diff --git a/include/st_convert_api.h b/include/st_convert_api.h
index c384c63c2..0b8a05543 100644
--- a/include/st_convert_api.h
+++ b/include/st_convert_api.h
@@ -249,6 +249,31 @@ static inline int st20_rfc4175_422be10_to_yuv422p8(struct st20_rfc4175_422_10_pg
   return st20_rfc4175_422be10_to_yuv422p8_simd(pg, y, b, r, w, h, MTL_SIMD_LEVEL_MAX);
 }
 
+/**
+ * Convert rfc4175_422be10 to yuv420p8 with the max optimized SIMD level.
+ *
+ * @param pg
+ *   Point to pg(rfc4175_422be10) data.
+ * @param y
+ *   Point to Y(yuv420p8) vector.
+ * @param b
+ *   Point to b(yuv420p8) vector.
+ * @param r
+ *   Point to r(yuv420p8) vector.
+ * @param w
+ *   The st2110-20(video) width.
+ * @param h
+ *   The st2110-20(video) height.
+ * @return
+ *   - 0 if successful.
+ *   - <0: Error code if convert fail.
+ */
+static inline int st20_rfc4175_422be10_to_yuv420p8(struct st20_rfc4175_422_10_pg2_be* pg,
+                                                   uint8_t* y, uint8_t* b, uint8_t* r,
+                                                   uint32_t w, uint32_t h) {
+  return st20_rfc4175_422be10_to_yuv420p8_simd(pg, y, b, r, w, h, MTL_SIMD_LEVEL_MAX);
+}
+
 /**
  * Convert rfc4175_422be12 to yuv422p12le with the max optimized SIMD level.
  *
diff --git a/include/st_convert_internal.h b/include/st_convert_internal.h
index e991189eb..93edaa7e2 100644
--- a/include/st_convert_internal.h
+++ b/include/st_convert_internal.h
@@ -266,6 +266,32 @@ int st20_rfc4175_422be10_to_yuv422p8_simd(struct st20_rfc4175_422_10_pg2_be* pg,
                                           uint8_t* y, uint8_t* b, uint8_t* r, uint32_t w,
                                           uint32_t h, enum mtl_simd_level level);
 
+/**
+ * Convert rfc4175_422be10 to yuv420p8 with required SIMD level.
+ * Note the level may downgrade to the SIMD which system really support.
+ *
+ * @param pg
+ *   Point to pg(rfc4175_422be10) data.
+ * @param y
+ *   Point to Y(yuv420p8) vector.
+ * @param b
+ *   Point to b(yuv420p8) vector.
+ * @param r
+ *   Point to r(yuv420p8) vector.
+ * @param w
+ *   The st2110-20(video) width.
+ * @param h
+ *   The st2110-20(video) height.
+ * @param level
+ *   simd level.
+ * @return
+ *   - 0 if successful.
+ *   - <0: Error code if convert fail.
+ */
+int st20_rfc4175_422be10_to_yuv420p8_simd(struct st20_rfc4175_422_10_pg2_be* pg,
+                                          uint8_t* y, uint8_t* b, uint8_t* r, uint32_t w,
+                                          uint32_t h, enum mtl_simd_level level);
+
 /**
  * Convert rfc4175_422be12 to yuv422p12le with required SIMD level.
  * Note the level may downgrade to the SIMD which system really support.
diff --git a/include/st_pipeline_api.h b/include/st_pipeline_api.h
index cb66fd401..433040731 100644
--- a/include/st_pipeline_api.h
+++ b/include/st_pipeline_api.h
@@ -146,6 +146,8 @@ enum st_frame_fmt {
    * transport frame without conversion. The frame should not have lines padding.
    */
   ST_FRAME_FMT_YUV422CUSTOM8 = 13,
+  /** YUV 420 planar 8bit */
+  ST_FRAME_FMT_YUV420PLANAR8 = 14,
   /** End of yuv format list, new yuv should be inserted before this */
   ST_FRAME_FMT_YUV_END,
 
