Overview of SRTMiniServer Version 2.7

Overview of SRTMiniServer Version 2.7

SRTMiniServer 2.7 is out (4 September 2026). The main change is control on the line you are actually running: decoder type and SRT latency are no longer global switches. Around that we tightened interlaced HEVC, opened RTMP Enhanced (HEVC) on the built-in RTMP server, and shortened TalkBack return audio.

1. Per-line decoder and SRT latency

Overview of SRTMiniServer Version 2.7

Software / HW / No Decode now sits on each line, not on the whole application. One incoming HEVC can stay on hardware decode, a Routing Pack line on No Decode, a third line on Software.

After an update from 2.6.x every line starts on Software. If you had HW decode before, switch it back where you need it. Routing Pack lines stay on No Decode.

Each line also has its own SRT latency control, so a short studio hop and a long LTE contribution no longer share one value.

2. Interlaced HEVC

Half-height field pictures are now weaved into a complete interlaced frame inside the application. Interlaced HEVC sources are handled consistently, instead of depending on how the decoder happened to emit fields.

The same release also fixes a decoding issue that produced a grey picture on some streams.

3. RTMP Enhanced (HEVC), including DJI Osmo Pocket 4

The built-in RTMP server now accepts RTMP Enhanced streams with HEVC. That includes the DJI Osmo Pocket 4. (more »)

4. Faster TalkBack return audio

Return audio to the field reporter is delivered with lower delay in TalkBack Manager. (more »)

5. Advanced Settings: parameter cloud

Overview of SRTMiniServer Version 2.7

All available parameters are listed next to the editor. Click a name to insert it with its default value; click again to remove it. The RESTART hint appears only for parameters that actually need a restart. (more »)

Ignore Audio and Multicast to are hidden by default. Show them from Advanced Settings with ui_ignore_audio = true and ui_multicast = true. While a control is hidden, its setting is not applied: audio is never ignored, and no multicast address is used. The change is applied without restarting the application; a running line picks it up after that line restarts.

6. FFmpeg 8.1.2 and a Caller Disconnect fix

The whole decoding pipeline is built on FFmpeg 8.1.2.

One more operational fix: pressing Disconnect on a Caller line pointed at an unreachable address did not stop the line — it kept coming back as active. That is fixed.

Download

Download SRTMiniServer 2.7 »>

Full notes: Release notes. Trial is 21 days. GaraninApps Team — srtminiserver.com