​Output DASH segments are longer than requested with google cloud transcoder - google-cloud-transcoder

Question

This job sets 2s segments for the video and audio streams. The video total duration is 134s, so I would expect about 67 segments. However we see in the MPD manifest that there are 45 video segments, and 54 audio segments (for each audio track).
Is this the expected behavior? Our player does buffer more than 2s at once.
Why is there a different number of video and audio segments?
Job Config
{
"config":{
"inputs":[
{
"key":"input0",
"uri":"gs://inputfile.mp4"
}
],
"editList":[
{
"key":"atom0",
"inputs":[
"input0"
],
"endTimeOffset":"134.549s",
"startTimeOffset":"0s"
}
],
"elementaryStreams":[
{
"videoStream":{
"codec":"h264",
"profile":"high",
"preset":"veryfast",
"widthPixels":224,
"pixelFormat":"yuv420p",
"bitrateBps":500000,
"rateControlMode":"crf",
"crfLevel":24,
"vbvSizeBits":500000,
"vbvFullnessBits":450000,
"gopDuration":"3s",
"entropyCoder":"cabac",
"frameRate":0.1
},
"key":"thumbnail-lowfps-stream"
},
{
"videoStream":{
"codec":"h264",
"profile":"high",
"preset":"veryfast",
"widthPixels":224,
"pixelFormat":"yuv420p",
"bitrateBps":500000,
"rateControlMode":"crf",
"crfLevel":24,
"vbvSizeBits":500000,
"vbvFullnessBits":450000,
"gopDuration":"3s",
"entropyCoder":"cabac",
"frameRate":1
},
"key":"thumbnail-mediumfps-stream"
},
{
"videoStream":{
"codec":"h264",
"profile":"high",
"preset":"veryfast",
"widthPixels":224,
"pixelFormat":"yuv420p",
"bitrateBps":500000,
"rateControlMode":"crf",
"crfLevel":24,
"vbvSizeBits":500000,
"vbvFullnessBits":450000,
"gopDuration":"3s",
"entropyCoder":"cabac",
"frameRate":5
},
"key":"thumbnail-highfps-stream"
},
{
"videoStream":{
"codec":"h264",
"profile":"high",
"preset":"veryfast",
"widthPixels":224,
"pixelFormat":"yuv420p",
"bitrateBps":500000,
"rateControlMode":"crf",
"crfLevel":24,
"vbvSizeBits":500000,
"vbvFullnessBits":450000,
"gopDuration":"3s",
"entropyCoder":"cabac",
"frameRate":2.23
},
"key":"thumbnail-dynamicfps-stream"
},
{
"videoStream":{
"codec":"h264",
"profile":"high",
"preset":"veryfast",
"widthPixels":400,
"pixelFormat":"yuv420p",
"bitrateBps":500000,
"rateControlMode":"crf",
"crfLevel":24,
"vbvSizeBits":500000,
"vbvFullnessBits":450000,
"gopDuration":"3s",
"entropyCoder":"cabac",
"frameRate":8.919
},
"key":"preview-ld-stream"
},
{
"videoStream":{
"codec":"h264",
"profile":"high",
"preset":"veryfast",
"widthPixels":852,
"pixelFormat":"yuv420p",
"bitrateBps":500000,
"rateControlMode":"crf",
"crfLevel":24,
"vbvSizeBits":500000,
"vbvFullnessBits":450000,
"gopDuration":"3s",
"entropyCoder":"cabac",
"frameRate":30
},
"key":"preview-hd-stream"
},
{
"audioStream":{
"codec":"aac",
"bitrateBps":128000,
"channelCount":2,
"channelLayout":[
"fl",
"fr"
],
"mapping":[
{
"key":"atom0",
"channels":[
{
"inputs":[
{
"key":"input0",
"track":1
}
]
},
{
"inputs":[
{
"key":"input0",
"track":1,
"channel":1
}
]
}
]
}
],
"sampleRateHertz":48000
},
"key":"audio-stream-1"
},
{
"audioStream":{
"codec":"aac",
"bitrateBps":128000,
"channelCount":2,
"channelLayout":[
"fl",
"fr"
],
"mapping":[
{
"key":"atom0",
