Merge "Enable 4x4 DCT/ADST transform unit test"

This commit is contained in:
Jingning Han 2013-11-14 09:26:33 -08:00 committed by Gerrit Code Review
commit 4995dbc713

View File

@ -13,178 +13,288 @@
#include <string.h>
#include "third_party/googletest/src/include/gtest/gtest.h"
#include "test/acm_random.h"
#include "test/clear_system_state.h"
#include "test/register_state_check.h"
#include "test/util.h"
extern "C" {
#include "vp9/common/vp9_entropy.h"
#include "./vp9_rtcd.h"
void vp9_idct4x4_16_add_c(const int16_t *input, uint8_t *output, int pitch);
}
#include "test/acm_random.h"
#include "vpx/vpx_integer.h"
#include "vpx_ports/mem.h"
using libvpx_test::ACMRandom;
namespace {
void fdct4x4(int16_t *in, int16_t *out, uint8_t* /*dst*/,
int stride, int /*tx_type*/) {
const int kNumCoeffs = 16;
typedef void (*fdct_t)(const int16_t *in, int16_t *out, int stride);
typedef void (*idct_t)(const int16_t *in, uint8_t *out, int stride);
typedef void (*fht_t) (const int16_t *in, int16_t *out, int stride,
int tx_type);
typedef void (*iht_t) (const int16_t *in, uint8_t *out, int stride,
int tx_type);
void fdct4x4_ref(const int16_t *in, int16_t *out, int stride, int tx_type) {
vp9_fdct4x4_c(in, out, stride);
}
void idct4x4_add(int16_t* /*in*/, int16_t *out, uint8_t *dst,
int stride, int /*tx_type*/) {
vp9_idct4x4_16_add_c(out, dst, stride);
}
void fht4x4(int16_t *in, int16_t *out, uint8_t* /*dst*/,
int stride, int tx_type) {
void fht4x4_ref(const int16_t *in, int16_t *out, int stride, int tx_type) {
vp9_short_fht4x4_c(in, out, stride, tx_type);
}
void iht4x4_add(int16_t* /*in*/, int16_t *out, uint8_t *dst,
int stride, int tx_type) {
vp9_iht4x4_16_add_c(out, dst, stride, tx_type);
}
class FwdTrans4x4Test : public ::testing::TestWithParam<int> {
class Trans4x4TestBase {
public:
virtual ~FwdTrans4x4Test() {}
virtual void SetUp() {
tx_type_ = GetParam();
if (tx_type_ == 0) {
fwd_txfm_ = fdct4x4;
inv_txfm_ = idct4x4_add;
} else {
fwd_txfm_ = fht4x4;
inv_txfm_ = iht4x4_add;
}
}
virtual ~Trans4x4TestBase() {}
protected:
void RunFwdTxfm(int16_t *in, int16_t *out, uint8_t *dst,
int stride, int tx_type) {
(*fwd_txfm_)(in, out, dst, stride, tx_type);
virtual void RunFwdTxfm(const int16_t *in, int16_t *out, int stride) = 0;
virtual void RunInvTxfm(const int16_t *out, uint8_t *dst, int stride) = 0;
void RunAccuracyCheck() {
ACMRandom rnd(ACMRandom::DeterministicSeed());
uint32_t max_error = 0;
int64_t total_error = 0;
const int count_test_block = 10000;
for (int i = 0; i < count_test_block; ++i) {
DECLARE_ALIGNED_ARRAY(16, int16_t, test_input_block, kNumCoeffs);
DECLARE_ALIGNED_ARRAY(16, int16_t, test_temp_block, kNumCoeffs);
DECLARE_ALIGNED_ARRAY(16, uint8_t, dst, kNumCoeffs);
DECLARE_ALIGNED_ARRAY(16, uint8_t, src, kNumCoeffs);
// Initialize a test block with input range [-255, 255].
