[Libav-user] A question about "Input contains (near) NaN/+-Inf"

[zzzzzz]

Hello,
    Here's a question. Sometimes when my code goes to "ret = avcodec_send_frame(enc_ctx, tmpFrame);", the value of ret is AVERROR(EINVAL). And it shows "Input contains (near) NaN/+-Inf".    Then I print the information of tmpFrame. Usually, the reason of this bug is that the format of tmpFrame is not AV_SAMPLE_FMT_FLTP. But as it shows in the picture followed, the format is 8, which is the value of AV_SAMPLE_FMT_FLTP. So are there any other reasons causing this bug?


    Sorry to bother you. Thanks for your help.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://ffmpeg.org/pipermail/libav-user/attachments/20210106/6583e28c/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 89A069DF at F0487A68.A1C6F55F.png.jpg
Type: image/jpeg
Size: 3191 bytes
Desc: not available
URL: <https://ffmpeg.org/pipermail/libav-user/attachments/20210106/6583e28c/attachment.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: BF343429 at 8FE61356.A1C6F55F.png.jpg
Type: image/jpeg
Size: 5892 bytes
Desc: not available
URL: <https://ffmpeg.org/pipermail/libav-user/attachments/20210106/6583e28c/attachment-0001.jpg>