Real-Time RTMP Ingest Setup and Usage
This article provides an overview of how to set up and use Phenix's Real-time RTMP ingest feature.
Since an RTMP stream is sent from an encoder to Phenix via RTMP, the encoder must be set up to send its output to Phenix. This means that, while the Phenix Customer Portal can be used to obtain information such as the stream key, the Portal’s publishing tab cannot be used to publish RTMP to the Phenix platform.
RTMP publishing requires a stream key. For instructions on how to find your stream key, see How do I find my stream key?
Encoder Configuration for RTMP Ingest
Your encoder's configuration must be set up to ensure that it is not introducing unnecessary latency. The number one issue that causes high latency is not using the correct low-latency settings on the local encoder.
For details on setting up your source, please refer to the page for your encoder (see the menu on the left).
To characterize the latency of the encoder, try measuring the lag using ffplay with the "nobuffer" setting. Using ffplay is always recommended to determine the baseline latency when evaluating encoder settings. The argument will be similar to:
ffplay -fflags nobuffer <url>
If the latency is not extremely low, that indicates that the encoder is not configured correctly for low latency.
Check that encoder settings are set to the lowest possible values that will provide acceptable results, such as baseline profile, CBR, 5 Mbps, 1-second GOP size.
Phenix supports H.264-encoded video as ingest, and recommends using Constrained Baseline Profile up to Level 5.2 with a maximum frame size of 36864 macro-blocks for best performance. Be sure to verify that features such as scene change detection are not enabled to keep the encoding consistent. Some video feeds using may appear to work initially, but will fail when certain unsupported features are used (e.g., when the encoding is optimized for a high motion scene).
The source bitrate must be less than 8Mbps, with no spikes in the bitrate exceeding that value. There must be zerolatency interleaving, and no grouping of audio or video packets. The audio and video packets should be in strict increasing order based on the timestamp in the signal.
The audio must be 48kHz AAC, either mono or stereo, MPEG-2 ADTS, not MPEG-4 LOAS/LATM.
When using single-bitrate output (i.e., when the multi-bitrate capability is not set), the video quality/resolution is not changed; that is, the same video quality input is the quality that is output.
For example, if FHD video is provided to Phenix via RTMP, and the output quality is set to SD, the output will be FHD.
Set bitrate output higher than the Phenix target bitrate for your desired quality level from the supported quality capabilities. Examples:
sdcapability is set in the stream key, Phenix targets 830kbps, so the RTMP encoder should be set to output 1,500kbps
hdcapability is set in the stream key, Phenix targets 1,600kbps, so the RTMP encoder should be set to output 2,500kbps.
Frame rate → set to 30
- Certain video encoding/decoding pipelines will buffer a certain number of frames
Keyframe interval → set to 1 second
For x264 encoding, CPU Usage Preset → set to veryfast
Profile → set to baseline
- OBS → set to zerolatency
- Wirecast , set x264 command line options →
In OBS, these recommended settings will look like the following:
For details on API usage, please refer to the RTMP REST API. To take advantage of Real-time RTMP, you must include the
An example RTMP URL is:
96characterStreamKey is the stream key to be used when publishing, which can be found in the Customer Portal under the Properties tab for a Channel or Room.
RTMP Push Instructions
<streamkey><separator><Phenix RTMP Ingest Options>
where the separator can be either be a semicolon
; or a pipe
|. See RTMP Ingest Options for additional details.
** Common Capabilities: **
- Set desired highest quality layer with
hd(high definition), or
fhd(full high definition)
- Include the
multi-bitratecapability to automatically transcode to multiple quality layers for ABR delivery
streamingcapability to create HLS and DASH formatted streams available if applicable
In OBS, the recommended settings will look like the following:
If you choose not to specify any capabilities, the following default
capabilities will be used:
If some capabilities are specified, but quality layer is not, then the
quality layer will be used.