IP reverse engineering for compatibility is explicitly allowed in many jurisdictions. We wouldn't need to reverse engineer the format if it were patented, that's the whole point of patents: Dolby would have to disclose the details. IIRC they simply don't publish any info and rely on obscurity...
Damn. I know truehd is basically meridian lossless with some extentions like ffmpeg can decode but I was hoping that the metadata track was similar between it and JOC.
I guess someone in a country that doesn't care about software patents has to reverse engineer it from the encoder :(
Boo, ok, figured it wouldn't be that easy.
Do you know of any work on reverse-engineering the Atmos metadata in either format? I'm assuming the E-AC-3 JOC is documented in whatever ATSC/SMPTE standard it got stuffed in.
By the way, for TrueHD/MLP files they tack on the object stream as I believe the 3rd or 4th stream. Might be possible to decode it or at least separate/remux/demux it
See: https://github.com/FFmpeg/FFmpeg/blob/eef763c7057a7f5f4b7dae7855d07b2a6da8b537/libavcodec/mlpdec.c#L388
The reason ffmpeg...
That MSI 2x M.2 M-key -> PCIe CEM x8 (why the fuck does it have an x16 connector...) doesn't look to have a pcie switch on it: they need a heatsink usually and are $$$$
It's very likely using X8->x4x4 bifurcation
90-100dB THD+N is acceptable for something that's not purpose built since the noise floor at my desk is fairly high. Could always use the apple USB-C dongle for an extra few dB!