Bitstream change - avoid overflow of identity txfm
1) Use 12-bit precision for square-root-2 scaling in identity txfms
2) Use generic method to compute the identity txfm
3) Test the identity config for tx64
BUG=aomedia:1327
Change-Id: I7350682bae9134b1f70a9ad80d5e64236f7728e2
diff --git a/test/av1_inv_txfm1d_test.cc b/test/av1_inv_txfm1d_test.cc
index b5a7d19..cf0bab5 100644
--- a/test/av1_inv_txfm1d_test.cc
+++ b/test/av1_inv_txfm1d_test.cc
@@ -50,7 +50,7 @@
// the maximum stage number of fwd/inv 1d dct/adst txfm is 12
const int8_t cos_bit = 13;
-const int8_t range_bit[12] = { 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32, 32 };
+const int8_t range_bit[12] = { 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20 };
void reference_idct_1d_int(const int32_t *in, int32_t *out, int size) {
double input[64];
