Hector Martin

Wow, just realized it's been this long without a release. I remember fixing 4d406367 two years ago, it's unfortunately it still isn't released...

The FFmpeg encoder doesn't explode, although it drops into the noise bands faster. No aliasing too. It doesn't do as good a job on the subsequent noise section though (it...

Setting OPUS_APPLICATION_RESTRICTED_LOWDELAY yields near "perfect" output (for 32k, anyway):  While OPUS_APPLICATION_VOIP and OPUS_APPLICATION_AUDIO both reproduce the problem, with slight differences in the length of the explosion. That suggests this...

Are you using an up to date libopus? I fixed [this one](https://github.com/xiph/opus/commit/4d40636748da8aefd2e17b18c6897199838af77f) two years ago, and it sounds exactly like what you describe. Looks like there hasn't been a libopus...

The waifu2x neural network actually uses 2x2 pixel duplication for the input (nearest neighbor), so yeah, it's technically a straight filter, not an upscaler.

Re: waifu2x using a box/nearest neighbor filter, see: https://github.com/nagadomi/waifu2x/blob/master/lib/reconstruct.lua#L192

I'm seeing the same issue. Here's a memory graph of two servers, both running carbon-relay-ng with an identical config. carbon-relay-ng was installed between May 28-29. Decreases in cached memory are...

Well, looks like carbon-relay-ng already imports pprof, so that was easy. ``` (pprof) top10 5.41GB of 5.45GB total (99.16%) Dropped 109 nodes (cum

That's possible, though I'm not currently interested in the instrumentation and just left that section untouched from the sample config file :-)

I'm not really very involved with the codebase these days, so I'll defer to @nikias. That said, given certain things I've heard, making the DFU protocol codepath more modular is...