profile
viewpoint
Thomas Daede tdaede Mozilla Mountain View

tdaede/anubis 3

x68k accelerator

tdaede/daala_zynq 3

Project to run Daala hardware decoder on Zedboard

tdaede/daala 2

Daala repository for HW assisted decoder

tdaede/dve 2

Distributed Video Encoder

tdaede/franklin-tools 2

ADT bootstrapping tools for the Franklin Ace 1000/1200

tdaede/aom 1

personal aom tree, do not use

tdaede/aperturephones 1

3d printed steampunk headphones

issue commentxiph/rav1e

Deploy Arm64 builds for macOS and Windows on (pre-)release

  • aarch64-apple-darwin requires bigSur in order to link correctly and bigSur is not yet stable on github that I know, we have plans to leverage the kind offer from https://www.macstadium.com/, but it isn't in place yet.
  • aarch64-pc-windows-msvc should just be a matter of adding the correct target on the Windows deploy.

@EwoutH if you send a patch for the latter would be great while I'll figure out how to set up the CI runner on the macstadium system.

EwoutH

comment created time in 11 minutes

pull request commentxiph/rav1e

Add few missing cdfs to the map

btw help checking all are covered would be welcome.

The following shows that tx_size_cdf and txfm_partition_cdf are uncovered.

for CDF in `grep -o 'pub .*_cdf[^:]*' src/context/cdf_context.rs | cut -f2 -d' '`
do grep -q "${CDF}_start =" src/context/cdf_context.rs || echo "$CDF"
done
lu-zero

comment created time in 4 hours

pull request commentxiph/rav1e

Deploy CI: Reformat, simplify Nasm installation, Cargo C 0.7.1

Coverage Status

Coverage decreased (-0.5%) to 73.515% when pulling 86cccd06e63b03b97639de866d6802a637472f65 on EwoutH:patch-35 into cae1980267937fa352be29edc95cb8d77ee4cbfb on xiph:master.

EwoutH

comment created time in 10 hours

issue openedxiph/rav1e

Deploy Arm64 builds for macOS and Windows on (pre-)release

It would be great to have pre-build binaries for the aarch64-apple-darwin and aarch64-pc-windows-msvc targets deployed to GitHub on releases and pre-releases.

created time in 10 hours

PR opened xiph/rav1e

Deploy CI: Reformat, simplify Nasm installation, Cargo C 0.7.1
  • Removes empty lines and reformat single-line run: commands
  • Simplify Nasm installation by using ilammy/setup-nasm also on Ubuntu
  • Update Cargo C to 0.7.1 (also for the regular CI workflow)

I tested the updated deploy config on this release.

+16 -52

0 comment

2 changed files

pr created time in 10 hours

pull request commentxiph/rav1e

CI: Run on pushes and PRs to 0.x maintenance branches

Coverage Status

Coverage decreased (-0.5%) to 73.515% when pulling d3199d580c71787aad8a872dc3d8f06919a9fbab on EwoutH:patch-34 into cae1980267937fa352be29edc95cb8d77ee4cbfb on xiph:master.

EwoutH

comment created time in 11 hours

PR opened xiph/rav1e

CI: Run on pushes and PRs to 0.x maintenance branches

Run the CI also on pushes and pull request to the 0.x maintenance branches (0.3, 0.4 etc.)

For this PR to have effect, it needs also to be merged to the 0.3 and 0.4 branches. Are separate PRs preferred, or can this commit be cherry picked?

+2 -0

0 comment

1 changed file

pr created time in 11 hours

push eventxiph/rav1e

Luca Barbato

commit sha cae1980267937fa352be29edc95cb8d77ee4cbfb

Remove the js api from the main tree It lives in a separate repository now.

view details

push time in 17 hours

PR merged xiph/rav1e

Remove the js api from the main tree

It lives in a separate repository now.

