ted in mkv, opus, and RTP stream.
In the last case, packets can be dropped with the tc command.
Thanks a lot Steiner for your original work, it helped a lot.
Thanks for your time.
Philip-Dylan Gleonec (1):
avcodec/libopusdec: Enable FEC/PLC
libavcodec/libopusd
t reconstruct, i.e. packets received twice and
packets with a length that is not a multiple of 2.5ms.
Signed-off-by: Philip-Dylan Gleonec
Co-developed-by: Steinar H. Gunderson
---
libavcodec/libopusdec.c | 105 +++-
1 file changed, 94 insertions(+), 11 delet
-Dylan Gleonec
---
libavcodec/libopusenc.c | 9 +
1 file changed, 9 insertions(+)
diff --git a/libavcodec/libopusenc.c b/libavcodec/libopusenc.c
index 70d17f802b..c18e8ae7fa 100644
--- a/libavcodec/libopusenc.c
+++ b/libavcodec/libopusenc.c
@@ -460,6 +460,15 @@ static int libopus_encode
> Could you elaborate?
> I would have expected that the normal use case is not have a
> lossy input and that the new feature is always useful if data
> was lost.
The use-case for FEC is typically RTP stream where audio is compressed
with opus. In that case, depending on the network conditions, pac
>> Am Di., 16. Feb. 2021 um 15:02 Uhr schrieb Philip-Dylan Gleonec
>> :
>>
>>>
>>> Adds FEC/PLC support to libopus. The lost packets are detected as a
>>> discontinuity in the audio stream. When a discontinuity is used, this
>>> patch tries
> I've added him to the CC list
Now done.
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
ion of
>> 120ms, a threshold can find how the pts is expressed.
>
>There is AVCodecContext.pkt_timebase.
Thanks for the feedback, I had missed that property.
I've attached a reworked version of the patch which makes use of it. It
is indeed more clean and precise. I've left a
Here is the reworked patch properly attached.
Sorry about the duplicate mail, I just noticed I had a mishap with my
mail client and the previous patch was scrubbed.
Signed-off-by: Philip-Dylan Gleonec
Co-developed-by: Steinar H. Gunderson
---
libavcodec/libopusdec.c | 107
Le 16/02/2021 à 15:04, Philip-Dylan Gleonec a écrit :
An estimation of packet loss is required by libopus to compute its FEC
data. Currently, this estimation is constant, and can not be changed
after configuration. This means an application using libopus through
ffmpeg can not adapt the packet
Le 18/02/2021 à 17:38, Philip-Dylan Gleonec a écrit :
Here is the reworked patch properly attached.
Sorry about the duplicate mail, I just noticed I had a mishap with my
mail client and the previous patch was scrubbed.
Hello,
Is someone interested in picking this up, or is there some further
Hello,
Please find attached a rebased patchset for the FEC implementation of
libopus. Following the received feedbacks, some improvements have been
done compared to the first version:
- remove a log when a packet is decoded without FEC
- add a check to only set libopus encoder packet loss estimat
packet loss can be estimated from RTCP packets.
Signed-off-by: Philip-Dylan Gleonec
---
libavcodec/libopusenc.c | 17 +
1 file changed, 17 insertions(+)
diff --git a/libavcodec/libopusenc.c b/libavcodec/libopusenc.c
index 45b23fcbb5..b9e2fc45e3 100644
--- a/libavcodec/libopusenc.c
t reconstruct, i.e. packets received twice and
packets with a length that is not a multiple of 2.5ms.
Signed-off-by: Philip-Dylan Gleonec
Co-developed-by: Steinar H. Gunderson
---
libavcodec/libopusdec.c | 105 +++-
1 file changed, 94 insertions(+), 11 delet
Hello,
Is there some interest in this patch ? If so, is there something I can
modify to improve it ?
Regards,
Philip-Dylan Gleonec
___
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe
Hello,
Please find attached a new version of the patchset, with the required
corrections. I also added the following changes:
- remove use of avc->channels (deprecated) in favor of avc->ch_layout
- rebase on master
The patches have been tested against FATE, and validated in use on a
rtp stream wi
packet loss can be estimated from RTCP packets.
Signed-off-by: Philip-Dylan Gleonec
---
libavcodec/libopusenc.c | 17 +
1 file changed, 17 insertions(+)
diff --git a/libavcodec/libopusenc.c b/libavcodec/libopusenc.c
index c884075ffe..26d2082ffa 100644
--- a/libavcodec/libopusenc.c
t reconstruct, i.e. packets received twice and
packets with a length that is not a multiple of 2.5ms.
Signed-off-by: Philip-Dylan Gleonec
Co-developed-by: Steinar H. Gunderson
---
libavcodec/libopusdec.c | 105 +++-
1 file changed, 94 insertions(+), 11 delet
17 matches
Mail list logo
