Hi,
Could someone give me an idea for which case onaddstream(event) would pass no
tracks at all ?
event.stream.getVideoTracks().length and event.stream.getAudioTracks().length
are = 0.
ICE, DTLS is OK for audio, RTP traffic (G.722) flows in both directions. Same
code used for Chrome is OK and
If you are family with C++ you could retrieve needed libs (webrtc, nicer etc.)
from FireFox or Chromium and build your own recording server. Or use WebRTC
Native code. If you prefer Java take Jitsi Video Bridge and hack it.
___
dev-media mailing list
de
> I'd like to know if it's possible to force Firefox Hello to use H.264 & G.722
> only.
To use H.264 & G.722 only you could modify remote SDP rtpmap, deleting
undesirable codec lines.
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists
Visual Studio (2013) debugger doesn't stop at breakpoints placed in some
modules like WebRTC and nICEr. As I googled, the reason of this depends on the
fact that xul.dll, contained these modules, "did not load at the default load
address" (as it's hinted in the module list).
Could somebody help
Attaching files is restriched by corporate security rules.
Could you hint me if Firefox supports separate (without BUNDLEing) media tracks
? Must they be attached to the same MediaStream (MID) or not ?
___
dev-media mailing list
dev-media@lists.mozilla.o
Thanks a lot !
After turning off the multi-processing that began to work.
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
I can't provide a service for reproducing that. But I encovered a bug in
RTCPeerConnection.onaddstream(event) : event.stream.getVideoTracks() and
event.stream.getAudioTracks() are empty while event.stream.getTracks() is not
and includes audio or video tracks.
Here's the code :
//--
If I add an extra m-line to remote descrition, I'm getting a "Wrong SDP: Offer
and answer have different number of m-lines (2 vs 3)" error. I don't understand
why must they be the same size. If I'm sending a single video stream what
should restrict me from receiving two ones ?
Actually the thir
Thanks, Eric,
> If Chrome is allowing the answerer to add a third m= line, then it's
> totally not
> complying with RFC 3264 and that's a defect in Chrome.
I didn't try to do add m-lines for Chrome. Because of Plan B I used an
additional ssrc line for that purpose. In the case of Unified Plan I
Hi All,
Is it possible to change encoding bitrate for video (VP8) in Firefox ?
Currently it'a about 300 kb/s, and I'd like to increase it.
For Chrome I'm using such an attribute to make it encode at 500 kb/s :
a=fmtp:100 x-google-start-bitrate=128; x-google-min-bitrate=380;
x-google-max-bitrat
I see a video with no lags and good FPS but not enough sharp and a little dirty.
Here Chrome stats graph (other side os Firefox) :
https://drive.google.com/file/d/0B6N7bihGDAcJTHpYd1lwaFhCcy1CRUthWE5NR2ZCc1ZVUW04/view?usp=docslist_api
Firefox statistics shows :
outbound_rtp_audio_0
Local: 1
That means that there's not manual API controls for setting the needed bitrates
in FireFox, right ? Only an automatic one. That's not good.
My setup is not p2p as AppRtc, it's a Chrome Client - Conferencing Server -
Firefox Client. Remote SDPs are created programmaticaly. Maybe really something
1.4 mbps is also sad. Too much load for the server. For Chrome I decreased it
from 2 to 0.5 - picture quality at 0.5 is good enough.
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
This question is mostly intended for Eric as being DTLS guru.
I have a case when in the end of DTLS handshaking between our server and
Firefox, the server sends Encryption Alert (this name I see in Wireshark), and
after receiving that the Firefox closes corresponding media pipeline. I
couldn't
Thanks, Eric.
Actually that's not me, it is an inherited code. It was working for other
applicaitons. You mean that shutdown should be called after whole session ends
but not only DTLS part.
___
dev-media mailing list
dev-media@lists.mozilla.org
https:
Hi,
Does Firefox supports multple BUNDLEs ? I have the following case. Three WebRTC
peers are connected via media server.
Audio and video multistreamed in two channels : one for all peers audio, second
- for all the video.
I have two nominated ICE connection, oprt 8000 for audio, 9000 - video.
R
Here's full two-pass log.
1) First Firefox entered a conference with a single peer. Firefox is offer,
conference server is answer. Audio and video are OK at both sides
ICE Stats
Local Candidate Remote Candidate
ICE State Prio
I also tried to set fixed ports to the last ("offer") remote SDP m-lines as
8000 and 9000 instead of 1 to help it be attached to correct ICE nominant. But
nothing changes.
