diff --git a/libswscale/Makefile b/libswscale/Makefile
index 5e03e6fa0a..4b8f9de425 100644
--- a/libswscale/Makefile
+++ b/libswscale/Makefile
@@ -25,5 +25,6 @@ OBJS-$(CONFIG_SHARED)        += log2_tab.o
 SLIBOBJS-$(HAVE_GNU_WINDRES) += swscaleres.o
 
 TESTPROGS = colorspace                                                  \
+            floatimg_cmp                                                \
             pixdesc_query                                               \
             swscale                                                     \
diff --git a/libswscale/tests/.gitignore b/libswscale/tests/.gitignore
index 1a26f038c4..c56abf0ee7 100644
--- a/libswscale/tests/.gitignore
+++ b/libswscale/tests/.gitignore
@@ -1,3 +1,4 @@
 /colorspace
+/floatimg_cmp
 /pixdesc_query
 /swscale
diff --git a/libswscale/tests/floatimg_cmp.c b/libswscale/tests/floatimg_cmp.c
new file mode 100644
index 0000000000..5c67594fb6
--- /dev/null
+++ b/libswscale/tests/floatimg_cmp.c
@@ -0,0 +1,296 @@
+/*
+ * This file is part of FFmpeg.
+ *
+ * FFmpeg is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; either
+ * version 2.1 of the License, or (at your option) any later version.
+ *
+ * FFmpeg is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <math.h>
+#include <inttypes.h>
+#include <float.h>
+
+#include "libavutil/avutil.h"
+#include "libavutil/imgutils.h"
+#include "libavutil/intfloat.h"
+#include "libavutil/intreadwrite.h"
+#include "libavutil/lfg.h"
+#include "libavutil/mem.h"
+#include "libavutil/parseutils.h"
+#include "libavutil/pixdesc.h"
+
+#include "libswscale/swscale.h"
+
+#define DEFAULT_W 96
+#define DEFAULT_H 96
+
+static const enum AVPixelFormat pix_fmts[] = {
+    AV_PIX_FMT_YUV444P16LE,
+    AV_PIX_FMT_YUV444P,
+    AV_PIX_FMT_YUV444P9LE, AV_PIX_FMT_YUV444P10LE,
+    AV_PIX_FMT_YUV444P12LE, AV_PIX_FMT_YUV444P14LE,
+    AV_PIX_FMT_RGB24,  AV_PIX_FMT_BGR24,
+    AV_PIX_FMT_RGBA,   AV_PIX_FMT_BGRA,
+    AV_PIX_FMT_ARGB,   AV_PIX_FMT_ABGR,
+    AV_PIX_FMT_0RGB,   AV_PIX_FMT_0BGR,
+    AV_PIX_FMT_RGB0,   AV_PIX_FMT_BGR0,
+    AV_PIX_FMT_RGB48LE,  AV_PIX_FMT_BGR48LE,
+    AV_PIX_FMT_RGBA64LE, AV_PIX_FMT_BGRA64LE,
+    AV_PIX_FMT_GBRP,   AV_PIX_FMT_GBRAP,
+    AV_PIX_FMT_GBRP9LE,
+    AV_PIX_FMT_GBRP10LE, AV_PIX_FMT_GBRAP10LE,
+    AV_PIX_FMT_GBRP12LE, AV_PIX_FMT_GBRAP12LE,
+    AV_PIX_FMT_GBRP14LE,
+    AV_PIX_FMT_GBRP16LE,  AV_PIX_FMT_GBRAP16LE
+};
+
+const char *usage =  "floatimg_cmp -pixel_format <pix_fmt> -size <image_size> -ref <testfile>\n";
+
+int main(int argc, char **argv)
+{
+    enum AVPixelFormat inFormat = AV_PIX_FMT_NONE;
+    enum AVPixelFormat dstFormat = AV_PIX_FMT_NONE;
+    const AVPixFmtDescriptor *desc;
+    uint8_t *ptr;
+    uint32_t *in, *out;
+
+    uint8_t *rgbIn[4]  = {NULL, NULL, NULL, NULL};
+    uint8_t *rgbOut[4] = {NULL, NULL, NULL, NULL};
+    int rgbStride[4];
+
+    uint8_t *dst[4] = {NULL, NULL, NULL, NULL};
+    int dstStride[4];
+
+    int i, x, y, p, size, count;
+    int res = -1;
+    int w = -1;
+    int h = -1;
+    union av_intfloat32 v0, v1;
+
+    double sum;
+    float minimum, maximum, diff;
+
+    struct SwsContext *sws = NULL;
+    AVLFG rand;
+    FILE *fp = NULL;
+
+    for (i = 1; i < argc; i += 2) {
+        if (argv[i][0] != '-' || i + 1 == argc)
+            goto bad_option;
+        if (!strcmp(argv[i], "-ref")) {
