Finish NVR integration: multi-camera live/playback views, storage controls, and network administration

[kingcinder]

Goal

Finish the NVR side of BossCamSuite as a simple, robust, operator-usable subsystem.

This is the next major productization target after the current camera-control and composite-editor milestone.

Required features

1. Multi-camera live view layouts

Add live camera viewing layouts for:

• 1 camera
• 3 camera
• 4 camera
• 5 camera
• 6 camera
• 9 camera

Requirements:

• simple layout switching
• dropdown-based camera assignment per pane
• stable live preview routing
• operator-friendly switching between layouts
• preserve currently proven camera-control/editor work

2. Recorded video playback layouts

Add recorded playback window layouts matching live view options:

• 1 camera
• 3 camera
• 4 camera
• 5 camera
• 6 camera
• 9 camera

Requirements:

• per-pane camera selection
• playback layout switching
• synchronized or clearly operator-controlled playback behavior where feasible
• simple playback transport controls

3. Playback time navigation

Implement operator-usable playback selection tools:

• calendar picker for year / month / day
• clock/time-of-day selector
• ability to narrow playback search by time
• camera-specific playback targeting
• simple search and jump flow

4. Storage location controls

Add options for selecting storage/output locations for:

• recorded video
• optional separate audio storage
• screenshot still storage

Requirements:

• configurable paths
• simple UI
• validation and error feedback
• do not silently fail on invalid paths

5. Camera network administration

Add nested menus / structured admin flows for camera networking, including:

• setup from reset
• reprogramming currently connected camera networking settings
• chaining map to connect cameras to:
  • other cameras
  • other AP points
  • broader network access points
• connection strength / signal metrics wherever available
• operator-friendly network administration flow

6. Standard NVR/DVR product features

Add whatever is commonplace / standard fare for NVR/DVR programs, while keeping the system simple, robust, and functional.

Interpret this as including likely/common items such as:

• camera assignment / naming
• snapshot capture
• basic recording controls
• playback navigation controls
• straightforward status visibility
• practical operator workflows

Do not bloat the UI just to claim completeness.

Constraints

• Keep it simple but robust and functional.
• Build on existing BossCamSuite architecture; do not discard current proven camera-control/editor work.
• Preserve anti-regression boundaries already recorded in issue Lock anti-regression boundaries for composite systems and proven firmware blockers #1.
• Do not regress existing structured editors or validated camera-control paths.
• Surface operator workflows cleanly; avoid turning the app back into an engineering console.

Acceptance criteria

• Live view supports 1/3/4/5/6/9 camera layouts
• Playback supports 1/3/4/5/6/9 camera layouts
• Camera selection per pane is operator-usable
• Calendar + clock playback targeting exists and works as a user-facing flow
• Storage locations for video / separate audio / screenshots are configurable
• Networking administration flows exist in structured UI
• Existing build/test remain green
• Existing camera-control functionality does not regress

Notes

This is a product-facing NVR integration milestone, not just backend plumbing. The result should feel like a real NVR/DVR application layer built on top of BossCamSuite’s existing device-control foundation.

[m13v]

the playback time navigation piece is where most NVR projects quietly die. I've watched three of these efforts in the last two years stall right here, because operators don't think in calendar+clock, they think in 'show me when the back gate opened around lunch'. once you build the calendar/clock UI you realize nobody uses it, they want a search box. the layouts are the easy part. the hard part is indexing events so playback selection isn't scrubbing through 24 hours of 9-pane grid. budget more time for the event index than the layout switcher, that ratio is usually backwards in the spec.

[kingcinder]

Thanks, that level of insight is genuinely appreciated, what would be a better approach: a modular overlay where each "highlighted" area can be labeled and motion events within them can then be flagged and thus filterable, or a more holistic and arguably less deterministic AI-size fits all? I'm leaning more towards the former rather than the latter. Curious to read your input on the matter.