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/
as in use. There's a bug to fix there.
>
> Best regards,
> Byron Campen
>
> On 1/27/16 6:03 AM, Alexander Abagian wrote:
> > Here's full two-pass log.
> >
> >
> > 1) First Firefox entered a conference with a single peer. Firefox is offer,
> > c
ly including a candidate that it doesn't end up using, I am not
> sure, but it is moot since bundle is in use (meaning that the transport
> info from the first audio m-section is authoritative). It is still a bug
> though.
>
> Best regards,
> Byron Campen
>
> On
Does Firefox support partial offers and answers ?
I encovered a problem with more than two participants in Firefox client.
The client works in the following manner :
1)
setLocalDescription [offer]
a=group:BUNDLE sdparta_0 sdparta_1
...
m=audio
a=mid:sdparta_0
...
m=video
a=mid:sdparta_1
...
se
Thank you for the answer.
I've changed the order of m-sections from all-audio-first (I've done it so
because have seen this requirement it in some RFC) to audio-video pairing with
adding a new m-secions to the end of the SDP, and it helps.
But anyway when I'm removing a participant, I'm receiv
Hi,
I have the following problem with RTCP in more than two participants conference.
I encodered that the media server rejects some RTCP packet when Firefox
entering. The reason is in the fact that Firefox uses different RTCP Sender
SSRC for different participants.
Is it possible to force Firefo
ry 22, 2016 at 5:12:17 PM UTC+3, Byron Campen wrote:
> On 2/22/16 8:00 AM, Alexander Abagian wrote:
> > Is it possible to force Firefox to use the same SSRC of the same media type
> > for each participant ?
> No, because if we did that we would be breaking spec pretty badly;
&
Hi,
I've got the following warning in FF 45 :
(ice/ERR) peer received no media stream attributes
(ice/ERR) peer specified too many components
(ice/WARNING) specified bogus candidate
(ice/WARNING) Error parsing attribute: candidate:1316968211 1 UDP 2130706430
91.224.14.66 8001 typ srflx generatio
Would it be OK if I use 0.0.0.0 as a raddr ?
On Thursday, March 17, 2016 at 8:31:57 PM UTC+3, Byron Campen wrote:
> On 3/17/16 12:18 PM, Alexander Abagian wrote:
> > Hi,
> >
> > I've got the following warning in FF 45 :
> >
> > (ice/ERR) peer received no medi
Hi,
I have the following case.
There's a conference with a three participants. One of them quits, and after
that ICE connection becomes broken.
Each m-section uses its own mid, audio and video are multistreamed in two
transport channels.
The order is always SetRemoteDescription["offer"] /
Set
Hi,
I have the following case.
There's a conference with a three participants. One of them quits, and after
that ICE connection becomes broken.
Each m-section uses its own mid, audio and video are multistreamed in two
transport channels.
The order is always SetRemoteDescription["offer"] /
Set
No, it is not. I just tried to set it to false, but nothing changes.
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
Hi,
I encovered that after m-sections count is greater than ~(25 + 25) for audio
and video, the addstream callback is no more called when new section are added.
The case is : a conference with a three participants, on of them (X) is
re-entering several times. M-sections count increases : old X
Hi all,
In random cases, when Firefox webrtc client connects to the media server, I'm
getting a DTLS handshake for the video channel (only) fails with DTLS v1.2
alert "illegal parameter". What is strange, that this alert is the first DTLS
message in the whole DTLS handshake session; and never a
Alert in case it is the DTLS server and never receives a Client Hello.
>
> Best regards
> Nils Ohlmeier
>
> > On Aug 31, 2016, at 11:58, Alexander Abagian wrote:
> >
> > Hi all,
> >
> > In random cases, when Firefox webrtc client connects to the medi
Firefox version is the latest stable 48.0.2.
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
11 PM UTC+3, Nils Ohlmeier wrote:
> > On Sep 1, 2016, at 04:51, Alexander Abagian wrote:
> >
> > Here's webrtc-internals. The pcap is almost the same, the only difference
> > is some 5-digit ports. Firefox ip is 192.168.125.39. Media server
> > (192.168.125.13
Hi all,
Trying to make a workaround for TIAS bandwidth control problem, I was using
degradationPreference for the resolution to be fixed. But there were no effect
- resolution still was changing in big range, from 1280x720 to 352x288.
