search

vitrivr / vitrivr-engine

vitrivr's next-generation retrieval engine. It is capable of extracting and retrieving a wider range of multimedia objects such as audio, video, images or 3d models.
https://vitrivr.org
MIT License
5 stars 3 forks source link

[BUG] Incorrect segmentation in FFMpegVideoDecoder when Audio is disabled #119

Open net-cscience-raphael opened 2 days ago

net-cscience-raphael commented 2 days ago

Description

If one disables audio on FFmpegVideoDecoder only one Segment on each source is emitted.

Steps To Reproduce

extract two video files with:

      "decoder": {
        "timeWindowMs": "1000",
        "audio": "false"
      },

Environment

rahelarnold98 commented 1 day ago

After the first segment is emitted, the flags videoReady and audioReady are set to false. But since audio = false, the audioReady flag remains false because there are no audio frames to trigger it again. So, no subsequent segments are actually emitted. Even if additional video frames are collected for further segments, the emit() function isn’t triggered.

In dev #de301d75e015992ea5be82fbeff59aba56a0577a, I changed the condition so that it only depends on videoReady if audio = false.