for (int j = 0; j < kNumCoeffs; ++j) {
src[j] = rnd.Rand8();
dst[j] = rnd.Rand8();
test_input_block[j] = src[j] - dst[j];
}
REGISTER_STATE_CHECK(RunFwdTxfm(test_input_block,
test_temp_block, pitch_));
REGISTER_STATE_CHECK(RunInvTxfm(test_temp_block, dst, pitch_));
for (int j = 0; j < kNumCoeffs; ++j) {
const uint32_t diff = dst[j] - src[j];
const uint32_t error = diff * diff;
if (max_error < error)
max_error = error;
total_error += error;
}
}
EXPECT_GE(1u, max_error)
<< "Error: 4x4 FHT/IHT has an individual round trip error > 1";
EXPECT_GE(count_test_block , total_error)
<< "Error: 4x4 FHT/IHT has average round trip error > 1 per block";
}
void RunInvTxfm(int16_t *in, int16_t *out, uint8_t *dst,
int stride, int tx_type) {
(*inv_txfm_)(in, out, dst, stride, tx_type);
void RunCoeffCheck() {
ACMRandom rnd(ACMRandom::DeterministicSeed());
const int count_test_block = 5000;
DECLARE_ALIGNED_ARRAY(16, int16_t, input_block, kNumCoeffs);
DECLARE_ALIGNED_ARRAY(16, int16_t, output_ref_block, kNumCoeffs);
DECLARE_ALIGNED_ARRAY(16, int16_t, output_block, kNumCoeffs);
for (int i = 0; i < count_test_block; ++i) {
// Initialize a test block with input range [-255, 255].
for (int j = 0; j < kNumCoeffs; ++j)
input_block[j] = rnd.Rand8() - rnd.Rand8();
fwd_txfm_ref(input_block, output_ref_block, pitch_, tx_type_);
REGISTER_STATE_CHECK(RunFwdTxfm(input_block, output_block, pitch_));
// The minimum quant value is 4.
for (int j = 0; j < kNumCoeffs; ++j)
EXPECT_EQ(output_block[j], output_ref_block[j]);
}
}
void RunMemCheck() {
ACMRandom rnd(ACMRandom::DeterministicSeed());
const int count_test_block = 5000;
DECLARE_ALIGNED_ARRAY(16, int16_t, input_block, kNumCoeffs);
DECLARE_ALIGNED_ARRAY(16, int16_t, input_extreme_block, kNumCoeffs);
DECLARE_ALIGNED_ARRAY(16, int16_t, output_ref_block, kNumCoeffs);
DECLARE_ALIGNED_ARRAY(16, int16_t, output_block, kNumCoeffs);
for (int i = 0; i < count_test_block; ++i) {
// Initialize a test block with input range [-255, 255].
for (int j = 0; j < kNumCoeffs; ++j) {
input_block[j] = rnd.Rand8() - rnd.Rand8();
input_extreme_block[j] = rnd.Rand8() % 2 ? 255 : -255;
}
if (i == 0)
for (int j = 0; j < kNumCoeffs; ++j)
input_extreme_block[j] = 255;
if (i == 1)
for (int j = 0; j < kNumCoeffs; ++j)
input_extreme_block[j] = -255;
fwd_txfm_ref(input_extreme_block, output_ref_block, pitch_, tx_type_);
REGISTER_STATE_CHECK(RunFwdTxfm(input_extreme_block,
output_block, pitch_));
// The minimum quant value is 4.
for (int j = 0; j < kNumCoeffs; ++j) {
EXPECT_EQ(output_block[j], output_ref_block[j]);
EXPECT_GE(4 * DCT_MAX_VALUE, abs(output_block[j]))
<< "Error: 16x16 FDCT has coefficient larger than 4*DCT_MAX_VALUE";
}
}
}
void RunInvAccuracyCheck() {
ACMRandom rnd(ACMRandom::DeterministicSeed());
const int count_test_block = 1000;
DECLARE_ALIGNED_ARRAY(16, int16_t, in, kNumCoeffs);
DECLARE_ALIGNED_ARRAY(16, int16_t, coeff, kNumCoeffs);
DECLARE_ALIGNED_ARRAY(16, uint8_t, dst, kNumCoeffs);
DECLARE_ALIGNED_ARRAY(16, uint8_t, src, kNumCoeffs);
for (int i = 0; i < count_test_block; ++i) {
// Initialize a test block with input range [-255, 255].
