Fix argument to finer_search_pixel_proj_error
Patch https://aomedia-review.googlesource.com/c/aom/+/20200, merged as
32d150b, converted several functions in pickrst.c to take a
"use_highbitdepth" flag as well as (or instead of) the actual bit
depth.
Unfortunately, I missed a call site and the code can end up passing
the number 8 as the flag for use_highbitdepth (and, since 8 != 0, this
ends up using the high bit depth patch).
BUG=aomedia:714
Change-Id: Ie4dbad92f57ea1bacc4d99aad15454d9e5b6ff47
diff --git a/av1/encoder/pickrst.c b/av1/encoder/pickrst.c
index 0a922fc..5f7682b 100644
--- a/av1/encoder/pickrst.c
+++ b/av1/encoder/pickrst.c
@@ -394,8 +394,8 @@
aom_clear_system_state();
encode_xq(exq, exqd);
err = finer_search_pixel_proj_error(src8, width, height, src_stride, dat8,
- dat_stride, bit_depth, flt1, width,
- flt2, width, 2, exqd);
+ dat_stride, use_highbitdepth, flt1,
+ width, flt2, width, 2, exqd);
if (besterr == -1 || err < besterr) {
bestep = ep;
besterr = err;