I was calling this from setLocalDescription success callback
Nils,
It's about bug #1308481 (https://bugzilla.mozilla.org/show_bug.cgi?id=1308481).
Randell,
I just tried the latest 53.01a x64 Nightly. CIF is really dropped, but 640x480
now has the same bad behavior as the CIF had : influence to increasing the
bitrate to 2000 kBps.
_
Hi All,
We've made a WebRTC load testing js app, starting a set of WebRTC clients in a
single page. Each of them has it's own RTCPeerConnection and everything other.
When it starts from Firefox, then only 15 of these clients become connected.
Any extra clients are dead.
It looks like there i
Here's webrtc-internals : https://yadi.sk/d/89E5gz71zKUEg
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
really working. Making the Frankenstein myself is too
big task for just a some load testing tool.
I guess I have to migrate to Chromium with this task.
On Thursday, November 24, 2016 at 12:28:27 AM UTC+3, Nils Ohlmeier wrote:
> Hi Alexander,
>
> > On Nov 23, 2016, at 08:55, Alexa
Hi Mozilla guys,
I encovered that incoming audio traffic packets has a SSRC than is not used in
remote SDP, but anyway the sound is audible. Is it a bug ?
See ssrc == 2891593250 below:
Remote SDP
v=0
o=- 392585836 2 IN IP4 127.0.0.1
s=-
t=0 0
a=sendrecv
a=group:BUNDLE audio-1234567890
a=group
here's complete log
https://yadi.sk/d/ehbdA8Vy3G6nNr
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
Nils,
Both of them are Firefox texts, but the attached one is a bad try, my
apologizes.
The first one looks unfamiliar because the sdp was created by the script but
not the WebRTC engine.
It illustrates that there's no inboundrtp SSRC, which is successfully receiving
and producing sound, in a
Hi,
In Firefox, answer SDP rtpmap always includes no more than a one codec
descrption. Why doesn't it lists all corresponding codecs as Chrome does ?
Is there a way to overcome it without re-negotiation ?
___
dev-media mailing list
dev-media@lists.mozi
Bad news.
But is it available to have different codecs for different m-sections of the
same type ?
I.e.
Remote SDP (offer) contains
m=audio
a=rtpmap:8 PCMA/8000
a=mid:audio-remote-1
m=audio
a=rtpmap:9 G722/8000
a=mid:audio-remote-2
And the local answer SDP from createAnswer() would be simila
I guess, Firefox has some problems in codec negotiating. See funny sdps below.
The fun is that remote SDP has rtpmaps limited in PCMA and PCMU only, but local
SDP, produced by createAnswer(), has G722 against them both.
Local SDP (answer)
=
v=0
o=mozilla...THIS_IS_SDPARTA-52.0.1 3673
What's about bug 1342727, 1332031 ?
https://bugzilla.mozilla.org/show_bug.cgi?id=1342727
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
Hi,
I've got a case where a local Firefox candidate has not got a remote paired
candidate, and the ICE testing procedure has not started for it. It's
94.25.182.143:1368 in this list.
The analogous candidate (94.25.182.143:10973) used for audio transmission has
succeeded, but this one (video) e
> On 4/17/17 12:40 PM, Alexander Abagian wrote:
> > Hi,
> >
> > I've got a case where a local Firefox candidate has not got a remote paired
> > candidate, and the ICE testing procedure has not started for it. It's
> > 94.25.182.143:1368 in this list.
>
Hi,
Is there any news about configuring Firefox webrtc local media/data port range ?
Many admins wish to limit this range.
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
At least, what is the current port limit if any ?
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
I guess server candidates 212.24.35.119:8001 and 212.24.35.119:8001 type was
mistakenly defined by the server as "host" and this could confuse ICE engine.
But anyway in this case checking state should not be kept "inprogress" for such
a long period (30 min or more).
_
Hi all,
I've got a about:webrtc log for a quite big webrtc conference with a lot of
lines in like these:
(ice/WARNING) ICE(PC:1498488600251000 (id=47
url=https://go.muppetshow.com/service/wconference)): peer (PC:1498488600251000
(id=47 url=https://go.muppetshow.com/service/wconference):default
Yes, other participants were chrome webrtc and proprietary clients. Usually
that works.
