diff --git a/libavcodec/pngenc.c b/libavcodec/pngenc.c
index 58acc3c409..bf61be1609 100644
--- a/libavcodec/pngenc.c
+++ b/libavcodec/pngenc.c
@@ -38,7 +38,6 @@ typedef struct PNGEncContext {
     uint8_t *bytestream;
     uint8_t *bytestream_start;
     uint8_t *bytestream_end;
-    AVFrame picture;
 
     int filter_type;
 
@@ -218,7 +217,7 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
                         const AVFrame *pict, int *got_packet)
 {
     PNGEncContext *s = avctx->priv_data;
-    AVFrame * const p= &s->picture;
+    const AVFrame * const p = pict;
     int bit_depth, color_type, y, len, row_size, ret, is_progressive;
     int bits_per_pixel, pass_row_size, enc_row_size;
     int64_t max_packet_size;
@@ -228,10 +227,6 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
     uint8_t *progressive_buf = NULL;
     uint8_t *top_buf = NULL;
 
-    *p = *pict;
-    p->pict_type= AV_PICTURE_TYPE_I;
-    p->key_frame= 1;
-
     is_progressive = !!(avctx->flags & CODEC_FLAG_INTERLACED_DCT);
     switch(avctx->pix_fmt) {
     case AV_PIX_FMT_RGBA64BE:
@@ -454,8 +449,13 @@ static av_cold int png_enc_init(AVCodecContext *avctx){
         avctx->bits_per_coded_sample = 8;
     }
 
-    avcodec_get_frame_defaults(&s->picture);
-    avctx->coded_frame= &s->picture;
+    avctx->coded_frame = av_frame_alloc();
+    if (!avctx->coded_frame)
+        return AVERROR(ENOMEM);
+
+    avctx->coded_frame->pict_type = AV_PICTURE_TYPE_I;
+    avctx->coded_frame->key_frame = 1;
+
     ff_dsputil_init(&s->dsp, avctx);
 
     s->filter_type = av_clip(avctx->prediction_method, PNG_FILTER_VALUE_NONE, PNG_FILTER_VALUE_MIXED);
@@ -472,6 +472,12 @@ static av_cold int png_enc_init(AVCodecContext *avctx){
     return 0;
 }
 
+static av_cold int png_enc_close(AVCodecContext *avctx)
+{
+    av_frame_free(&avctx->coded_frame);
+    return 0;
+}
+
 #define OFFSET(x) offsetof(PNGEncContext, x)
 #define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
 static const AVOption options[] = {
@@ -494,6 +500,7 @@ AVCodec ff_png_encoder = {
     .id             = AV_CODEC_ID_PNG,
     .priv_data_size = sizeof(PNGEncContext),
     .init           = png_enc_init,
+    .close          = png_enc_close,
     .encode2        = encode_frame,
     .capabilities   = CODEC_CAP_FRAME_THREADS | CODEC_CAP_INTRA_ONLY,
     .pix_fmts       = (const enum AVPixelFormat[]){
diff --git a/libavcodec/sunrastenc.c b/libavcodec/sunrastenc.c
index 0879ce62cb..a55e3d439a 100644
--- a/libavcodec/sunrastenc.c
+++ b/libavcodec/sunrastenc.c
@@ -198,6 +198,12 @@ static int sunrast_encode_frame(AVCodecContext *avctx,  AVPacket *avpkt,
     return 0;
 }
 
+static av_cold int sunrast_encode_close(AVCodecContext *avctx)
+{
+    av_frame_free(&avctx->coded_frame);
+    return 0;
+}
+
 static const AVCodecDefault sunrast_defaults[] = {
      { "coder", "rle" },
      { NULL },
@@ -210,6 +216,7 @@ AVCodec ff_sunrast_encoder = {
     .id             = AV_CODEC_ID_SUNRAST,
     .priv_data_size = sizeof(SUNRASTContext),
     .init           = sunrast_encode_init,
+    .close          = sunrast_encode_close,
     .encode2        = sunrast_encode_frame,
     .defaults       = sunrast_defaults,
     .pix_fmts       = (const enum AVPixelFormat[]){ AV_PIX_FMT_BGR24,
diff --git a/libavcodec/targaenc.c b/libavcodec/targaenc.c
index 108e1485e4..d4483ec398 100644
--- a/libavcodec/targaenc.c
+++ b/libavcodec/targaenc.c
@@ -170,11 +170,31 @@ static int targa_encode_frame(AVCodecContext *avctx, AVPacket *pkt,
     return 0;
 }
 