"channels":[
{
"inputs":[
{
"key":"input0",
"track":2
}
]
},
{
"inputs":[
{
"key":"input0",
"track":2,
"channel":1
}
]
}
]
}
],
"sampleRateHertz":48000
},
"key":"audio-stream-2"
},
{
"audioStream":{
"codec":"aac",
"bitrateBps":128000,
"channelCount":2,
"channelLayout":[
"fl",
"fr"
],
"mapping":[
{
"key":"atom0",
"channels":[
{
"inputs":[
{
"key":"input0",
"track":3
}
]
},
{
"inputs":[
{
"key":"input0",
"track":3,
"channel":1
}
]
}
]
}
],
"sampleRateHertz":48000
},
"key":"audio-stream-3"
},
{
"audioStream":{
"codec":"aac",
"bitrateBps":128000,
"channelCount":2,
"channelLayout":[
"fl",
"fr"
],
"mapping":[
{
"key":"atom0",
"channels":[
{
"inputs":[
{
"key":"input0",
"track":4
}
]
},
{
"inputs":[
{
"key":"input0",
"track":4,
"channel":1
}
]
}
]
}
],
"sampleRateHertz":48000
},
"key":"audio-stream-4"
}
],
"muxStreams":[
{
"key":"thumbnail-lowfps-mux",
"fileName":"small_low_fps.mp4",
"container":"mp4",
"elementaryStreams":[
"thumbnail-lowfps-stream"
]
},
{
"key":"thumbnail-mediumfps-mux",
"fileName":"small_mid_fps.mp4",
"container":"mp4",
"elementaryStreams":[
"thumbnail-mediumfps-stream"
]
},
{
"key":"thumbnail-highfps-mux",
"fileName":"small_high_fps.mp4",
"container":"mp4",
"elementaryStreams":[
"thumbnail-highfps-stream"
]
},
{
"key":"thumbnail-dynamicfps-mux",
"fileName":"small_dynamic_fps.mp4",
"container":"mp4",
"elementaryStreams":[
"thumbnail-dynamicfps-stream"
]
},
{
"key":"preview-ld-mux",
"fileName":"preview_seek.mp4",
"container":"mp4",
"elementaryStreams":[
"preview-ld-stream"
]
},
{
"key":"preview-hd-mux-dash",
"fileName":"preview-hd-mux-dash.m4s",
"container":"fmp4",
"elementaryStreams":[
"preview-hd-stream"
],
"segmentSettings":{
"segmentDuration":"2s"
}
},
{
"key":"audio-stream-1-mux",
"fileName":"audio-stream-1-mux.m4s",
"container":"fmp4",
"elementaryStreams":[
"audio-stream-1"
],
"segmentSettings":{
"segmentDuration":"2s"
}
},
{
"key":"audio-stream-2-mux",
"fileName":"audio-stream-2-mux.m4s",
"container":"fmp4",
"elementaryStreams":[
"audio-stream-2"
],
"segmentSettings":{
"segmentDuration":"2s"
}
},
{
"key":"audio-stream-3-mux",
"fileName":"audio-stream-3-mux.m4s",
"container":"fmp4",
"elementaryStreams":[
"audio-stream-3"
],
"segmentSettings":{
"segmentDuration":"2s"
}
},
{
"key":"audio-stream-4-mux",
"fileName":"audio-stream-4-mux.m4s",
"container":"fmp4",
"elementaryStreams":[
"audio-stream-4"
],
"segmentSettings":{
"segmentDuration":"2s"
}
}
],
"manifests":[
{
"fileName":"preview_main.mpd",
"type":"DASH",
"muxStreams":[
"preview-hd-mux-dash",
"audio-stream-1-mux",
"audio-stream-2-mux",
"audio-stream-3-mux",
"audio-stream-4-mux"
]
}
],
"output":{
"uri":"gs://output-bucket"
},
"spriteSheets":[
{
"format":"jpeg",
"filePrefix":"small_mid_fps_sprite-",
"spriteWidthPixels":224,
"spriteHeightPixels":126,
"columnCount":1,
"rowCount":1,
"startTimeOffset":"0s",
"interval":"1s"
}
]
},
}

Answer 1

The reason this happen is due to gopDuration==3s, and segmentDuration==2s.
gopDuration has to be <= segmentDuration, and, at the same time, segmentDuration has to be divisible by gopDuration.
Once you set gopDuration==2s, you should get what you want.