diff --git a/src/encoder_state-bitstream.c b/src/encoder_state-bitstream.c
index 705460a1..0bcc0f9b 100644
--- a/src/encoder_state-bitstream.c
+++ b/src/encoder_state-bitstream.c
@@ -1298,6 +1298,12 @@ void kvz_encoder_state_write_bitstream_slice_header(
     }
   }
 
+  // ToDo: depquant
+
+  if (state->encoder_control->cfg.signhide_enable) {
+    WRITE_U(stream, 1, 1, "sh_sign_data_hiding_used_flag");
+  }
+
   if (state->frame->slicetype != KVZ_SLICE_I) {
 
     // BT Size set only with non-I-frames, in I-frames the size is 32x32
diff --git a/src/rdo.c b/src/rdo.c
index 83e8d742..2b2d33f7 100644
--- a/src/rdo.c
+++ b/src/rdo.c
@@ -638,7 +638,7 @@ void kvz_rdoq_sign_hiding(
  * coding engines using probability models like CABAC
  * From HM 12.0
  */
-/*
+
 // ToDo: implement new RDOQ
 void kvz_rdoq(encoder_state_t * const state, coeff_t *coef, coeff_t *dest_coeff, int32_t width,
            int32_t height, int8_t type, int8_t scan_mode, int8_t block_type, int8_t tr_depth)
@@ -699,7 +699,7 @@ void kvz_rdoq(encoder_state_t * const state, coeff_t *coef, coeff_t *dest_coeff,
     default: assert(0 && "There should be 1, 4, 16 or 64 coefficient groups");
   }
 
-  cabac_ctx_t *base_coeff_group_ctx = &(cabac->ctx.cu_sig_coeff_group_model[type]);
+  cabac_ctx_t *base_coeff_group_ctx = &(cabac->ctx.sig_coeff_group_model[type]);
   cabac_ctx_t *baseCtx              = (type == 0) ? &(cabac->ctx.cu_sig_model_luma[0][0]) : &(cabac->ctx.cu_sig_model_chroma[0][0]);
 
   struct {
@@ -974,7 +974,7 @@ void kvz_rdoq(encoder_state_t * const state, coeff_t *coef, coeff_t *dest_coeff,
     kvz_rdoq_sign_hiding(state, qp_scaled, scan, &sh_rates, best_last_idx_p1, coef, dest_coeff);
   }
 }
-*/
+
 
 /**
  * Calculate cost of actual motion vectors using CABAC coding
diff --git a/src/rdo.h b/src/rdo.h
index d867cc4a..29848e65 100644
--- a/src/rdo.h
+++ b/src/rdo.h
@@ -36,8 +36,8 @@
 extern const uint32_t kvz_g_go_rice_range[5];
 extern const uint32_t kvz_g_go_rice_prefix_len[5];
 
-/*void  kvz_rdoq(encoder_state_t *state, coeff_t *coef, coeff_t *dest_coeff, int32_t width,
-           int32_t height, int8_t type, int8_t scan_mode, int8_t block_type, int8_t tr_depth);*/
+void  kvz_rdoq(encoder_state_t *state, coeff_t *coef, coeff_t *dest_coeff, int32_t width,
+           int32_t height, int8_t type, int8_t scan_mode, int8_t block_type, int8_t tr_depth);
 
 uint32_t kvz_get_coeff_cost(const encoder_state_t * const state,
                             const coeff_t *coeff,
diff --git a/src/strategies/avx2/quant-avx2.c b/src/strategies/avx2/quant-avx2.c
index b50f6505..3909947c 100644
--- a/src/strategies/avx2/quant-avx2.c
+++ b/src/strategies/avx2/quant-avx2.c
@@ -349,8 +349,8 @@ void kvz_quant_avx2(const encoder_state_t * const state, const coeff_t * __restr
   int32_t height, int8_t type, int8_t scan_idx, int8_t block_type)
 {
   const encoder_control_t * const encoder = state->encoder_control;
-  //const uint32_t log2_block_size = kvz_g_convert_to_bit[width] + 2;
-  //const uint32_t * const scan = kvz_g_sig_last_scan[scan_idx][log2_block_size - 1];
+  const uint32_t log2_block_size = kvz_g_convert_to_bit[width] + 2;
+  const uint32_t * const scan = kvz_g_sig_last_scan[scan_idx][log2_block_size - 1];
 
