Bink Register Frame Buffer8 New Access

You now label each registered buffer with a frame_number_tag. When Bink finishes decoding, it calls your sync_callback with that tag. This allows triple-buffered async decode without polling.

while (playing) 
    BinkWait(my_bink_handle); // Wait for next frame
    // Bink writes directly to my_8bit_buffer, then calls callback
    // Render pass uses buffer + palette texture 
    SubmitDraw();

The classic pipeline:
Decode video → write to frame buffer → wait for vblank → flip bink register frame buffer8 new

The sync_flags field lets you insert hardware-specific memory barriers: You now label each registered buffer with a

Historically, the signature looked something like this (pseudo-code from Bink v1.x): The classic pipeline: Decode video → write to

void BinkRegisterFrameBuffer8(HBINK bink, void* buffer, 
                              int width, int height, 
                              int stride, int palette_handle);

You gave Bink a pointer to an 8bpp buffer, and Bink would decode frames directly into that memory. The palette_handle referenced a previously registered palette.

Allocate a CPU-visible, write-combined buffer for the 8-bit indices. Ensure it is aligned to cache lines (64 or 128 bytes).

void* my_8bit_buffer = vkAllocateMemory( ..., 
    VK_MEMORY_PROPERTY_HOST_VISIBLE_BIT | 
    VK_MEMORY_PROPERTY_HOST_COHERENT_BIT);