for (int j = 0; j < kNumCoeffs; ++j) {
src[j] = rnd.Rand8();
dst[j] = rnd.Rand8();
in[j] = src[j] - dst[j];
}
fwd_txfm_ref(in, coeff, pitch_, tx_type_);
REGISTER_STATE_CHECK(RunInvTxfm(coeff, dst, pitch_));
for (int j = 0; j < kNumCoeffs; ++j) {
const uint32_t diff = dst[j] - src[j];
const uint32_t error = diff * diff;
EXPECT_GE(1u, error)
<< "Error: 16x16 IDCT has error " << error
<< " at index " << j;
}
}
}
int pitch_;
int tx_type_;
void (*fwd_txfm_)(int16_t *in, int16_t *out, uint8_t *dst,
int stride, int tx_type);
void (*inv_txfm_)(int16_t *in, int16_t *out, uint8_t *dst,
int stride, int tx_type);
fht_t fwd_txfm_ref;
};
TEST_P(FwdTrans4x4Test, SignBiasCheck) {
ACMRandom rnd(ACMRandom::DeterministicSeed());
DECLARE_ALIGNED_ARRAY(16, int16_t, test_input_block, 16);
DECLARE_ALIGNED_ARRAY(16, int16_t, test_output_block, 16);
const int pitch = 4;
int count_sign_block[16][2];
const int count_test_block = 1000000;
class Trans4x4DCT
: public Trans4x4TestBase,
public PARAMS(fdct_t, idct_t, int) {
public:
virtual ~Trans4x4DCT() {}
memset(count_sign_block, 0, sizeof(count_sign_block));
for (int i = 0; i < count_test_block; ++i) {
// Initialize a test block with input range [-255, 255].
for (int j = 0; j < 16; ++j)
test_input_block[j] = rnd.Rand8() - rnd.Rand8();
virtual void SetUp() {
fwd_txfm_ = GET_PARAM(0);
inv_txfm_ = GET_PARAM(1);
tx_type_ = GET_PARAM(2);
pitch_ = 4;
fwd_txfm_ref = fdct4x4_ref;
}
virtual void TearDown() { libvpx_test::ClearSystemState(); }
RunFwdTxfm(test_input_block, test_output_block, NULL, pitch, tx_type_);
for (int j = 0; j < 16; ++j) {
if (test_output_block[j] < 0)
++count_sign_block[j][0];
else if (test_output_block[j] > 0)
++count_sign_block[j][1];
}
protected:
void RunFwdTxfm(const int16_t *in, int16_t *out, int stride) {
fwd_txfm_(in, out, stride);
}
void RunInvTxfm(const int16_t *out, uint8_t *dst, int stride) {
inv_txfm_(out, dst, stride);
}
for (int j = 0; j < 16; ++j) {
const bool bias_acceptable = (abs(count_sign_block[j][0] -
count_sign_block[j][1]) < 10000);
EXPECT_TRUE(bias_acceptable)
<< "Error: 4x4 FDCT/FHT has a sign bias > 1%"
<< " for input range [-255, 255] at index " << j
<< " tx_type " << tx_type_;
}
fdct_t fwd_txfm_;
idct_t inv_txfm_;
};
memset(count_sign_block, 0, sizeof(count_sign_block));
for (int i = 0; i < count_test_block; ++i) {
// Initialize a test block with input range [-15, 15].
for (int j = 0; j < 16; ++j)
test_input_block[j] = (rnd.Rand8() >> 4) - (rnd.Rand8() >> 4);
RunFwdTxfm(test_input_block, test_output_block, NULL, pitch, tx_type_);
for (int j = 0; j < 16; ++j) {
if (test_output_block[j] < 0)
++count_sign_block[j][0];
else if (test_output_block[j] > 0)
++count_sign_block[j][1];
}
}
for (int j = 0; j < 16; ++j) {
const bool bias_acceptable = (abs(count_sign_block[j][0] -
count_sign_block[j][1]) < 100000);
EXPECT_TRUE(bias_acceptable)
<< "Error: 4x4 FDCT/FHT has a sign bias > 10%"
<< " for input range [-15, 15] at index " << j;
}
TEST_P(Trans4x4DCT, AccuracyCheck) {
RunAccuracyCheck();
}
TEST_P(FwdTrans4x4Test, RoundTripErrorCheck) {
ACMRandom rnd(ACMRandom::DeterministicSeed());
int max_error = 0;
int total_error = 0;
const int count_test_block = 1000000;
for (int i = 0; i < count_test_block; ++i) {
DECLARE_ALIGNED_ARRAY(16, int16_t, test_input_block, 16);
DECLARE_ALIGNED_ARRAY(16, int16_t, test_temp_block, 16);
DECLARE_ALIGNED_ARRAY(16, uint8_t, dst, 16);
DECLARE_ALIGNED_ARRAY(16, uint8_t, src, 16);
for (int j = 0; j < 16; ++j) {
src[j] = rnd.Rand8();
dst[j] = rnd.Rand8();
}
// Initialize a test block with input range [-255, 255].