   int32_t qp_scaled = kvz_get_scaled_qp(type, state->qp, (encoder->bitdepth - 8) * 6);
   const uint32_t log2_tr_size = kvz_g_convert_to_bit[width] + 2;
@@ -359,7 +359,7 @@ void kvz_quant_avx2(const encoder_state_t * const state, const coeff_t * __restr
   const int32_t transform_shift = MAX_TR_DYNAMIC_RANGE - encoder->bitdepth - log2_tr_size; //!< Represents scaling through forward transform
   const int32_t q_bits = QUANT_SHIFT + qp_scaled / 6 + transform_shift;
   const int32_t add = ((state->frame->slicetype == KVZ_SLICE_I) ? 171 : 85) << (q_bits - 9);
-  //const int32_t q_bits8 = q_bits - 8;
+  const int32_t q_bits8 = q_bits - 8;
 
   uint32_t ac_sum = 0;
   int32_t last_cg = -1;
@@ -433,7 +433,7 @@ void kvz_quant_avx2(const encoder_state_t * const state, const coeff_t * __restr
   temp = _mm_add_epi32(temp, _mm_shuffle_epi32(temp, _MM_SHUFFLE(1, 0, 3, 2)));
   temp = _mm_add_epi32(temp, _mm_shuffle_epi32(temp, _MM_SHUFFLE(0, 1, 0, 1)));
   ac_sum += _mm_cvtsi128_si32(temp);
-  /*
+  
   // Signhiding disabled in VVC
   if (!encoder->cfg.signhide_enable || ac_sum < 2)
     return;
@@ -505,7 +505,7 @@ void kvz_quant_avx2(const encoder_state_t * const state, const coeff_t * __restr
 #undef VEC_WIDTH
 #undef SCAN_SET_SIZE
 #undef LOG2_SCAN_SET_SIZE
-  */
+  
 }
 
 #if KVZ_BIT_DEPTH == 8
@@ -658,7 +658,7 @@ int kvz_quantize_residual_avx2(encoder_state_t *const state,
   }
 
   // Quantize coeffs. (coeff -> coeff_out)
-  /*
+  
   if (state->encoder_control->cfg.rdoq_enable &&
       (width > 4 || !state->encoder_control->cfg.rdoq_skip))
   {
@@ -666,7 +666,7 @@ int kvz_quantize_residual_avx2(encoder_state_t *const state,
     tr_depth += (cur_cu->part_size == SIZE_NxN ? 1 : 0);
     kvz_rdoq(state, coeff, coeff_out, width, width, (color == COLOR_Y ? 0 : 2),
       scan_order, cur_cu->type, tr_depth);
-  } else*/ {
+  } else {
     kvz_quant(state, coeff, coeff_out, width, width, (color == COLOR_Y ? 0 : 2),
       scan_order, cur_cu->type);
   }
diff --git a/src/strategies/generic/encode_coding_tree-generic.c b/src/strategies/generic/encode_coding_tree-generic.c
index 8b6f4bba..1888f0dd 100644
--- a/src/strategies/generic/encode_coding_tree-generic.c
+++ b/src/strategies/generic/encode_coding_tree-generic.c
@@ -273,6 +273,7 @@ void kvz_encode_coeff_nxn_generic(encoder_state_t * const state,
         quant_state = (quant_state_transition_table >> ((quant_state << 2) + ((coeff_abs & 1) << 1))) & 3;
         if (coeff_abs) {
           num_non_zero++;
+          first_nz_pos_in_cg = scan_pos;
           last_nz_pos_in_cg = MAX(last_nz_pos_in_cg, scan_pos);
           coeff_signs <<= 1;
           if (coeff[blk_pos] < 0) coeff_signs++;
diff --git a/src/strategies/generic/quant-generic.c b/src/strategies/generic/quant-generic.c
index 5a29e2ff..0006663d 100644
--- a/src/strategies/generic/quant-generic.c
+++ b/src/strategies/generic/quant-generic.c
@@ -38,8 +38,8 @@ void kvz_quant_generic(const encoder_state_t * const state, coeff_t *coef, coeff
   int32_t height, int8_t type, int8_t scan_idx, int8_t block_type)
 {
   const encoder_control_t * const encoder = state->encoder_control;
-  //const uint32_t log2_block_size = kvz_g_convert_to_bit[width] + 2;
-  //const uint32_t * const scan = kvz_g_sig_last_scan[scan_idx][log2_block_size - 1];
+  const uint32_t log2_block_size = kvz_g_convert_to_bit[width] + 2;
+  const uint32_t * const scan = kvz_g_sig_last_scan[scan_idx][log2_block_size - 1];
 