diff --git a/lib/src/st2110/pipeline/st20_pipeline_rx.c b/lib/src/st2110/pipeline/st20_pipeline_rx.c
index ca360de29..5c74c9201 100644
--- a/lib/src/st2110/pipeline/st20_pipeline_rx.c
+++ b/lib/src/st2110/pipeline/st20_pipeline_rx.c
@@ -127,6 +127,14 @@ static int rx_st20p_packet_convert(void* priv, void* frame,
     uint8_t* r = (uint8_t*)framebuff->dst.addr[2] +
                  framebuff->dst.linesize[2] * meta->row_number + meta->row_offset;
     ret = st20_rfc4175_422be10_to_yuv422p8(src, y, b, r, meta->pg_cnt, 2);
+  } else if (ctx->ops.output_fmt == ST_FRAME_FMT_YUV420PLANAR8) {
+    uint8_t* y = (uint8_t*)framebuff->dst.addr[0] +
+                 framebuff->dst.linesize[0] * meta->row_number + meta->row_offset * 2;
+    uint8_t* b = (uint8_t*)framebuff->dst.addr[1] +
+                 framebuff->dst.linesize[1] * meta->row_number + meta->row_offset;
+    uint8_t* r = (uint8_t*)framebuff->dst.addr[2] +
+                 framebuff->dst.linesize[2] * meta->row_number + meta->row_offset;
+    ret = st20_rfc4175_422be10_to_yuv420p8(src, y, b, r, meta->pg_cnt, 2);
   }
 
   return ret;
diff --git a/lib/src/st2110/st_convert.c b/lib/src/st2110/st_convert.c
index 8341d1a0e..784d86723 100644
--- a/lib/src/st2110/st_convert.c
+++ b/lib/src/st2110/st_convert.c
@@ -111,6 +111,33 @@ static int convert_rfc4175_422be10_to_yuv422p8(struct st_frame* src,
   return ret;
 }
 
+static int convert_rfc4175_422be10_to_yuv420p8(struct st_frame* src,
+                                               struct st_frame* dst) {
+  int ret = 0;
+  struct st20_rfc4175_422_10_pg2_be* be10 = NULL;
+  uint8_t* y = NULL;
+  uint8_t* b = NULL;
+  uint8_t* r = NULL;
+  uint32_t h = st_frame_data_height(dst);
+
+  if (!has_lines_padding(src, dst)) {
+    be10 = src->addr[0];
+    y = dst->addr[0];
+    b = dst->addr[1];
+    r = dst->addr[2];
+    ret = st20_rfc4175_422be10_to_yuv420p8(be10, y, b, r, dst->width, h);
+  } else {
+    for (uint32_t line = 0; line < h; line++) {
+      be10 = src->addr[0] + src->linesize[0] * line;
+      y = dst->addr[0] + dst->linesize[0] * line;
+      b = dst->addr[1] + dst->linesize[1] * line;
+      r = dst->addr[2] + dst->linesize[2] * line;
+      ret = st20_rfc4175_422be10_to_yuv420p8(be10, y, b, r, dst->width, 1);
+    }
+  }
+  return ret;
+}
+
 static int convert_rfc4175_422be10_to_v210(struct st_frame* src, struct st_frame* dst) {
   int ret = 0;
   struct st20_rfc4175_422_10_pg2_be* be10 = NULL;
@@ -504,6 +531,11 @@ static const struct st_frame_converter converters[] = {
         .dst_fmt = ST_FRAME_FMT_YUV422PLANAR8,
         .convert_func = convert_rfc4175_422be10_to_yuv422p8,
     },
+    {
+        .src_fmt = ST_FRAME_FMT_YUV422RFC4175PG2BE10,
+        .dst_fmt = ST_FRAME_FMT_YUV420PLANAR8,
+        .convert_func = convert_rfc4175_422be10_to_yuv420p8,
+    },
     {
         .src_fmt = ST_FRAME_FMT_YUV422RFC4175PG2BE10,
         .dst_fmt = ST_FRAME_FMT_V210,
@@ -1329,6 +1361,45 @@ int st20_rfc4175_422be10_to_yuv422p8_simd(struct st20_rfc4175_422_10_pg2_be* pg,
   return st20_rfc4175_422be10_to_yuv422p8_scalar(pg, y, b, r, w, h);
 }
 
