On Mon, 17 Oct 2011, Måns Rullgård wrote:
Martin Storsjö <[email protected]> writes:
---
tests/fate2.mak | 4 ++
tests/ref/fate/g722enc | 94 ++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 98 insertions(+), 0 deletions(-)
create mode 100644 tests/ref/fate/g722enc
diff --git a/tests/fate2.mak b/tests/fate2.mak
index 0bb6654..4653da5 100644
--- a/tests/fate2.mak
+++ b/tests/fate2.mak
@@ -133,6 +133,10 @@ fate-gsm-ms: CMD = framecrc -i $(SAMPLES)/gsm/ciao.wav
FATE_TESTS += fate-g722dec-1
fate-g722dec-1: CMD = framecrc -i $(SAMPLES)/g722/conf-adminmenu-162.g722
+FATE_TESTS += fate-g722enc
+fate-g722enc: tests/data/asynth-16000-1.sw
+fate-g722enc: CMD = framecrc -ar 16000 -ac 1 -f s16le -i
tests/data/asynth-16000-1.sw -acodec g722 -ac 1
+
FATE_TESTS += fate-msmpeg4v1
fate-msmpeg4v1: CMD = framecrc -flags +bitexact -dct fastint -idct simple -i
$(SAMPLES)/msmpeg4v1/mpg4.avi -an
diff --git a/tests/ref/fate/g722enc b/tests/ref/fate/g722enc
new file mode 100644
index 0000000..77f794c
--- /dev/null
+++ b/tests/ref/fate/g722enc
@@ -0,0 +1,94 @@
+0, 0, 512, 0xec3e85e4
+0, 2880, 512, 0xb2e48424
+0, 5760, 512, 0x4f37851a
+0, 8640, 512, 0xea8d85ef
Is there any point in per-frame checksums here, or would a single
checksum of the full output be just as good? In other words, is it
likely for only a few of the frames to change and would knowing which
ones did provide useful information? If not, this is just wasting space.
Good point - it should be bitexact. If it fails, I'm not really sure if it
helps to have per-frame crcs (the codec doesn't have any inherent frame
size in itself, so the cutting is quite arbitrary). So if any difference
is found in any configuration, I guess it requires just as much
investigation either way.
// Martin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
