diff --git a/src/climatebenchpress/compressor/compressors/__init__.py b/src/climatebenchpress/compressor/compressors/__init__.py
index b523c75..09d1a93 100644
--- a/src/climatebenchpress/compressor/compressors/__init__.py
+++ b/src/climatebenchpress/compressor/compressors/__init__.py
@@ -1,6 +1,7 @@
 __all__ = [
     "BitRound",
     "BitRoundPco",
+    "FFmpeg",
     "Jpeg2000",
     "Sperr",
     "StochRound",
@@ -14,6 +15,7 @@
 from . import abc as abc
 from .bitround import BitRound
 from .bitround_pco import BitRoundPco
+from .ffmpeg import FFmpeg
 from .jpeg2000 import Jpeg2000
 from .sperr import Sperr
 from .stochround import StochRound
diff --git a/src/climatebenchpress/compressor/compressors/ffmpeg.py b/src/climatebenchpress/compressor/compressors/ffmpeg.py
new file mode 100644
index 0000000..bda7cc7
--- /dev/null
+++ b/src/climatebenchpress/compressor/compressors/ffmpeg.py
@@ -0,0 +1,75 @@
+__all__ = ["FFmpeg"]
+
+import numpy as np
+import numcodecs.astype
+import numcodecs.compat
+import numcodecs_wasm_fixed_offset_scale
+import numcodecs_wasm_round
+from numcodecs.abc import Codec
+from numcodecs_combinators.stack import CodecStack
+
+from .abc import Compressor
+
+
+class FFmpeg(Compressor):
+    name = "ffmpeg"
+    description = "FFmpeg"
+
+    @staticmethod
+    def abs_bound_codec(
+        error_bound,
+        *,
+        data_min=None,
+        data_max=None,
+        **kwargs,
+    ):
+        assert data_min is not None, "data_min must be provided"
+        assert data_max is not None, "data_max must be provided"
+
+        max_pixel_val = 2**12 - 1  # maximum pixel value for our integer encoding.
+
+        data_range = data_max - data_min
+
+        return CodecStack(
+            # increase precision for better rounding during linear quantization
+            numcodecs.astype.AsType(
+                encode_dtype="float64",
+                decode_dtype="float32",
+            ),
+            # remap from [min, max] to [0, max_pixel_val]
+            numcodecs_wasm_fixed_offset_scale.FixedOffsetScale(
+                offset=data_min,
+                scale=data_range / max_pixel_val,
+            ),
+            # round and truncate to integer values
+            numcodecs_wasm_round.Round(precision=1),
+            numcodecs.astype.AsType(
+                encode_dtype="uint32",
+                decode_dtype="float64",
+            ),
+            # apply the ffmpeg codec
+            FFmpegCodec(),
+        )
+
+
+class FFmpegCodec(Codec):
+    codec_id = "ffmpeg"
+
+    def __init__(self, *args, **kwargs):
+        # handle config
+        pass
+
+    def encode(self, buf) -> bytes:
+        a = numcodecs.compat.ensure_ndarray(buf)
+
+        # use ffmpeg to encode the array as a video
+
+        return b"encoded"
+
+    def decode(self, buf, out=None):
+        b = numcodecs.compat.ensure_bytes(buf)
+
+        # use ffmpeg to decode the data
+        decoded = np.array()
+
+        return numcodecs.compat.ndarray_copy(decoded, out)