+static int st20_rfc4175_422be10_to_yuv420p8_scalar(struct st20_rfc4175_422_10_pg2_be* pg,
+                                                   uint8_t* y, uint8_t* b, uint8_t* r,
+                                                   uint32_t w, uint32_t h) {
+  uint32_t line_pg_cnt = w / 2; /* two pgs in one convert */
+
+  for (uint32_t i = 0; i < (h / 2); i++) { /* 2 lines each loop */
+    /* first line */
+    for (uint32_t j = 0; j < line_pg_cnt; j++) {
+      *b++ = pg->Cb00;
+      *y++ = (pg->Y00 << 2) | (pg->Y00_ >> 2);
+      *r++ = (pg->Cr00 << 4) | (pg->Cr00_ >> 2);
+      *y++ = (pg->Y01 << 6) | (pg->Y01_ >> 2);
+      pg++;
+    }
+    /* second line, no u and v */
+    for (uint32_t j = 0; j < line_pg_cnt; j++) {
+      *y++ = (pg->Y00 << 2) | (pg->Y00_ >> 2);
+      *y++ = (pg->Y01 << 6) | (pg->Y01_ >> 2);
+      pg++;
+    }
+  }
+
+  return 0;
+}
+
+int st20_rfc4175_422be10_to_yuv420p8_simd(struct st20_rfc4175_422_10_pg2_be* pg,
+                                          uint8_t* y, uint8_t* b, uint8_t* r, uint32_t w,
+                                          uint32_t h, enum mtl_simd_level level) {
+  enum mtl_simd_level cpu_level = mtl_get_simd_level();
+  int ret;
+
+  MTL_MAY_UNUSED(cpu_level);
+  MTL_MAY_UNUSED(level);
+  MTL_MAY_UNUSED(ret);
+
+  /* the last option */
+  return st20_rfc4175_422be10_to_yuv420p8_scalar(pg, y, b, r, w, h);
+}
+
 int st20_rfc4175_422le10_to_v210_scalar(uint8_t* pg_le, uint8_t* pg_v210, uint32_t w,
                                         uint32_t h) {
   uint32_t pg_count = w * h / 2;
diff --git a/lib/src/st2110/st_fmt.c b/lib/src/st2110/st_fmt.c
index 98621711f..47d5e497b 100644
--- a/lib/src/st2110/st_fmt.c
+++ b/lib/src/st2110/st_fmt.c
@@ -352,6 +352,13 @@ static const struct st_frame_fmt_desc st_frame_fmt_descs[] = {
         .planes = 1,
         .sampling = ST_FRAME_SAMPLING_422,
     },
+    {
+        /* ST_FRAME_FMT_YUV420PLANAR8 */
+        .fmt = ST_FRAME_FMT_YUV420PLANAR8,
+        .name = "YUV420PLANAR8",
+        .planes = 3,
+        .sampling = ST_FRAME_SAMPLING_420,
+    },
     {
         /* ST_FRAME_FMT_RGBRFC4175PG4BE10 */
         .fmt = ST_FRAME_FMT_RGBRFC4175PG4BE10,
@@ -522,6 +529,7 @@ size_t st_frame_size(enum st_frame_fmt fmt, uint32_t width, uint32_t height,
       size = pixels * 3; /* 8 bits RGB pixel in a 24 bits */
       break;
     case ST_FRAME_FMT_YUV420CUSTOM8:
+    case ST_FRAME_FMT_YUV420PLANAR8:
       size = st20_frame_size(ST20_FMT_YUV_420_8BIT, width, height);
       break;
     default:
diff --git a/python/example/misc_util.py b/python/example/misc_util.py
index 34a0d21df..880564d44 100644
--- a/python/example/misc_util.py
+++ b/python/example/misc_util.py
@@ -131,6 +131,19 @@ def ptr_to_yuv422p8(ptr, width, height):
     return y, u, v
 