for (int j = 0; j < 16; ++j)
test_input_block[j] = src[j] - dst[j];
const int pitch = 4;
RunFwdTxfm(test_input_block, test_temp_block, dst, pitch, tx_type_);
for (int j = 0; j < 16; ++j) {
if (test_temp_block[j] > 0) {
test_temp_block[j] += 2;
test_temp_block[j] /= 4;
test_temp_block[j] *= 4;
} else {
test_temp_block[j] -= 2;
test_temp_block[j] /= 4;
test_temp_block[j] *= 4;
}
}
// inverse transform and reconstruct the pixel block
RunInvTxfm(test_input_block, test_temp_block, dst, pitch, tx_type_);
for (int j = 0; j < 16; ++j) {
const int diff = dst[j] - src[j];
const int error = diff * diff;
if (max_error < error)
max_error = error;
total_error += error;
}
}
EXPECT_GE(1, max_error)
<< "Error: FDCT/IDCT or FHT/IHT has an individual roundtrip error > 1";
EXPECT_GE(count_test_block, total_error)
<< "Error: FDCT/IDCT or FHT/IHT has average "
<< "roundtrip error > 1 per block";
TEST_P(Trans4x4DCT, CoeffCheck) {
RunCoeffCheck();
}
INSTANTIATE_TEST_CASE_P(VP9, FwdTrans4x4Test, ::testing::Range(0, 4));
TEST_P(Trans4x4DCT, MemCheck) {
RunMemCheck();
}
TEST_P(Trans4x4DCT, InvAccuracyCheck) {
RunInvAccuracyCheck();
}
class Trans4x4HT
: public Trans4x4TestBase,
public PARAMS(fht_t, iht_t, int) {
public:
virtual ~Trans4x4HT() {}
virtual void SetUp() {
fwd_txfm_ = GET_PARAM(0);
inv_txfm_ = GET_PARAM(1);
tx_type_ = GET_PARAM(2);
pitch_ = 4;
fwd_txfm_ref = fht4x4_ref;
}
virtual void TearDown() { libvpx_test::ClearSystemState(); }
protected:
void RunFwdTxfm(const int16_t *in, int16_t *out, int stride) {
fwd_txfm_(in, out, stride, tx_type_);
}
void RunInvTxfm(const int16_t *out, uint8_t *dst, int stride) {
inv_txfm_(out, dst, stride, tx_type_);
}
fht_t fwd_txfm_;
iht_t inv_txfm_;
};
TEST_P(Trans4x4HT, AccuracyCheck) {
RunAccuracyCheck();
}
TEST_P(Trans4x4HT, CoeffCheck) {
RunCoeffCheck();
}
TEST_P(Trans4x4HT, MemCheck) {
RunMemCheck();
}
TEST_P(Trans4x4HT, InvAccuracyCheck) {
RunInvAccuracyCheck();
}
using std::tr1::make_tuple;
INSTANTIATE_TEST_CASE_P(
C, Trans4x4DCT,
::testing::Values(
make_tuple(&vp9_fdct4x4_c, &vp9_idct4x4_16_add_c, 0)));
INSTANTIATE_TEST_CASE_P(
C, Trans4x4HT,
::testing::Values(
make_tuple(&vp9_short_fht4x4_c, &vp9_iht4x4_16_add_c, 0),
make_tuple(&vp9_short_fht4x4_c, &vp9_iht4x4_16_add_c, 1),
make_tuple(&vp9_short_fht4x4_c, &vp9_iht4x4_16_add_c, 2),
make_tuple(&vp9_short_fht4x4_c, &vp9_iht4x4_16_add_c, 3)));
#if HAVE_SSE2
INSTANTIATE_TEST_CASE_P(
SSE2, Trans4x4DCT,
::testing::Values(
make_tuple(&vp9_fdct4x4_sse2,
&vp9_idct4x4_16_add_sse2, 0)));
INSTANTIATE_TEST_CASE_P(
SSE2, Trans4x4HT,
::testing::Values(
make_tuple(&vp9_short_fht4x4_sse2, &vp9_iht4x4_16_add_sse2, 0),
make_tuple(&vp9_short_fht4x4_sse2, &vp9_iht4x4_16_add_sse2, 1),
make_tuple(&vp9_short_fht4x4_sse2, &vp9_iht4x4_16_add_sse2, 2),
make_tuple(&vp9_short_fht4x4_sse2, &vp9_iht4x4_16_add_sse2, 3)));
#endif
} // namespace