More detailed errors - Exception: add mask expects relative coordinates only

[Mincka]

Hi,

I have this error Exception: add mask expects relative coordinates only when I try to use the latest dev version.
I read in the topic for new UI of 0.14 that's because one my mask has bigger coordinates than the detect stream.

The only issue is that I have dozen of complex masks and it will take very long time to check them all.

Would it be possible to show with the error message:

• Content of the coordinates with the issue
• And / or the line of the config file
• And / or another reference to the mask or zone and the affected camera

Thank you

[NickM-27]

There's no way to know the line number or the camera while building the config model here. It's possible we can show the original line

[hawkeye217]

Just put up a PR to give you the content of the problematic coordinates. #11292

[Mincka]

I removed the camera one by one and I think I found the problematic mask.
This camera is pulled Wyze Bridge and I believed at first that the resolution had changed, but that's not the case and Frigate sees it as 1280x720.

  Frigate_Entree:
    ffmpeg:
      output_args:
        record: preset-record-generic-audio-aac
      inputs:
        - path: rtsp://10.0.0.84:9554/entre
          roles:
            - detect
            - record
            - audio
    audio:
      enabled: true
      listen:
        - bark
        - fire_alarm
        - scream
        - yell
    snapshots:
      enabled: true
      retain:
        default: 1
    record:
      enabled: true
      retain:
        days: 0.5
      events:
        retain:
          default: 0.5
    objects:
      track:
        - dog
        - person
    motion:
      mask:
        - 867,720,1280,720,1280,666,859,667

As soon as I had this mask, Frigate fails to start.

[NickM-27]

I would suggest removing the mask and see what the detect resolution is when pulling up /api/config

[Mincka]

You were right, the detect resolution is 360p. 👍
It's confirmed by my option in the Wyze bridge that is set to QUALITY=SD30 to 360p...
I was mislead by the image I saved in Frigate, which has the 1280x720p resolution.
There's a "height": 720 in live in the json, what is it? Is it true resolution or some kind of upscale?

It would be perfect if it was possible to catch this discrepancy (coordinates higher than the detect resolution), and throw a more precise exception.

I still need to check my other masks since I still get the issue with other cameras. 😥
Thankfully, this will be a problem of the past when everything will be migrated to relative coordinates.