 
+def ptr_to_yuv420p8(ptr, width, height):
+    y_size = width * height
+    u_size = width * height // 4
+
+    frame = np.frombuffer(ptr, dtype=np.uint8)
+
+    y = frame[:y_size].reshape((height, width))
+    u = frame[y_size : y_size + u_size].reshape((height // 2, width // 2))
+    v = frame[y_size + u_size :].reshape((height // 2, width // 2))
+
+    return y, u, v
+
+
 def yuv422p10le_to_yuv422(ptr, width, height):
     y_size = width * height
     u_size = width * height // 2
@@ -170,6 +183,28 @@ def downscale_yuv422(y, u, v, scale_factor):
     return y_downscaled, u_downscaled, v_downscaled
 
 
+def downscale_yuv420(y, u, v, scale_factor):
+    height, width = y.shape
+
+    y_downscaled = cv2.resize(
+        y,
+        (width // scale_factor, height // scale_factor),
+        interpolation=cv2.INTER_LINEAR,
+    )
+    u_downscaled = cv2.resize(
+        u,
+        (width // (2 * scale_factor), height // (2 * scale_factor)),
+        interpolation=cv2.INTER_LINEAR,
+    )
+    v_downscaled = cv2.resize(
+        v,
+        (width // (2 * scale_factor), height // (2 * scale_factor)),
+        interpolation=cv2.INTER_LINEAR,
+    )
+
+    return y_downscaled, u_downscaled, v_downscaled
+
+
 def display_yuv422(y, u, v):
     u = cv2.resize(u, (y.shape[1], y.shape[0]), interpolation=cv2.INTER_LINEAR)
     v = cv2.resize(v, (y.shape[1], y.shape[0]), interpolation=cv2.INTER_LINEAR)
@@ -180,6 +215,18 @@ def display_yuv422(y, u, v):
     cv2.waitKey(1)
 
 
+def display_yuv420(y, u, v):
+    u = cv2.resize(u, (y.shape[1], y.shape[0]), interpolation=cv2.INTER_LINEAR)
+    v = cv2.resize(v, (y.shape[1], y.shape[0]), interpolation=cv2.INTER_LINEAR)
+
+    yuv = cv2.merge([y, u, v])
+
+    # display_img = cv2.cvtColor(yuv, cv2.COLOR_YUV2BGR_I420)
+    display_img = cv2.cvtColor(yuv, cv2.COLOR_YUV2BGR)
+    cv2.imshow(sys.argv[0], display_img)
+    cv2.waitKey(1)
+
+
 def destroy():
     cv2.destroyAllWindows()
 
@@ -200,6 +247,22 @@ def frame_display_yuv422p8(frame, display_scale_factor):
     display_yuv422(y_scale, u_scale, v_scale)
 
 
+def frame_display_yuv420p8(frame, display_scale_factor):
+    # Pack frame pointer from frame addr
+    ptr = (ctypes.c_ubyte * (frame.data_size)).from_address(
+        mtl.st_frame_addr_cpuva(frame, 0)
+    )
+    width = frame.width
+    height = frame.height
+
+    # Convert to yuv420 planar 8
+    y, u, v = ptr_to_yuv420p8(ptr, width, height)
+    # Downscale
+    y_scale, u_scale, v_scale = downscale_yuv420(y, u, v, display_scale_factor)
+    # Convert to RGB and display
+    display_yuv420(y_scale, u_scale, v_scale)
+
+
 def frame_display_yuv422p10le(frame, display_scale_factor):
     # Pack frame pointer from frame addr
     ptr = (ctypes.c_ubyte * (frame.data_size)).from_address(
@@ -259,6 +322,8 @@ def frame_display(mtl_handle, frame, display_scale_factor):
         frame_display_yuv422p8(frame, display_scale_factor)
     elif frame.fmt == mtl.ST_FRAME_FMT_YUV422RFC4175PG2BE10:
         frame_display_rfc4175be10(mtl_handle, frame, display_scale_factor)
+    elif frame.fmt == mtl.ST_FRAME_FMT_YUV420PLANAR8:
+        frame_display_yuv420p8(frame, display_scale_factor)
     else:
         print(f"Unknown fmt: {mtl.st_frame_fmt_name(frame.fmt)}")