   int32_t qp_scaled = kvz_get_scaled_qp(type, state->qp, (encoder->bitdepth - 8) * 6);
   const uint32_t log2_tr_size = kvz_g_convert_to_bit[width] + 2;
@@ -48,7 +48,7 @@ void kvz_quant_generic(const encoder_state_t * const state, coeff_t *coef, coeff
   const int32_t transform_shift = MAX_TR_DYNAMIC_RANGE - encoder->bitdepth - log2_tr_size; //!< Represents scaling through forward transform
   const int32_t q_bits = QUANT_SHIFT + qp_scaled / 6 + transform_shift;
   const int32_t add = ((state->frame->slicetype == KVZ_SLICE_I) ? 171 : 85) << (q_bits - 9);
-  //const int32_t q_bits8 = q_bits - 8;
+  const int32_t q_bits8 = q_bits - 8;
 
   uint32_t ac_sum = 0;
 
@@ -67,8 +67,8 @@ void kvz_quant_generic(const encoder_state_t * const state, coeff_t *coef, coeff
     q_coef[n] = (coeff_t)(CLIP(-32768, 32767, level));
 
   }
-  /*
-  // Signhiding disabled in VVC
+
+  // Signhiding
   if (!encoder->cfg.signhide_enable || ac_sum < 2) return;
 
   int32_t delta_u[LCU_WIDTH*LCU_WIDTH >> 2];
@@ -165,7 +165,6 @@ void kvz_quant_generic(const encoder_state_t * const state, coeff_t *coef, coeff
 #undef SCAN_SET_SIZE
 #undef LOG2_SCAN_SET_SIZE
   }
-  */
 }
 
 /**
@@ -220,7 +219,7 @@ int kvz_quantize_residual_generic(encoder_state_t *const state,
   }
 
   // Quantize coeffs. (coeff -> coeff_out)
-  /*
+  
   if (state->encoder_control->cfg.rdoq_enable &&
       (width > 4 || !state->encoder_control->cfg.rdoq_skip))
   {
@@ -228,7 +227,7 @@ int kvz_quantize_residual_generic(encoder_state_t *const state,
     tr_depth += (cur_cu->part_size == SIZE_NxN ? 1 : 0);
     kvz_rdoq(state, coeff, coeff_out, width, width, (color == COLOR_Y ? 0 : 2),
       scan_order, cur_cu->type, tr_depth);
-  } else*/ {
+  } else {
   
     kvz_quant(state, coeff, coeff_out, width, width, (color == COLOR_Y ? 0 : 2),
       scan_order, cur_cu->type);
diff --git a/tests/test_intra.sh b/tests/test_intra.sh
index 58eba361..d3d93b85 100755
--- a/tests/test_intra.sh
+++ b/tests/test_intra.sh
@@ -9,3 +9,6 @@ set -eu
 common_args='256x128 10 -p1 --preset=ultrafast --threads=0 --no-wpp --no-tmvp --no-deblock --sao=0 --alf=full --cpuid=0'
 valgrind_test $common_args --rd=1
 valgrind_test $common_args --rd=2 --no-transform-skip --qp 37
+valgrind_test $common_args --rd=2 --no-transform-skip --qp 37 --rdoq
+valgrind_test $common_args --rd=2 --no-transform-skip --qp 37 --signhide
+valgrind_test $common_args --rd=2 --no-transform-skip --qp 37 --signhide --rdoq