+            fp = fopen(argv[i + 1], "rb");
+            if (!fp) {
+                fprintf(stderr, "could not open '%s'\n", argv[i + 1]);
+                goto end;
+            }
+        } else if (!strcmp(argv[i], "-size")) {
+            res = av_parse_video_size(&w, &h, argv[i + 1]);
+            if (res < 0) {
+                fprintf(stderr, "invalid video size %s\n",  argv[i + 1]);
+                goto end;
+            }
+        } else if (!strcmp(argv[i], "-pixel_format")) {
+            inFormat = av_get_pix_fmt(argv[i + 1]);
+            if (inFormat == AV_PIX_FMT_NONE) {
+                fprintf(stderr, "invalid pixel format %s\n", argv[i + 1]);
+                goto end;
+            }
+        } else {
+bad_option:
+            fprintf(stderr, "%s", usage);
+            fprintf(stderr, "bad option or argument missing (%s)\n", argv[i]);
+            goto end;
+        };
+    }
+
+    if (!fp) {
+        inFormat = AV_PIX_FMT_GBRPF32LE;
+        w = DEFAULT_W;
+        h = DEFAULT_H;
+    }
+
+    if (w <= 0 || h <= 0) {
+        fprintf(stderr, "%s", usage);
+        fprintf(stderr, "invalid -video_size\n");
+        goto end;
+    }
+
+    if (inFormat == AV_PIX_FMT_NONE) {
+        fprintf(stderr, "%s", usage);
+        fprintf(stderr, "invalid input pixel format\n");
+        goto end;
+    }
+
+    desc = av_pix_fmt_desc_get(inFormat);
+    if (!(desc->flags & AV_PIX_FMT_FLAG_FLOAT)) {
+        fprintf(stderr, "input pixel format not floating point.\n");
+        goto end;
+    }
+
+    res = av_image_fill_linesizes(rgbStride, inFormat, w);
+    if (res < 0) {
+        fprintf(stderr, "av_image_fill_linesizes failed\n");
+        goto end;
+    }
+    for (p = 0; p < 4; p++) {
+        rgbStride[p] = FFALIGN(rgbStride[p], 16);
+        if (rgbStride[p]) {
+            rgbIn[p] = av_mallocz(rgbStride[p] * h + 16);
+            rgbOut[p] = av_mallocz(rgbStride[p] * h + 16);
+        }
+        if (rgbStride[p] && (!rgbIn[p] || !rgbOut[p])) {
+            goto end;
+        }
+    }
+
+    for (i = 0; i < FF_ARRAY_ELEMS(pix_fmts); i++) {
+        dstFormat = pix_fmts[i];
+        if (fp) {
+            fseek(fp, 0, SEEK_SET);
+            for (p = 0; p < 4; p++) {
+                if (!rgbStride[p])
+                    continue;
+
+                ptr = rgbIn[p];
+                for (y = 0; y < h; y++) {
+                    size = fread(ptr, 1, w*4, fp);
+                    if (size != w*4) {
+                        fprintf(stderr, "read error: %d\n", size);
+                        goto end;
+                    }
+                    ptr += rgbStride[p];
+                }
+            }
+        } else {
+            // fill src with random values between 0.0 - 1.0
+            av_lfg_init(&rand, 1);
+            for (p = 0; p < 4; p++) {
+                if (!rgbStride[p])
+                    continue;
+
+                for (y = 0; y < h; y++) {
+                    in = (uint32_t*)(rgbIn[p] + y * rgbStride[p]);
+                    for (x = 0; x < w; x++) {
+                        v0.f =  (float)av_lfg_get(&rand)/(float)(UINT32_MAX);
+                        *in++ = AV_RL32(&v0.i);
+                    }
+                }
+            }
+        }
+
+        // setup intermediate image
+        for (p = 0; p < 4; p++) {