"Frigate_Entree": {
            "audio": {
                "enabled": true,
                "enabled_in_config": true,
                "listen": [
                    "bark",
                    "fire_alarm",
                    "scream",
                    "yell"
                ],
                "max_not_heard": 30,
                "min_volume": 500,
                "num_threads": 2
            },
            "best_image_timeout": 60,
            "birdseye": {
                "enabled": true,
                "mode": "continuous",
                "order": 0
            },
            "detect": {
                "annotation_offset": 0,
                "enabled": true,
                "fps": 5,
                "height": 360,
                "max_disappeared": 25,
                "min_initialized": 2,
                "stationary": {
                    "interval": 50,
                    "max_frames": {
                        "objects": {}
                    },
                    "threshold": 50
                },
                "width": 640
            },
            "enabled": true,
            "ffmpeg": {
                "global_args": [
                    "-hide_banner",
                    "-loglevel",
                    "warning",
                    "-threads",
                    "2"
                ],
                "hwaccel_args": "preset-vaapi",
                "input_args": "preset-rtsp-generic",
                "inputs": [
                    {
                        "global_args": [],
                        "hwaccel_args": [],
                        "input_args": [],
                        "path": "rtsp://10.0.0.84:9554/entre",
                        "roles": [
                            "record",
                            "detect",
                            "audio"
                        ]
                    }
                ],
                "output_args": {
                    "detect": [
                        "-threads",
                        "2",
                        "-f",
                        "rawvideo",
                        "-pix_fmt",
                        "yuv420p"
                    ],
                    "record": "preset-record-generic-audio-aac"
                },
                "retry_interval": 10
            },
            "ffmpeg_cmds": [
                {
                    "cmd": "ffmpeg -hide_banner -loglevel warning -threads 2 -hwaccel_flags allow_profile_mismatch -hwaccel vaapi -hwaccel_device /dev/dri/renderD128 -hwaccel_output_format vaapi -user_agent FFmpeg Frigate/0.14.0-2be15b6 -avoid_negative_ts make_zero -fflags +genpts+discardcorrupt -rtsp_transport tcp -timeout 5000000 -use_wallclock_as_timestamps 1 -i rtsp://10.0.0.84:9554/entre -f segment -segment_time 10 -segment_format mp4 -reset_timestamps 1 -strftime 1 -c:v copy -c:a aac /tmp/cache/Frigate_Entree@%Y%m%d%H%M%S%z.mp4 -r 5 -vf fps=5,scale_vaapi=w=640:h=360:format=nv12,hwdownload,format=nv12,format=yuv420p -threads 2 -f rawvideo -pix_fmt yuv420p pipe:",
                    "roles": [
                        "record",
                        "detect",
                        "audio"
                    ]
                }
            ],
            "live": {
                "height": 720,
                "quality": 8,
                "stream_name": "Frigate_Entree"
            },
            "motion": {
                "contour_area": 10,
                "delta_alpha": 0.2,
                "enabled": true,
                "enabled_in_config": true,
                "frame_alpha": 0.01,
                "frame_height": 100,
                "improve_contrast": true,
                "lightning_threshold": 0.8,
                "mask": "",
                "mqtt_off_delay": 30,
                "raw_mask": null,
                "threshold": 30
            },
            "mqtt": {
                "bounding_box": true,
                "crop": true,
                "enabled": true,
                "height": 270,
                "quality": 70,
                "required_zones": [],
                "timestamp": true
            },
            "name": "Frigate_Entree",
            "objects": {
                "filters": {
                    "amazon": {
                        "max_area": 24000000,
                        "max_ratio": 24000000,
                        "min_area": 0,
                        "min_ratio": 0,
                        "min_score": 0.7,
                        "threshold": 0.7
                    },
                    "car": {
                        "max_area": 24000000,
                        "max_ratio": 24000000,
                        "min_area": 0,
                        "min_ratio": 0,
                        "min_score": 0.5,
                        "threshold": 0.8
                    },
                    "dog": {
                        "max_area": 24000000,
                        "max_ratio": 24000000,
                        "min_area": 0,
                        "min_ratio": 0,
                        "min_score": 0.5,
                        "threshold": 0.8
                    },
                    "face": {
                        "max_area": 24000000,
                        "max_ratio": 24000000,
                        "min_area": 0,
                        "min_ratio": 0,
                        "min_score": 0.7,
                        "threshold": 0.7
                    },
                    "fedex": {
                        "max_area": 24000000,
                        "max_ratio": 24000000,
                        "min_area": 0,
                        "min_ratio": 0,
                        "min_score": 0.7,
                        "threshold": 0.7
                    },
                    "license_plate": {
                        "max_area": 24000000,
                        "max_ratio": 24000000,
                        "min_area": 0,
                        "min_ratio": 0,
                        "min_score": 0.7,
                        "threshold": 0.7
                    },
                    "person": {
                        "max_area": 24000000,
                        "max_ratio": 24000000,
                        "min_area": 0,
                        "min_ratio": 0,
                        "min_score": 0.5,
                        "threshold": 0.85
                    },
                    "ups": {
                        "max_area": 24000000,
                        "max_ratio": 24000000,
                        "min_area": 0,
                        "min_ratio": 0,
                        "min_score": 0.7,
                        "threshold": 0.7
                    }
                },
                "mask": "",
                "track": [
                    "dog",
                    "person"
                ]
            },
            "onvif": {
                "autotracking": {
                    "calibrate_on_startup": false,
                    "enabled": false,
                    "enabled_in_config": false,
                    "movement_weights": [],
                    "required_zones": [],
                    "return_preset": "home",
                    "timeout": 10,
                    "track": [
                        "person"
                    ],
                    "zoom_factor": 0.3,
                    "zooming": "disabled"
                },
                "host": "",
                "port": 8000
            },
            "record": {
                "enabled": true,
                "enabled_in_config": true,
                "events": {
                    "post_capture": 5,
                    "pre_capture": 5,
                    "retain": {
                        "default": 0.5,
                        "mode": "motion",
                        "objects": {}
                    }
                },
                "expire_interval": 60,
                "export": {
                    "timelapse_args": "-vf setpts=0.04*PTS -r 30"
                },
                "preview": {
                    "quality": "medium"
                },
                "retain": {
                    "days": 0.5,
                    "mode": "all"
                },
                "sync_recordings": false
            },
            "review": {
                "alerts": {
                    "labels": [
                        "person",
                        "car"
                    ],
                    "required_zones": []
                },
                "detections": {
                    "required_zones": []
                }
            },
            "snapshots": {
                "bounding_box": true,
                "clean_copy": true,
                "crop": false,
                "enabled": true,
                "quality": 70,
                "required_zones": [],
                "retain": {
                    "default": 1,
                    "mode": "motion",
                    "objects": {}
                },
                "timestamp": false
            },
            "timestamp_style": {
                "color": {
                    "blue": 255,
                    "green": 255,
                    "red": 255
                },
                "format": "%m/%d/%Y %H:%M:%S",
                "position": "tl",
                "thickness": 2
            },
            "ui": {
                "dashboard": true,
                "order": 0
            },
            "zones": {}
        },

[Mincka]

The other camera with an issue has a stranger case.
That's a Reolink RLC-410W, and I read the Fluent stream, which is 640x360.
My mask were limited to 640x360, BUT, in the API, I see it's detected as 640x352, for some reason.
It made the boot failed since 360 > 352.

            "detect": {
                "annotation_offset": 0,
                "enabled": true,
                "fps": 5,
                "height": 352,
                "max_disappeared": 50,
                "min_initialized": 2,
                "stationary": {
                    "interval": 50,
                    "max_frames": {
                        "objects": {}
                    },
                    "threshold": 50
                },
                "width": 640
            },

[NickM-27]

in general this is why we suggest setting detect resolution in the config. In this case this is what opencv is reading the stream dimensions as