X-Git-Url: https://git.nihav.org/?a=blobdiff_plain;f=nihav-realmedia%2Fsrc%2Fcodecs%2Frv40dsp.rs;h=9d7e44354c1953f1722053932ac2f76e86c5b81c;hb=bb0e22f77fd2c0f859ab342cf566a6151557745e;hp=030321c30acd10307a22d8514de807c445fbdbcb;hpb=5641dccfbf2a70d589cf094a0d4ed5a10f919f00;p=nihav.git

diff --git a/nihav-realmedia/src/codecs/rv40dsp.rs b/nihav-realmedia/src/codecs/rv40dsp.rs
index 030321c..9d7e443 100644
--- a/nihav-realmedia/src/codecs/rv40dsp.rs
+++ b/nihav-realmedia/src/codecs/rv40dsp.rs
@@ -1,6 +1,6 @@
 use nihav_core::frame::{FrameType, NAVideoBuffer};
-use nihav_core::codecs::MV;
-use nihav_core::codecs::blockdsp::edge_emu;
+use nihav_codec_support::codecs::MV;
+use nihav_codec_support::codecs::blockdsp::edge_emu;
 use super::rv3040::{RV34DSP, RV34MBInfo};
 
 fn clip8(a: i16) -> u8 {
@@ -38,6 +38,12 @@ macro_rules! filter {
                      -5 * el!($s, $o + 2 * $step)
                         + el!($s, $o + 3 * $step) + 32) >> 6) as i16)
         );
+    (33; $s: ident, $o: expr, $stride: expr) => (
+            clip8(((  el!($s, $o)
+                    + el!($s, $o + 1)
+                    + el!($s, $o + $stride)
+                    + el!($s, $o + 1 + $stride) + 2) >> 2) as i16)
+        );
 }
 
 macro_rules! mc_func {
@@ -130,6 +136,15 @@ macro_rules! mc_func {
             $ofilt(dst, didx, dstride, &buf, 2*bstride, $size);
         }
         );
+    (mc33; $name: ident, $size: expr) => (
+        fn $name (dst: &mut [u8], mut didx: usize, dstride: usize, src: &[u8], mut sidx: usize, sstride: usize) {
+            for _ in 0..$size {
+                for x in 0..$size { dst[didx + x] = filter!(33; src, sidx + x, sstride); }
+                sidx += sstride;
+                didx += dstride;
+            }
+        }
+        );
 }
 mc_func!(copy; copy_16, 16);
 mc_func!(copy; copy_8,   8);
@@ -161,8 +176,8 @@ mc_func!(cm03; luma_mc_31_16, 16, luma_mc_01_16);
 mc_func!(cm03; luma_mc_31_8,   8, luma_mc_01_8);
 mc_func!(cm03; luma_mc_32_16, 16, luma_mc_02_16);
 mc_func!(cm03; luma_mc_32_8,   8, luma_mc_02_8);
-mc_func!(cm03; luma_mc_33_16, 16, luma_mc_03_16);
-mc_func!(cm03; luma_mc_33_8,   8, luma_mc_03_8);
+mc_func!(mc33; luma_mc_33_16, 16);
+mc_func!(mc33; luma_mc_33_8,   8);
 
 const RV40_CHROMA_BIAS: [[u16; 4]; 4] = [
     [ 0, 4, 8, 4 ],
@@ -280,12 +295,12 @@ fn rv40_weak_loop_filter4(pix: &mut [u8], mut off: usize, step: usize, stride: u
         pix[off - step] = clip8(p0 + diff);
         pix[off       ] = clip8(q0 - diff);
 
-        if filter_p1 && ((p1 - p0).wrapping_abs() <= beta) {
+        if filter_p1 && ((p1 - p2).wrapping_abs() <= beta) {
             let p1_diff = ((p1 - p0) + (p1 - p2) - diff) >> 1;
             pix[off - 2*step] = clip8(p1 - clip_symm(p1_diff, lim_p1));
         }
 
-        if filter_q1 && ((q1 - q0).wrapping_abs() <= beta) {
+        if filter_q1 && ((q1 - q2).wrapping_abs() <= beta) {
             let q1_diff = ((q1 - q0) + (q1 - q2) + diff) >> 1;
             pix[off + step] = clip8(q1 - clip_symm(q1_diff, lim_q1));
         }
@@ -299,6 +314,7 @@ fn rv40_weak_loop_filter4_h(pix: &mut [u8], off: usize, stride: usize,
                             lim_p0q0: i16, lim_p1: i16, lim_q1: i16) {
     rv40_weak_loop_filter4(pix, off, stride, 1, filter_p1, filter_q1, alpha, beta, lim_p0q0, lim_p1, lim_q1);
 }
+#[allow(clippy::eq_op)]
 fn rv40_weak_loop_filter4_v(pix: &mut [u8], off: usize, stride: usize,
                             filter_p1: bool, filter_q1: bool, alpha: i16, beta: i16,
                             lim_p0q0: i16, lim_p1: i16, lim_q1: i16) {
@@ -334,12 +350,12 @@ fn rv40_weak_loop_filter4_v(pix: &mut [u8], off: usize, stride: usize,
         ch[3 - 1] = clip8(p0 + diff);
         ch[3    ] = clip8(q0 - diff);
 