Closes #2553

+1 -1885

1 comment

37 changed files

lu-zero

pr closed time in 17 hours

issue closedxiph/rav1e

Move `rav1e_js` out of the main encoder tree

Hi,

Right now rav1e_js living in the main tree is superfluous to the encoder and its functionality and also additional CI run-time, I would like to propose to move $root/rav1e_js as a whole to a separate repo. So it is better and cleaner. Either we could keep it in rust-av or maybe in Xiph itself, I think it might be better if it lives in rust-av like rav1e-by-gop and later moves to Xiph if it is matured and required.

closed time in 17 hours

vibhoothiiaanand

issue commentxiph/vorbis

cross compilation broken on android

I'm not very good at developing for Android.

I can only advise you to build the required libraries using CMake, install in a specific location and then find them using find_package.

pmp-p

comment created time in a day

issue commentxiph/vorbis

cross compilation broken on android

Hi, I am having this hame issue

[cmake] Could NOT find Ogg (missing: OGG_LIBRARY OGG_INCLUDE_DIR)

I am using

libogg -> 1.3.4 libvorbis -> 1.3.7

and in my CMakeLists.txt:

.... add_subdirectory("ext/libogg") add_subdirectory("ext/libvorbis") ....

pmp-p

comment created time in a day

pull request commentxiph/rav1e

Remove the js api from the main tree

Coverage Status

Coverage increased (+7.9%) to 74.042% when pulling 54f8f1478311464068c9d6c7b63a6bae1c918ca1 on rust-av:remove-rav1e-js into eab1cd61c1c6a7d65d49726df62a99bd0a97b8fd on xiph:master.

lu-zero

comment created time in a day

PR opened xiph/rav1e

Remove the js api from the main tree

It lives in a separate repository now.

+1 -1885

0 comment

37 changed files

pr created time in a day

PR opened xiph/rav1e

rollback the CDFContext by field WorkInProgress

In theory we change the cdf only using symbol_with_update so we can recycle the FieldMap to copy back the single fields using copy_nonoverlapping.

In practice something does not quite work yet.

+176 -101

0 comment

4 changed files

pr created time in a day

pull request commentxiph/rav1e

Add few missing cdefs to the map

btw help checking all are covered would be welcome.

lu-zero

comment created time in 2 days

pull request commentxiph/rav1e

Add few missing cdefs to the map

Coverage Status

Coverage increased (+0.05%) to 66.156% when pulling e91d2dce0ed6e7f29a0919d9feb2067003d4778e on rust-av:missing-cdf into eab1cd61c1c6a7d65d49726df62a99bd0a97b8fd on xiph:master.

lu-zero

comment created time in 2 days

pull request commentxiph/rav1e

Add few missing cdefs to the map

nit: re the commit message, CDFs (CDEF is unrelated here).

lu-zero

comment created time in 2 days

PR opened xiph/rav1e

Add few missing cdefs to the map

We should make sure that all the fields are instrumented

+22 -0

0 comment

1 changed file

pr created time in 2 days

issue commentxiph/rnnoise

Has anyone built this and tested the demo in Windows>

I guess #88 indicate that someone is working on a Windows edition. It does not adjust the file opening mode, but perhaps it should.

sumanrbt

comment created time in 2 days

issue commentxiph/rnnoise

Mozilla Public License?

As it is perfectly possible to use BSD lisenced software in commercial products, I suspect this request is based on some misunderstanding. @sonialouise, care to explain a bit more what kind of problem you believe relicensing will solve?

sonialouise

comment created time in 2 days

issue commentxiph/rnnoise

Please tag releases

Having release tarballs and tags would also make it easier to package rnnoise in Distributions. I've asked for a Debian package in https://bugs.debian.org/980839 and did not know what to put in the version entry. Just a minor issue but exemplify the use of a version number.

darix

comment created time in 2 days

issue commentxiph/rav1e

Unaligned memmove is > 20% of encode time at -s0

according to callgrind:

