Why don't we have a "Pre-load & Demux" mode for modern media players?

[zunjue209]

Hi everyone,
I’ve been using standard media players for years, but there is one core limitation in almost all of them (including MPV and PotPlayer) that has bothered me for a very long time. I strongly believe it’s time to rethink it, especially now that our hardware capabilities are heavily underutilized.
The Problem with the Current "Streaming-First" Paradigm
Right now, media players are strictly built on a "streaming-first" logic—they read, demux, and decode frame-by-frame as the video plays. This was perfect 15 years ago when we were running on slow HDDs, 4GB of RAM, and limited VRAM. To achieve "instant opening times," players avoid reading the rest of the file.
But as a user, this creates severe limitations. For example, we cannot see the full-track audio waveform on the seekbar. To see where the video gets quiet, where the loud peaks are, or where someone stops talking, we are forced to drag the file into heavy video editing software (like Premiere or CapCut). In those editing apps, basic playback controls (subtitles, audio track switching, global shortcuts) are clunky.
The Proposal: A Video-Editing Style "Pre-load & Index" Feature
Since modern PC performance has heavily overflowed (we have blazing-fast PCIe NVMe SSDs, 32GB+ of RAM, and high-performance CPUs), we don't care about a 1- or 2-second delay when opening a video anymore.
I propose adding an optional toggle in the settings: "Enable Full Media Pre-loading".
When enabled, the player would act like a lightweight NLE (Non-Linear Editor) upon opening a file:
1. Background Demux & Decode: It spends 0.5 to 2 seconds quickly stripping the container, decoding the audio stream, and storing it in a temporary cache (RAM or a designated temp folder on SSD).
2. Static Waveform Generation: It renders a full, beautiful audio waveform across the entire length of the seekbar immediately.
3. Deep Memory Indexing: It indexes the video frames heavily in the background.
The Massive Benefits of "Trading Space for Experience"
While this approach generates temporary cache files and takes up more RAM/disk space, modern hardware can handle this with absolute ease. The benefits would be game-changing:
• Visual Seeking: Users can instantly "snack" or skip to high-energy parts of a video/lecture just by looking at the sound waves on the timeline.
• Flawless Reverse Play & Scrubbing: Since the streams are fully loaded and indexed, dragging the seekbar backward or forward would be completely lag-free and butter-smooth—eliminating the typical "I-frame choking" or audio pop-ups when jumping around.
• Instant Audio/Subtitle Switching: Zero-latency switching for multi-track Blu-ray files, because everything is already separated in the background.
• Smart Playback Features: It opens the door for features like "Auto-skip silence blocks" (crucial for tutorials and recordings) since the player already knows the exact layout of the entire audio track.
Closing Thoughts
We have entered an era where local storage and computing power are practically infinite for standard video playback. We no longer need to be stingy with RAM or temporary cache files.
I know this requires a fundamental shift in how a player’s caching/demuxing layer works, but I believe it is the missing link that would bridge the gap between traditional media players and modern user habits.
What are your thoughts on this? Is this technically feasible as an extension, a Lua script (for MPV), or a built-in experimental engine option?
Looking forward to hearing from the developers and the community!

[guidocella]

--cache