109.94.22.255:56154/udp(serverreflexive)212.24.35.119:8004/udp(host)
succeeded 288226015837044220 truetrue
109.94.22.255:57539/udp(serverreflexive)212.24.35.119:9004/udp(host
Hi all,
Is it possible to turn off audio-video synchronizing ?
Trying to solve no-video bug I've found log messages like
2017-07-03 20:36:08.682000 UTC - [Main Thread]: D/signaling
[main|PeerConnectionMedia] PeerConnectionMedia.cpp:1615: Syncing 19D04C00 to
204E33E0, re.com-231471651 to re.com
Hi,
Is there an internal limit for video channel or m-sections count in the
RTCPeerConnection ? I have a case when the seventh (usually) participant video
in a WebRTC conference client is not shown. Sometimes it could be 8, and if
they were added very quickly by special load client, then it's 1
Yes, the server works as SFU. The browser sends single audio and singe video
stream to it, and receives multiple streams from the server.
We usually use this conference system for our own meetings with ~20-25 people
and this is OK for common core i5 PC.
__
Thank you,
I've submitted #1402510 with an ability to test it. Will attach the logs soon.
On Friday, September 22, 2017 at 9:33:50 PM UTC+3, Randell Jesup wrote:
> On 9/22/2017 2:21 PM, Alexander Abagian wrote:
> > Yes, the server works as SFU. The browser sends single audio and si
On Friday, September 22, 2017 at 9:33:50 PM UTC+3, Randell Jesup wrote:
> On 9/22/2017 2:21 PM, Alexander Abagian wrote:
> > Yes, the server works as SFU. The browser sends single audio and singe
> > video stream to it, and receives multiple streams from the server.
> >
&g
Hi,
Does anybody know if multiple spatial or temporal layers are enabled in Firefox
VP9 implementation ?
Thanks,
Alex
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
I see in the code that it could be enabled by the command line flag similar to
Chrome. The question rest is how to be sure that multiple layers are turned on ?
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-
Hi,
I want to make Firefox stream both screen sharing and common webcam video
simultaneously. Is it possible at all ? Chrome lets do it with an external
screen sharing plugin, but here both of the should be produced by getUserMedia
and I haven't find a way how to make it. mediaDevices.enumerate
Thanks, that works. getUserMedia appears to be reenterable.
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
Hi,
I encovered that if RTCRtpSender.track.readyState === "ended" then after
RTCRtpSender.replaceTrack(newTrack) this state stays "ended" despite
newTrack.readyState is "live". Is it a correct behavior ?
___
dev-media mailing list
dev-media@lists.mozil
Here's test function (sorry for C++ style). "senders" assumed to contain a
single sender with a single stream.
After replaceTrack() sender.track.readyState stays "ended" while
newTrack.readyState is "live".
var ReplaceTrack = function(senders_, newTracks_) {
var wasReplaced = false;
Hi,
I'm faced with a UI problem concerning screen sharing dialog in FF 59. This
dialog offers two options : "Allow" and "Don't Allow" and a checkbox Remember
this decision.
First of all, the checkbox is useless, if checked, "Allow" is disabled and
"Firefox can not allow permanent access to you
Hi,
I'm faced with a UI problem concerning screen sharing dialog in FF 59. This
dialog offers two options : "Allow" and "Don't Allow" and a checkbox Remember
this decision.
First of all, the checkbox is useless, if checked, "Allow" is disabled and
"Firefox can not allow permanent access to y
Hi,
Is there any support for QoS in Firefox WebRTC now ?
Google has googDscp now and encodings[].priority setting ability in upcoming
release.
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
Hi,
I know that Firefox does not have such a useful thing as
-disable-webrtc-encryption. Could somebody recommend a way to modify Firefox
sources to build a custom build with SRTP off ? Best of all if DTLS would stay
on but SRTP does not encrypt RTP. I need to investigate H.264 in Wireshark.
T
ct. Escpecially, if it is a special custom build.
On Friday, March 1, 2019 at 11:23:22 PM UTC+3, Nils Ohlmeier wrote:
> Hi Alexander,
>
> > On 1Mar, 2019, at 09:00, Alexander Abagian <> wrote:
> > I know that Firefox does not have such a useful thing as
> > -di
DTLS - for testing purpose, it'll be better if the server and the client use
the same code blocks as it is usual.
___
dev-media mailing list
dev-media@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-media
70 matches
Mail list logo