--------------------------------------------------------------------------------
         Ir  file:function
--------------------------------------------------------------------------------
298,088,019  ???:__memcpy_avx_unaligned_erms [/lib64/libc-2.32.so]
253,325,238  /home/lu_zero/Sources/rav1e/src/me.rs:rav1e::me::full_pixel_me [/home/lu_zero/Sources/rav1e/target/release/rav1e]
243,043,970  .//home/lu_zero/Sources/rav1e/src/x86/sad_sse2.asm:0x00000000002b4d80 [/home/lu_zero/Sources/rav1e/target/release/rav1e]
123,545,730  /home/lu_zero/Sources/rav1e/src/asm/x86/dist/mod.rs:rav1e::me::full_pixel_me
122,318,416  /home/lu_zero/Sources/rav1e/src/rdo.rs:rav1e::rdo::cdef_dist_wxh_8x8 [/home/lu_zero/Sources/rav1e/target/release/rav1e]
 80,727,559  /home/lu_zero/Sources/rav1e/src/tiling/plane_region.rs:rav1e::me::full_pixel_me
 75,656,947  /home/lu_zero/Sources/rav1e/src/context/transform_unit.rs:rav1e::context::transform_unit::<impl rav1e::context::cdf_context::ContextWriter>::get_nz_map_contexts [/home/lu_zero/Sources/rav1e/target/release/rav1e]

Digging further:

image

And further:

image

So we have BlockContext::rollback() and ContextWriter::rollback().

Ideally we could instrument the CdfContext struct further. It is 17774 bytes.

negge

comment created time in 2 days

pull request commentxiph/rnnoise

Windows and CMake support

This pull request also remove two symbols from the list of exported symbols in the library.

Anyway, I suspect the different parts of this pull request should be handled and evaluated separately. At least I believe the introduction of a parallel build system should be decided independently of moving several local variables to the heap. Perhaps better to make separate pull reqeusts for each set of changes?

almogh52

comment created time in 2 days

issue openedxiph/rav1e

Unaligned memmove is > 20% of encode time at -s0

$ perf record target/release/rav1e Bosphorus_1920x1080_120fps_420_8bit_YUV.y4m -o /dev/null --tiles 4 --threads 4 --limit 64 -s0
Couldn't synthesize bpf events.
>  Using y4m decoder: 1920x1080p @ 30/1 fps, 4:2:0, 8-bit
>  Encoding settings: keyint_min=12 keyint_max=240 quantizer=100 bitrate=0 min_quantizer=0 low_latency=false tune=Psychovisual rdo_lookahead_frames=40 min_block_size=4x4 max_block_size=64x64 multiref=true fast_deblock=false reduced_tx_set=false tx_domain_distortion=false tx_domain_rate=false encode_bottomup=true rdo_tx_decision=true prediction_modes=Complex-All include_near_mvs=true no_scene_detection=false cdef=true use_satd_subpel=true non_square_partition=true enable_timing_info=false fine_directional_intra=true
>  CPU Feature Level: AVX2
>  Using 4 tiles (2x2)
>  encoded 64/64 frames, 0.100 fps, 2342.27 Kb/s, est. size: 0.60 MB, est. time: 0s,  elap. time: 10m 38s                        
>  ----------
>  Key frame:             1 | avg QP:  79.00 | avg size:  104170 B
>  Inter frame:          63 | avg QP: 138.00 | avg size:    8260 B
>  Intra only frame:      0 | avg QP:   0.00 | avg size:       0 B
>  Switching frame:       0 | avg QP:   0.00 | avg size:       0 B
>  ----
[ perf record: Woken up 1229 times to write data ]
[ perf record: Captured and wrote 308.871 MB perf.data (8096427 samples) ]