+            av_freep(&dst[p]);
+        }
+
+        res = av_image_fill_linesizes(dstStride, dstFormat, w);
+        if (res < 0) {
+            fprintf(stderr, "av_image_fill_linesizes failed\n");
+            goto end;
+        }
+        for (p = 0; p < 4; p++) {
+            dstStride[p] = FFALIGN(dstStride[p], 16);
+            if (dstStride[p]) {
+                dst[p] = av_mallocz(dstStride[p] * h + 16);
+            }
+            if (dstStride[p] && !dst[p]) {
+                goto end;
+            }
+        }
+
+        // srcFormat -> dstFormat
+        sws = sws_getContext(w, h, inFormat, w, h,
+                            dstFormat, SWS_BILINEAR, NULL, NULL, NULL);
+        if (!sws) {
+            fprintf(stderr, "Failed to get %s -> %s\n", av_get_pix_fmt_name(inFormat), av_get_pix_fmt_name(dstFormat) );
+            goto end;
+        }
+
+        res = sws_scale(sws, (const uint8_t *const *)rgbIn, rgbStride, 0, h, dst, dstStride);
+        if (res < 0 || res != h) {
+            fprintf(stderr, "sws_scale failed\n");
+            res = -1;
+            goto end;
+        }
+        sws_freeContext(sws);
+
+        // dstFormat -> srcFormat
+        sws = sws_getContext(w, h, dstFormat, w, h,
+                            inFormat, SWS_BILINEAR, NULL, NULL, NULL);
+        if(!sws) {
+            fprintf(stderr, "Failed to get %s -> %s\n", av_get_pix_fmt_name(dstFormat), av_get_pix_fmt_name(inFormat) );
+            goto end;
+        }
+
+        res = sws_scale(sws, (const uint8_t *const *)dst, dstStride, 0, h, rgbOut, rgbStride);
+        if (res < 0 || res != h) {
+            fprintf(stderr, "sws_scale failed\n");
+            res = -1;
+            goto end;
+        }
+        sws_freeContext(sws);
+        sws = NULL;
+
+        minimum = FLT_MAX;
+        maximum = -FLT_MAX;
+        count = 0;
+        sum = 0.0;
+
+        for (p = 0; p < 4; p++) {
+            if (!rgbStride[p])
+                continue;
+
+            for (y = 0; y < h; y++) {
+                in = (uint32_t*)(rgbIn[p] + y * rgbStride[p]);
+                out = (uint32_t*)(rgbOut[p] + y * rgbStride[p]);
+                for (x = 0; x < w; x++) {
+                    if (desc->flags & AV_PIX_FMT_FLAG_BE) {
+                        v0.i = AV_RB32(in);
+                        v1.i = AV_RB32(out);
+                    } else {
+                        v0.i = AV_RL32(in);
+                        v1.i = AV_RL32(out);
+                    }
+
+                    diff = fabsf(v0.f - v1.f);
+                    sum += diff;
+                    minimum = FFMIN(minimum, diff);
+                    maximum = FFMAX(maximum, diff);
+
+                    count++;
+                    in++;
+                    out++;
+                }
+            }
+        }
+
+        fprintf(stdout, "%s -> %s -> %s\n", av_get_pix_fmt_name(inFormat), av_get_pix_fmt_name(dstFormat), av_get_pix_fmt_name(inFormat) );
+        fprintf(stdout, "avg diff: %f\nmin diff: %f\nmax diff: %f\n", sum / count, minimum, maximum);
+        res = 0;
+    }
+
+end:
+    sws_freeContext(sws);
+    for (p = 0; p < 4; p++) {
+        av_freep(&rgbIn[p]);
+        av_freep(&rgbOut[p]);
+        av_freep(&dst[p]);
+    }
+    if (fp)
+        fclose(fp);
+
+    return res;
+}
diff --git a/tests/fate/libswscale.mak b/tests/fate/libswscale.mak
