was designed for exactly this kind of distributed workflow. Because the software runs inside a standard web browser, producers in Sydney can spin up a virtual mixer for a guest in Perth in minutes, with no physical surface to ship. What makes the configuration unusual, however, is that the monitoring loop must travel back down the same network and out through the guest's local audio interface. Mis-configure even one stage of that round trip and latency creeps in, breaking the conversational feel that radio depends on.
This walkthrough focuses on configuring RƎLAY for delay-free headphone monitoring specifically. The aim is a return feed that arrives in the guest's ears within the perceptual threshold of human conversation, while still leaving the producer full control over mix-minus, talkback and off-air cues.
Australian broadcasters work across a continent that spans three time zones and four primary regulatory jurisdictions under the ACMA, so remote workflows have to be robust enough to survive long-haul NBN links, congested hotel Wi-Fi at Sydney Olympic Park during an AFL final, and the occasional mobile-tethered interview from a regional fair in Tamworth. The settings below assume those everyday realities.
What Latency Actually Sounds Like
Most people notice round-trip delays of around 150 milliseconds as a slight echo; from 250 milliseconds upward, conversations begin to overlap and people instinctively start talking over each other. For a remote guest on a radio panel, anything above 200 milliseconds makes on-air interviews feel stilted and forces the host into unnatural pauses. The goal with RƎLAY is to keep the headphone return path under 100 milliseconds end-to-end, which means every stage of the signal chain has to be audited.
The usual suspects are the browser audio engine, the network codec, the local sound device and the headphone amplifier. RƎLAY uses WebRTC under the hood for peer-to-peer voice, which is good news because it negotiates jitter buffers dynamically. The bad news is that a default buffer setting tuned for music playback will add 40 to 60 milliseconds of round-trip overhead before the producer has even pressed a fader. The trick is to constrain that buffer aggressively while keeping enough headroom to absorb a Wi-Fi hiccup.
Preparing the Local Audio Path
Before touching the RƎLAY interface, the guest's machine needs to be set up correctly. A dedicated USB or Thunderbolt audio interface will always outperform a built-in laptop sound card, both in driver latency and in clock stability. Interfaces from RME, Universal Audio or Focusrite commonly sit in Australian home studios, and most expose an ASIO buffer in the 64-128 sample range that pairs well with RƎLAY's processing block size.
Headphones matter too. Closed-back designs from Beyerdynamic, Sennheiser or Shure prevent the microphone from picking up the return feed, which would otherwise leak into the on-air signal and confuse the AutoMix stage at the studio end. Where spill is unavoidable, the acoustic feedback behaviour of AutoMix is worth reading about, because it informs how aggressively the host can ride faders during a fast-moving interview.
Building the Mix-Minus in RƎLAY VRX
The RƎLAY Virtual Radio Mixer is where the program feed is assembled and the return path is shaped. Each guest gets a dedicated fader strip, and the rule of thumb for delay-free monitoring is to feed the guest everything except their own microphone. Building this mix-minus inside the browser keeps the round trip short, because the signal never has to traverse a hardware mixer in the host's rack.
Within RƎLAY, the cleanest approach is to route the off-air program bus straight to the guest's return channel and bypass any unnecessary DSP. EQ, compression and de-essing should be applied to the on-air bus only, not to the monitor path, since every plugin adds a few samples of delay. The built-in talkback channel can be left enabled so the producer can whisper to the guest off-air without breaking the mix-minus.
Network and Codec Tuning
RƎLAY communicates with the studio over standard IP, and the codec choice has a bigger impact on perceived delay than most operators expect. Opus at 64 kbps gives excellent speech quality with a much shorter algorithmic delay than AAC or older MPEG codecs. Pairing Opus with a constrained jitter buffer of 20-40 milliseconds in the browser settings brings the typical return path down to the 60-90 millisecond range that feels natural in conversation.
Australian networks add their own flavour. An NBN fibre-to-the-premises link in suburban Melbourne will deliver different results from a 4G/5G connection in the Hunter Valley, so the codec bitrate should be set to adapt dynamically. Where the guest is travelling, a wired ethernet adapter plugged into a hotel desk is almost always preferable to the in-house Wi-Fi shared with hundreds of other guests at a convention centre.
Workflow Patterns for Australian Broadcasters
Different networks use remote guests in different ways, and the RƎLAY configuration can be tuned accordingly. A quick comparison of common monitoring path approaches used across Australian radio today highlights the trade-offs.
| Monitoring Path | Typical Round-Trip Latency | Setup Complexity | Best Fit In Australia |
|---|---|---|---|
| Built-in laptop sound card | 180-260 ms | Low | Backup only, not for live crosses |
| USB audio interface with low ASIO buffer | 70-110 ms | Medium | Most Triple J and ABC Local remote setups |
| RƎLAY browser mix-minus with Opus | 60-90 ms | Medium | Drive-time interviews, podcast remotes |
| Power Core hardware return path via AoIP | 20-40 ms | High | Studio-grade links between capital cities |
| External hardware mixer at the guest end | 90-150 ms | High | Legacy setups, rarely recommended |
The middle three rows are where most Australian broadcasters land in practice. Power Core hardware returns are reserved for the most demanding on-air scenarios, such as networked FM coverage across Adelaide, Melbourne and Hobart, where any glitch is heard by tens of thousands of listeners simultaneously.
The headline to remember is simple: keep the browser's audio buffer small, run RƎLAY's mix-minus path without unnecessary DSP, and put the guest's local audio interface on the shortest ASIO setting the machine will hold. Done that way, the conversation in their headphones sounds like the person next to them, which is exactly what makes remote guests feel like they are actually in the studio.
