similar to: adaptive bandwidth

Hi Kelvin, The audio bandpass setting is only configurable when the encoder is instantiated (eg: start of a Voip call ) , but you can change the bitrate anytime.For example if you can read incoming RTCP packets , you can check if there's reported packet loss , and then lower the bitrate accordingly.Yes, the app has to be aware of the packetloss ?percentage. Cheers,Dragos From: Kelvin

I am using libopus for my implementation. I wonder if anybody in the list have any experience on how to make libopus dynamically adjust its bitrate? On Mar 3, 2015 10:42 PM, "Benjamin Schwartz" <benjamin.m.schwartz at gmail.com> wrote: > It sounds like your software isn't adjusting the opus bitrate in response > to network conditions. For example, many WebRTC

Thanks Dragos, I assume I will be setting those parameters during initialization of encoder right? Question is, if connection gets too lossy, how will opus adapt to it? Can it automatically shift bitrate down to minimize impact? Mark from IRC suggests that the app has to be aware of the losses and change it on the fly. Has anybody on the list tried this? Kelvin Chua On Wed, Mar 4, 2015 at 5:53

Hi Kelvin, You can use something like :opus_encoder_ctl(enc,OPUS_SET_BITRATE(bitrate));opus_encoder_ctl(enc,OPUS_SET_MAX_BANDWIDTH(bandpass)); bandpass is the audio bandpass?, eg: OPUS_BANDWIDTH_WIDEBAND . You will need to calculate the codec bitrate from the available network bitrate (by taking into account the size of the IP+UDP+RTP headers which is 40 bytes ) . By default the audio bandwidth

given opus as a variable bitrate codec applied to voip rtp, i can verify that the bitrate really changes by a few kbps between max and min. as i understood, the bitrate variation is dependent on the audio source. are there any other factors which would affect this varying bitrate? like for example: packet losses, jitter, latency, etc. Will it automatically shift to lower bitrate / sampling rate

having a hard time communicating on IRC, thank you gmaxwell, very informative. anyway, we were discussing the proper implementation of FEC on the decoder side. well, encoder side is just a boolean thing so that's alright. i gave an example where the receiver lost 5 rtp packets, 1 2 3 4 and 5 during which, we call opus_decode with a null pointer and fec=0 for every packet lost. now, when it

hi guys, i'm using sept 30 cvs and oh323 5.5 i'm having 5 second latecy(on only 1 audio path) when a call is transferred.... the scenario is this: sip--------->asterisk----->h323:operator (who then transfers the call) ---------------->h323:destination ------------------audio path 5-second latency---------------->

hi guys, have anybody tried using audiocodes sip fxs against asterisk? how's the device fairing? ~kelvin -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20030725/ae4b2f25/attachment.htm

* compiled from cvs, i am trying call ip phones in callmanager 3.2 10.17.0.2 is my callmanager i noticed from network dumps that instead of sending rtp to the ip phone, * sends it to 10.17.0.2! thereby causing no audio from * to ip phone. audio from ip phone to * is ok. only callmanager calls fail. netmeeting works ok... here is the debug, thanks for any info ~kelvin H323 debug enabled --

hi guys! i'm going to share a workaround for authentication from msn messenger, you have to change two lines in chan_sip.c msn messenger is known to look for the correct realm in authentication, therefore, change the realm in chan_sip.c, line 2061 and line 2910 (release 0.4.0) i hope the realm can be parsed from extensions.conf in the next release... ~kelvin =) -------------- next part

hi guys, i'm encountering one way audio on cvs using netmeeting and chan_h323.so is there a quick fix or workaround for this? compiled using openh323 1.12 pwlib 1.5 i also saw this in earlier version of openh323 and pwlib.... thanks for any info ~kelvin -------------- next part -------------- An HTML attachment was scrubbed... URL:

hi michael, here are the core dumps. only kphone works when 0.5.5 and * cvs. audiocodes and msn messenger all cause seg faults when calling ccm thru * (or vice-versa) ~kelvin [chan_oh323.so] => (OpenH323 Channel Driver) == Parsing '/etc/asterisk/rtp.conf': Found == Parsing '/etc/asterisk/oh323.conf': Found 0:00.004 OpenH323 Wrapper OpenH323 Wrapper

Can I say that RPART is a modified algo of CART and PARTY a modified of CHAID? Thanks. ---- Chua Siang Li Consultant - Operations Research Acceval Pte Ltd Tel: 6297 8740 Email: siang.li.chua at acceval-intl.com Website: www.acceval-intl.com This message and any attachments (the "message"...{{dropped:12}}

hi guys, do you have any suggestions on how to connect 2 TE410P via E1? (for simulation and testing purposes) asterisk1 --> TE410P ----> ? ---------> ? ---->TE410P -->asterisk2 -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.digium.com/pipermail/asterisk-users/attachments/20030907/698cd499/attachment.htm

Hello there. I have fitted a rpart model. > rpartModel <- rpart(y~., data=data.frame(y=y,x=x),method="class", ....) and can use rpart$where to find out the terminal nodes that each observations belongs. Now, I have a set of new data and used predict.rpart which seems to give only the predicted value with no information similar to rpart$where. May I know how

Hello there. I uses the following codes for the purpose of variable selection. > lmModel <- lm(y~.,data.frame(y=y, x=x)) > step <- stepAIC(lmModel, direction="both") > step$anova Stepwise Model Path Analysis of Deviance Table Initial Model: y ~ x.Market.Price + x.Quantity + x.Country + x.Incoterm + x.Channel + x.PaymentTerm Final

Hi. I uses rpart to build a regression tree. Y is continuous. Now, I try to predict on a new set of data. In the new set of data, one of my x (call Incoterm, a factor) has a new level. I wonder why the error below appears as the guide says "For factor predictors, if an observation contains a level not used to grow the tree, it is left at the deepest possible node and

I have a .csv table named mailing.csv as below. It consist a receiver, subject and sender. Receiver subject sender 1 Adrian Cole RE: [WHIRR-117] Composable services Tom White 2 Adrian Cole RE: [WHIRR-117] Composable services Tom White 3 Adrian Cole RE: [WHIRR-117] Composable services Adrian Cole 4 Adrian Cole RE: [WHIRR-117]

Hi. I need some advises on how to use R to find pi (i is the index) with the following objective function and constraint: max (sum i)[ f(ai, bi, pi) * g(ci, di, pi) * Di ] s.t. (sum i)[ f(ai, bi, pi) * Di * pi] / (sum i)[ f(ai, bi, pi) * Di ] <= constant f and g are diffentiable. So, I am thinking of optim with method = "BFGS"? But wonder how to include the

I am not sure what you mean by the ful SIP signaling? Here is the trace for the sip.conf which works successfully. Below that, I will include the trace for the pjsip.conf which it seems Vitelity isn't accepting the ACK in response to the OK ---- SIP --- <--- Transmitting SIP request (1004 bytes) to UDP:64.2.142.93:5060 ---> INVITE sip:8005555555 at 64.2.142.93 SIP/2.0 Via: SIP/2.0/UDP