00:07 <cone-515> ffmpeg James Almer master:2aab4e4cc0b4: avformat/iamf_writer: disallow Opus extradata with mapping family other than 0

00:07 <cone-515> ffmpeg James Almer master:9ce065c90dec: avformat/iamf_parse: sanitize audio_roll_distance values

00:08 <cone-515> ffmpeg James Almer release/7.0:fdd3e3504eba: avformat/iamf_parse: sanitize audio_roll_distance values

00:08 <cone-515> ffmpeg James Almer release/7.0:1ef18d022373: avformat/iamf_writer: disallow Opus extradata with mapping family other than 0

09:22 <cone-700> ffmpeg Tong Wu master:5c8523cef119: lavc/hw_base_encode: correct the timestamp when input_order = decode_delay

12:10 <elenril> mkver: ping yuvj

12:17 <Lynne> at this point OSS audio was deprecated and unsed for less time than us deprecating YUVJ

12:17 <elenril> isn't OSS still the thing to use on bsd?

12:23 <Lynne> I forgot about bsd :)

12:24 <Lynne> on linux the kernel version was deprecated for 10+ years, but universally distros enabled it as the padsp compat layer wasn't great

12:24 <psykose> poor dre :(

13:36 <kurosu> jamrial: feels like some code behind #ifdef for lcevc could avoid that and only depend on setting need_lcevc &=CONFIG_LIBLCEVC_DEC. Really, it may only make sense if there would be a native decoder

13:37 <jamrial> kurosu: true, but this way i remove a couple branches

13:40 <kurosu> They are likely highly predictable. But doesn't matter much one way or another

13:44 <kurosu> I imagine the real goal will be as an EL to HEVC (Br TV3.0 yadi yada)

14:24 <another|> hmm.. so a post popped up in my mastodon feed, claiming aac_he_v2 was broken with libfdk-aac > 0.1.6

14:25 <another|> does anyone have a build with fdk handy and can reproduce?

14:25 <another|> ref: https://chaos.social/@schenklklopfer/112818587012150780

15:54 <Traneptora> another|: can reproduce

15:55 <Traneptora> https://0x0.st/X9Fv.txt

15:58 <Traneptora> using 1.0.6-1 from arch repos for fdk-aac

16:11 <another|> couple of notes: fdkaac seems to just be a frontend and is irrelevant. libfdk-aac being the relevant lib package

16:13 <Lynne> wbs maintains the package

16:15 <another|> well, this is weird. what does `N-95774-g97fd5d3363` actually mean

16:15 <jamrial> that's the ffmpeg commit

16:15 <another|> I thought it counted commits since origin tag N

16:15 <another|> plus git commit hash

16:16 <another|> but this is just 4 commits back from HEAD

16:17 <another|> and my compile from HEAD has a number ~20k higher

16:18 <Traneptora> libfdk-aac version 2.0.3-1

16:27 <another|> I can reproduce this exact behaviour. however, `c=stereo` fixes it

16:27 <another|> I think SBR needs 2 channels?

16:28 <Traneptora> another|: yea, adding -ac 2 fixes it

16:28 <another|> ah, no

16:29 <another|> the error message is misleading

16:29 <another|> aac_he works fine. what actually fails is parametric stereo

16:30 <another|> e.g. the thing that makes it he_v2

16:30 <another|> s/e.g./i.e./

16:34 <another|> the name kinda gives it away :D

16:42 <Traneptora> another|:I thought SBR was made it he_v2

16:43 <Traneptora> ah, nvm

16:43 <Traneptora> AAC_HE is SBR, and HEv2 is SBR-PS

16:43 <Traneptora> > The usage profile HE-AAC v1 uses spectral band replication (SBR) to enhance the modified discrete cosine transform (MDCT) compression efficiency in the frequency domain.[3] The usage profile HE-AAC v2 couples SBR with Parametric Stereo (PS) to further enhance the compression efficiency of stereo signals.

16:43 <Traneptora> there you go

17:24 <another|> hmm.. looking at b63df9b52705790fb0c6702254a63710d3f007d2

17:25 <another|> >Note that option is available when fdk-aac version (AACENCODER_LIB_VL0.AACENCODER_LIB_VL1.AACENCODER_LIB_VL2) > (4.0.0).

17:25 <another|> current version seems to be 2.0.3 though?

17:28 <another|> ah, okay. that's an internal version number in fdk