-        if filter_p1 && ((p1 - p0).wrapping_abs() <= beta) {
+        if filter_p1 && ((p1 - p2).wrapping_abs() <= beta) {
             let p1_diff = ((p1 - p0) + (p1 - p2) - diff) >> 1;
             ch[3 - 2] = clip8(p1 - clip_symm(p1_diff, lim_p1));
         }
 
-        if filter_q1 && ((q1 - q0).wrapping_abs() <= beta) {
+        if filter_q1 && ((q1 - q2).wrapping_abs() <= beta) {
             let q1_diff = ((q1 - q0) + (q1 - q2) + diff) >> 1;
             ch[3 + 1] = clip8(q1 - clip_symm(q1_diff, lim_q1));
         }
@@ -364,7 +380,7 @@ fn sfilter(a: i16, b: i16, c: i16, d: i16, e: i16, dither: i16, clip: bool, lims
         } else if val > c + lims {
             c + lims
         } else {
-            c
+            val
         }
     } else {
         val
@@ -397,7 +413,7 @@ fn rv40_strong_loop_filter4(pix: &mut [u8], mut off: usize, step: usize, stride:
         let q3 = el!(pix, off + 3*step);
 
         let np0 = sfilter(p2, p1, p0, q0, q1,     RV40_DITHER_L[dmode + i], fmode != 0, lims);
-        let nq0 = sfilter(    p1, p0, q0, q1, q0, RV40_DITHER_R[dmode + i], fmode != 0, lims);
+        let nq0 = sfilter(    p1, p0, q0, q1, q2, RV40_DITHER_R[dmode + i], fmode != 0, lims);
 
         let np1 = sfilter(p3, p2, p1, np0, q0,              RV40_DITHER_L[dmode + i], fmode != 0, lims);
         let nq1 = sfilter(             p0, nq0, q1, q2, q3, RV40_DITHER_R[dmode + i], fmode != 0, lims);
@@ -408,8 +424,8 @@ fn rv40_strong_loop_filter4(pix: &mut [u8], mut off: usize, step: usize, stride:
         pix[off +   step] = nq1 as u8;
 
         if !chroma {
-            let np2 = sfilter(np0, np1, p2, p3, np1, 64, false, 0);
-            let nq2 = sfilter(nq0, nq1, q2, q3, q2,  64, false, 0);
+            let np2 = sfilter(np0, np1, p2, p3, p2, 64, false, 0);
+            let nq2 = sfilter(nq0, nq1, q2, q3, q2, 64, false, 0);
             pix[off - 3*step] = np2 as u8;
             pix[off + 2*step] = nq2 as u8;
         }
@@ -457,6 +473,7 @@ fn rv40_loop_strength_h(pix: &[u8], off: usize, stride: usize,
     rv40_loop_strength(pix, off, stride, 1, beta, beta2, edge)
 }
 
+#[allow(clippy::eq_op)]
 fn rv40_loop_strength_v(pix: &[u8], off: usize, stride: usize,
                         beta: i16, beta2: i16, edge: bool) -> (bool, bool, bool) {
     let src = &pix[off - 3..][..stride * 3 + 3 + 3];
@@ -564,7 +581,7 @@ const RV40_EDGE1: [isize; 4] = [ 0, 2, 2, 2 ];
 const RV40_EDGE2: [isize; 4] = [ 0, 3, 3, 3 ];
 
 impl RV34DSP for RV40DSP {
-    fn loop_filter(&self, frame: &mut NAVideoBuffer<u8>, _ftype: FrameType, mbinfo: &[RV34MBInfo], mb_w: usize, row: usize) {
+    fn loop_filter(&self, frame: &mut NAVideoBuffer<u8>, _ftype: FrameType, mbinfo: &[RV34MBInfo], mb_w: usize, mb_h: usize, row: usize) {
         // todo proper B-frame filtering?
         let mut offs:   [usize; 3] = [0; 3];
         let mut stride: [usize; 3] = [0; 3];
@@ -577,7 +594,7 @@ impl RV34DSP for RV40DSP {
             offs[comp] = frame.get_offset(comp) + start * stride[comp];
         }
 
-        let mut data = frame.get_data_mut();
+        let data = frame.get_data_mut().unwrap();
         let dst: &mut [u8] = data.as_mut_slice();
 
         let mut mb_pos: usize = row * mb_w;
@@ -810,7 +827,7 @@ impl RV34DSP for RV40DSP {
         let size: usize = if use16 { 16 } else { 8 };
         let dstride = frame.get_stride(0);
         let doffset = frame.get_offset(0) + (if !avg { x + y * dstride } else { 0 });
-        let mut data = frame.get_data_mut();
+        let data = frame.get_data_mut().unwrap();
         let dst: &mut [u8] = data.as_mut_slice();
 
         let (w_, h_) = prev_frame.get_dimensions(0);
@@ -840,7 +857,7 @@ impl RV34DSP for RV40DSP {
         let size: usize = if use8 { 8 } else { 4 };
         let dstride = frame.get_stride(comp);
         let doffset = frame.get_offset(comp) + (if !avg { x + y * dstride } else { 0 });
-        let mut data = frame.get_data_mut();
+        let data = frame.get_data_mut().unwrap();
         let dst: &mut [u8] = data.as_mut_slice();
 
         let (w_, h_) = prev_frame.get_dimensions(comp);