diff options
Diffstat (limited to 'video/out/gpu/video.c')
| -rw-r--r-- | video/out/gpu/video.c | 51 |
1 files changed, 34 insertions, 17 deletions
diff --git a/video/out/gpu/video.c b/video/out/gpu/video.c index 46d9026742..13e5b06918 100644 --- a/video/out/gpu/video.c +++ b/video/out/gpu/video.c @@ -373,8 +373,9 @@ const struct m_sub_options gl_video_conf = { SCALER_OPTS("tscale", SCALER_TSCALE), OPT_INTRANGE("scaler-lut-size", scaler_lut_size, 0, 4, 10), OPT_FLAG("scaler-resizes-only", scaler_resizes_only, 0), - OPT_FLAG("linear-scaling", linear_scaling, 0), OPT_FLAG("correct-downscaling", correct_downscaling, 0), + OPT_FLAG("linear-downscaling", linear_downscaling, 0), + OPT_FLAG("linear-upscaling", linear_upscaling, 0), OPT_FLAG("sigmoid-upscaling", sigmoid_upscaling, 0), OPT_FLOATRANGE("sigmoid-center", sigmoid_center, 0, 0.0, 1.0), OPT_FLOATRANGE("sigmoid-slope", sigmoid_slope, 0, 1.0, 20.0), @@ -423,6 +424,8 @@ const struct m_sub_options gl_video_conf = { OPT_REPLACED("opengl-fbo-format", "fbo-format"), OPT_REPLACED("opengl-dumb-mode", "gpu-dumb-mode"), OPT_REPLACED("opengl-gamma", "gamma-factor"), + OPT_REMOVED("linear-scaling", "Split into --linear-upscaling and " + "--linear-downscaling"), {0} }, .size = sizeof(struct gl_video_opts), @@ -1103,8 +1106,14 @@ static void cleanup_binds(struct gl_video *p) // Sets the appropriate compute shader metadata for an implicit compute pass // bw/bh: block size -static void pass_is_compute(struct gl_video *p, int bw, int bh) +static void pass_is_compute(struct gl_video *p, int bw, int bh, bool flexible) { + if (p->pass_compute.active && flexible) { + // Avoid overwriting existing block sizes when using a flexible pass + bw = p->pass_compute.block_w; + bh = p->pass_compute.block_h; + } + p->pass_compute = (struct compute_info){ .active = true, .block_w = bw, @@ -1248,7 +1257,7 @@ static void finish_pass_tex(struct gl_video *p, struct ra_tex **dst_tex, // If RA_CAP_PARALLEL_COMPUTE is set, try to prefer compute shaders // over fragment shaders wherever possible. if (!p->pass_compute.active && (p->ra->caps & RA_CAP_PARALLEL_COMPUTE)) - pass_is_compute(p, 16, 16); + pass_is_compute(p, 16, 16, true); if (p->pass_compute.active) { gl_sc_uniform_image2D_wo(p->sc, "out_image", *dst_tex); @@ -1744,7 +1753,7 @@ static void pass_dispatch_sample_polar(struct gl_video *p, struct scaler *scaler if (shmem_req > p->ra->max_shmem) goto fallback; - pass_is_compute(p, bw, bh); + pass_is_compute(p, bw, bh, false); pass_compute_polar(p->sc, scaler, img.components, bw, bh, iw, ih); return; @@ -2326,13 +2335,18 @@ static void pass_scale_main(struct gl_video *p) // Pre-conversion, like linear light/sigmoidization GLSLF("// scaler pre-conversion\n"); - bool use_linear = p->opts.linear_scaling || p->opts.sigmoid_upscaling; + bool use_linear = false; + if (downscaling) { + use_linear = p->opts.linear_downscaling; - // Linear light downscaling results in nasty artifacts for HDR curves due - // to the potentially extreme brightness differences severely compounding - // any ringing. So just scale in gamma light instead. - if (mp_trc_is_hdr(p->image_params.color.gamma) && downscaling) - use_linear = false; + // Linear light downscaling results in nasty artifacts for HDR curves + // due to the potentially extreme brightness differences severely + // compounding any ringing. So just scale in gamma light instead. + if (mp_trc_is_hdr(p->image_params.color.gamma)) + use_linear = false; + } else if (upscaling) { + use_linear = p->opts.linear_upscaling || p->opts.sigmoid_upscaling; + } if (use_linear) { p->use_linear = true; @@ -2485,7 +2499,7 @@ static void pass_colormanage(struct gl_video *p, struct mp_colorspace src, bool if (detect_peak) { pass_describe(p, "detect HDR peak"); - pass_is_compute(p, 8, 8); // 8x8 is good for performance + pass_is_compute(p, 8, 8, true); // 8x8 is good for performance gl_sc_ssbo(p->sc, "PeakDetect", p->hdr_peak_ssbo, "uint counter;" "uint frame_idx;" @@ -3488,9 +3502,9 @@ static bool check_dumb_mode(struct gl_video *p) return false; // otherwise, use auto-detection - if (o->target_prim || o->target_trc || o->linear_scaling || - o->correct_downscaling || o->sigmoid_upscaling || o->interpolation || - o->blend_subs || o->deband || o->unsharp) + if (o->target_prim || o->target_trc || o->correct_downscaling || + o->linear_downscaling || o->linear_upscaling || o->sigmoid_upscaling || + o->interpolation || o->blend_subs || o->deband || o->unsharp) return false; // check remaining scalers (tscale is already implicitly excluded above) for (int i = 0; i < SCALER_COUNT; i++) { @@ -3519,7 +3533,7 @@ static void check_gl_features(struct gl_video *p) bool have_ssbo = ra->caps & RA_CAP_BUF_RW; bool have_fragcoord = ra->caps & RA_CAP_FRAGCOORD; - const char *auto_fbo_fmts[] = {"rgba16", "rgba16f", "rgba16hf", + const char *auto_fbo_fmts[] = {"rgba16f", "rgba16hf", "rgba16", "rgb10_a2", "rgba8", 0}; const char *user_fbo_fmts[] = {p->opts.fbo_format, 0}; const char **fbo_fmts = user_fbo_fmts[0] && strcmp(user_fbo_fmts[0], "auto") @@ -3646,8 +3660,11 @@ static void check_gl_features(struct gl_video *p) p->opts.target_trc != MP_CSP_TRC_AUTO || p->use_lut_3d; // mix() is needed for some gamma functions - if (!have_mglsl && (p->opts.linear_scaling || p->opts.sigmoid_upscaling)) { - p->opts.linear_scaling = false; + if (!have_mglsl && (p->opts.linear_downscaling || + p->opts.linear_upscaling || p->opts.sigmoid_upscaling)) + { + p->opts.linear_downscaling = false; + p->opts.linear_upscaling = false; p->opts.sigmoid_upscaling = false; MP_WARN(p, "Disabling linear/sigmoid scaling (GLSL version too old).\n"); } |