$ perf report
# Overhead  Command  Shared Object       Symbol                                                                                                                                                                   
# ........  .......  ..................  .........................................................................................................................................................................
#
    20.33%  rav1e    libc-2.30.so        [.] __memmove_avx_unaligned_erms
    11.39%  rav1e    rav1e               [.] rav1e::encoder::encode_tx_block
     9.44%  rav1e    rav1e               [.] <rav1e::ec::WriterBase<S> as rav1e::ec::Writer>::symbol_with_update
     7.59%  rav1e    rav1e               [.] rav1e::rdo::cdef_dist_wxh_8x8
     5.29%  rav1e    rav1e               [.] rav1e::context::block_unit::<impl rav1e::context::cdf_context::ContextWriter>::write_coeffs_lv_map
     3.72%  rav1e    rav1e               [.] rav1e::context::transform_unit::<impl rav1e::context::cdf_context::ContextWriter>::get_nz_map_contexts
     3.60%  rav1e    rav1e               [.] rav1e::asm::x86::transform::forward::forward_transform_avx2
     2.68%  rav1e    rav1e               [.] rav1e::encoder::encode_block_post_cdef
     2.22%  rav1e    rav1e               [.] rav1e::rdo::compute_distortion
     1.85%  rav1e    rav1e               [.] <rav1e::ec::WriterBase<S> as rav1e::ec::Writer>::bit
     1.77%  rav1e    rav1e               [.] __rust_probestack
     1.43%  rav1e    rav1e               [.] rav1e::rdo::luma_chroma_mode_rdo::_$u7b$$u7b$closure$u7d$$u7d$::hfd13f0685c9817e4
     1.21%  rav1e    rav1e               [.] rav1e::quantize::QuantizationContext::update
     1.03%  rav1e    rav1e               [.] rav1e::asm::x86::quantize::dequantize_avx2
     0.94%  rav1e    rav1e               [.] rav1e::partition::get_intra_edges
     0.82%  rav1e    rav1e               [.] rav1e::context::block_unit::BlockContext::get_txb_ctx
     0.78%  rav1e    rav1e               [.] rav1e::predict::PredictionMode::predict_inter_single
     0.77%  rav1e    rav1e               [.] rav1e::encoder::write_tx_blocks
     0.76%  rav1e    rav1e               [.] rav1e::asm::x86::transform::forward::daala_fdct32
     0.72%  rav1e    rav1e               [.] rav1e::asm::x86::transform::forward::daala_fdct_ii_16
     0.64%  rav1e    rav1e               [.] rav1e::transform::forward_shared::Txfm2DFlipCfg::fwd
     0.62%  rav1e    rav1e               [.] rav1e::rdo::rdo_tx_size_type
     0.61%  rav1e    rav1e               [.] core::cmp::PartialOrd::le
     0.61%  rav1e    rav1e               [.] rav1e::asm::x86::transform::forward::daala_fdct8
     0.59%  rav1e    rav1e               [.] rav1e::context::block_unit::<impl rav1e::context::cdf_context::ContextWriter>::fill_neighbours_ref_counts
     0.55%  rav1e    rav1e               [.] rav1e::encoder::motion_compensate
     0.54%  rav1e    rav1e               [.] rav1e::asm::x86::transform::inverse::inverse_transform_add
     0.51%  rav1e    libc-2.30.so        [.] __memset_avx2_unaligned_erms
     0.50%  rav1e    rav1e               [.] rav1e::predict::PredictionMode::predict_inter

created time in 2 days

startedtdaede/midiori

started time in 2 days

Pull request review commentxiph/rnnoise

assorted fixes and performance improvements

 void compute_gru(const GRULayer *gru, float *state, const float *input);  void compute_rnn(RNNState *rnn, float *gains, float *vad, const float *input); -#endif /* _MLP_H_ */+#endif /* _RNN_H_ */

I am not able to evaluate the rest of the changes, but applied this one to get it out of the way.

sleepybishop

comment created time in 2 days

push eventxiph/rnnoise

Petter Reinholdtsen

commit sha 1cbdbcf1283499bbb2230a6b0f126eb9b236defd

Correct #endif in rnn.h.

view details

push time in 2 days

push eventxiph/rnnoise

Luca Barbato

commit sha eca42a4b9035a9fd66095183cc195a9f488c3e12

Add rnnoise_get_frame_size() It provides a mean to know how many samples are processed at time.

view details

Luca Barbato

commit sha 085d8f484af6141b1b88281a4043fb9215cead01

Document the API

view details

push time in 2 days

PR merged xiph/rnnoise

Add rnnoise_get_frame_size()

It provides a mean to know how many samples are processed at time.

+46 -0

5 comments

2 changed files

lu-zero

pr closed time in 2 days

more