+static av_cold int targa_encode_init(AVCodecContext *avctx)
+{
+    avctx->coded_frame = av_frame_alloc();
+    if (!avctx->coded_frame)
+        return AVERROR(ENOMEM);
+
+    avctx->coded_frame->key_frame = 1;
+    avctx->coded_frame->pict_type = AV_PICTURE_TYPE_I;
+
+    return 0;
+}
+
+static av_cold int targa_encode_close(AVCodecContext *avctx)
+{
+    av_frame_free(&avctx->coded_frame);
+    return 0;
+}
+
 AVCodec ff_targa_encoder = {
     .name           = "targa",
     .long_name      = NULL_IF_CONFIG_SMALL("Truevision Targa image"),
     .type           = AVMEDIA_TYPE_VIDEO,
     .id             = AV_CODEC_ID_TARGA,
+    .init           = targa_encode_init,
+    .close          = targa_encode_close,
     .encode2        = targa_encode_frame,
     .pix_fmts       = (const enum AVPixelFormat[]){
         AV_PIX_FMT_BGR24, AV_PIX_FMT_BGRA, AV_PIX_FMT_RGB555LE, AV_PIX_FMT_GRAY8, AV_PIX_FMT_PAL8,
diff --git a/libavcodec/tiffenc.c b/libavcodec/tiffenc.c
index f5d04eaf2b..7b1e5104bd 100644
--- a/libavcodec/tiffenc.c
+++ b/libavcodec/tiffenc.c
@@ -52,7 +52,6 @@ static const uint8_t type_sizes2[14] = {
 typedef struct TiffEncoderContext {
     AVClass *class;                         ///< for private options
     AVCodecContext *avctx;
-    AVFrame picture;
 
     int width;                              ///< picture width
     int height;                             ///< picture height
@@ -195,9 +194,9 @@ static int encode_strip(TiffEncoderContext *s, const int8_t *src,
     }
 }
 
-static void pack_yuv(TiffEncoderContext *s, uint8_t *dst, int lnum)
+static void pack_yuv(TiffEncoderContext *s, const AVFrame *p,
+                     uint8_t *dst, int lnum)
 {
-    AVFrame *p = &s->picture;
     int i, j, k;
     int w       = (s->width - 1) / s->subsampling[0] + 1;
     uint8_t *pu = &p->data[1][lnum / s->subsampling[1] * p->linesize[1]];
@@ -223,24 +222,12 @@ static void pack_yuv(TiffEncoderContext *s, uint8_t *dst, int lnum)
     }
 }
 
-static av_cold int encode_init(AVCodecContext *avctx)
-{
-    TiffEncoderContext *s = avctx->priv_data;
-
-    avctx->coded_frame            = &s->picture;
-    avctx->coded_frame->pict_type = AV_PICTURE_TYPE_I;
-    avctx->coded_frame->key_frame = 1;
-    s->avctx = avctx;
-
-    return 0;
-}
-
 static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
                         const AVFrame *pict, int *got_packet)
 {
     const AVPixFmtDescriptor *desc = av_pix_fmt_desc_get(avctx->pix_fmt);
     TiffEncoderContext *s = avctx->priv_data;
-    AVFrame *const p = &s->picture;
+    const AVFrame *const p = pict;
     int i;
     uint8_t *ptr;
     uint8_t *offset;
@@ -252,8 +239,6 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
     int is_yuv = 0, alpha = 0;
     int shift_h, shift_v;
 
-    *p = *pict;
-
     s->width          = avctx->width;
     s->height         = avctx->height;
     s->subsampling[0] = 1;
@@ -373,7 +358,7 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
         zn               = 0;
         for (j = 0; j < s->rps; j++) {
             if (is_yuv) {
-                pack_yuv(s, s->yuv_line, j);
+                pack_yuv(s, p, s->yuv_line, j);
                 memcpy(zbuf + zn, s->yuv_line, bytes_per_row);
                 j += s->subsampling[1] - 1;
             } else
@@ -409,7 +394,7 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
             s->strip_offsets[i / s->rps] = ptr - pkt->data;
         }
         if (is_yuv) {
-            pack_yuv(s, s->yuv_line, i);
+            pack_yuv(s, p, s->yuv_line, i);
             ret = encode_strip(s, s->yuv_line, ptr, bytes_per_row, s->compr);
             i  += s->subsampling[1] - 1;
         } else
@@ -497,10 +482,26 @@ fail:
     return ret < 0 ? ret : 0;
 }
 
+static av_cold int encode_init(AVCodecContext *avctx)
+{
+    TiffEncoderContext *s = avctx->priv_data;
+
+    avctx->coded_frame = av_frame_alloc();
+    if (!avctx->coded_frame)
+        return AVERROR(ENOMEM);
+
+    avctx->coded_frame->pict_type = AV_PICTURE_TYPE_I;
+    avctx->coded_frame->key_frame = 1;
+    s->avctx = avctx;
+
+    return 0;
+}
+
 static av_cold int encode_close(AVCodecContext *avctx)
 {
     TiffEncoderContext *s = avctx->priv_data;
 
+    av_frame_free(&avctx->coded_frame);
     av_freep(&s->strip_sizes);
     av_freep(&s->strip_offsets);
     av_freep(&s->yuv_line);
@@ -536,8 +537,8 @@ AVCodec ff_tiff_encoder = {
     .id             = AV_CODEC_ID_TIFF,
     .priv_data_size = sizeof(TiffEncoderContext),
     .init           = encode_init,
-    .encode2        = encode_frame,
     .close          = encode_close,
+    .encode2        = encode_frame,
     .pix_fmts       = (const enum AVPixelFormat[]) {
         AV_PIX_FMT_RGB24, AV_PIX_FMT_PAL8, AV_PIX_FMT_GRAY8,
         AV_PIX_FMT_GRAY8A, AV_PIX_FMT_GRAY16LE,