index 68eb159fec..5ec5f34cc4 100644
--- a/tests/fate/libswscale.mak
+++ b/tests/fate/libswscale.mak
@@ -2,6 +2,10 @@ FATE_LIBSWSCALE += fate-sws-pixdesc-query
 fate-sws-pixdesc-query: libswscale/tests/pixdesc_query$(EXESUF)
 fate-sws-pixdesc-query: CMD = run libswscale/tests/pixdesc_query$(EXESUF)
 
+FATE_LIBSWSCALE += fate-sws-floatimg-cmp
+fate-sws-floatimg-cmp: libswscale/tests/floatimg_cmp$(EXESUF)
+fate-sws-floatimg-cmp: CMD = run libswscale/tests/floatimg_cmp$(EXESUF)
+
 FATE_LIBSWSCALE += $(FATE_LIBSWSCALE-yes)
 FATE-$(CONFIG_SWSCALE) += $(FATE_LIBSWSCALE)
 fate-libswscale: $(FATE_LIBSWSCALE)
diff --git a/tests/ref/fate/sws-floatimg-cmp b/tests/ref/fate/sws-floatimg-cmp
new file mode 100644
index 0000000000..24204254c4
--- /dev/null
+++ b/tests/ref/fate/sws-floatimg-cmp
@@ -0,0 +1,120 @@
+gbrpf32le -> yuv444p16le -> gbrpf32le
+avg diff: 0.003852
+min diff: 0.000000
+max diff: 0.006638
+gbrpf32le -> yuv444p -> gbrpf32le
+avg diff: 0.004316
+min diff: 0.000000
+max diff: 0.012704
+gbrpf32le -> yuv444p9le -> gbrpf32le
+avg diff: 0.004053
+min diff: 0.000001
+max diff: 0.009402
+gbrpf32le -> yuv444p10le -> gbrpf32le
+avg diff: 0.003960
+min diff: 0.000000
+max diff: 0.008123
+gbrpf32le -> yuv444p12le -> gbrpf32le
+avg diff: 0.003878
+min diff: 0.000000
+max diff: 0.007011
+gbrpf32le -> yuv444p14le -> gbrpf32le
+avg diff: 0.003868
+min diff: 0.000000
+max diff: 0.006729
+gbrpf32le -> rgb24 -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> bgr24 -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> rgba -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> bgra -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> argb -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> abgr -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> 0rgb -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> 0bgr -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> rgb0 -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> bgr0 -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> rgb48le -> gbrpf32le
+avg diff: 0.003851
+min diff: 0.000000
+max diff: 0.007076
+gbrpf32le -> bgr48le -> gbrpf32le
+avg diff: 0.003851
+min diff: 0.000000
+max diff: 0.007076
+gbrpf32le -> rgba64le -> gbrpf32le
+avg diff: 0.003851
+min diff: 0.000000
+max diff: 0.007076
+gbrpf32le -> bgra64le -> gbrpf32le
+avg diff: 0.003851
+min diff: 0.000000
+max diff: 0.007076
+gbrpf32le -> gbrp -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> gbrap -> gbrpf32le
+avg diff: 0.004122
+min diff: 0.000000
+max diff: 0.008975
+gbrpf32le -> gbrp9le -> gbrpf32le
+avg diff: 0.007737
+min diff: 0.000000
+max diff: 0.014009
+gbrpf32le -> gbrp10le -> gbrpf32le
+avg diff: 0.007662
+min diff: 0.000000
+max diff: 0.013605
+gbrpf32le -> gbrap10le -> gbrpf32le
+avg diff: 0.007662
+min diff: 0.000000
+max diff: 0.013605
+gbrpf32le -> gbrp12le -> gbrpf32le
+avg diff: 0.007622
+min diff: 0.000000
+max diff: 0.013335
+gbrpf32le -> gbrap12le -> gbrpf32le
+avg diff: 0.007622
+min diff: 0.000000
+max diff: 0.013335
+gbrpf32le -> gbrp14le -> gbrpf32le
+avg diff: 0.007620
+min diff: 0.000000
+max diff: 0.013232
+gbrpf32le -> gbrp16le -> gbrpf32le
+avg diff: 0.007680
+min diff: 0.000000
+max diff: 0.013275
+gbrpf32le -> gbrap16le -> gbrpf32le
+avg diff: 0.007680
+min diff: 0.000000
+max diff: 0.013275