Re: OT: jack_iodelay outside jack
Matt Keys <[email protected]> Tue, 29 Jun 2021 23:36:46 +0000
| Newsgroups | gmane.comp.audio.ardour.user |
|---|---|
| Message-ID | <BN6PR19MB0114EBE8EFD4710A28830140AF029@BN6PR19MB0114.namprd19.prod.outlook.com> |
--===============7957778772144479537== Content-Language: en-US Content-Type: multipart/alternative; boundary="_000_BN6PR19MB0114EBE8EFD4710A28830140AF029BN6PR19MB0114namp_" --_000_BN6PR19MB0114EBE8EFD4710A28830140AF029BN6PR19MB0114namp_ Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Thank you, this is good info. The intention behind the controlled environement was hopes of establishing = an accurate baseline for use use when factoring latency of uncompressed/com= pressed between physically remote clients. SIP/webrtc ride UDP (RTP) simila= r to jacktrip. Opus 48k sample 96kbps I presume is fairly similar to jacktr= ip 48k sample / 128 frames per sec, or perhaps adjustable to match on eithe= r side. They're also very close in terms of troubleshooting tools using thi= ngs like iperf to gauge delay between two sites. ________________________________ From: Ardour-Users <[email protected]> on behalf of Chr= is Caudle <[email protected]> Sent: Sunday, June 27, 2021 11:07 PM To: [email protected] <[email protected]> Subject: Re: [Ardour-Users] OT: jack_iodelay outside jack On Sun, June 27, 2021 8:09 pm, Matt Keys wrote: > I've asked in other forums without success. Have you tried Hydrogen Audio? Those guys have done a lot of listening tests. Have you checked the opus information direct from the web site? This page has a comparison that shows opus at 96k is high quality, but still perceptible: https://opus-codec.org/comparison/ Latency seems to be in the 5ms-20ms range for high bit rate opus. > I'm trying to understand which is better for tx/rx of highest quality and > lowest latency audio to multple remote peers - uncompressed audio How remote? If on a LAN you control then multicast uncompressed is both highest quality and lowest latency. > example jacktrrip 2 channel 48k / 128 frames), or something like opus at > 48k 2 channel (for example webrtc using opus)? The network protocol built into jackd can use opus (well, celt, the immediate precursor to opus); you could start with that, compare uncompressed and compressed using netjack1. https://jackaudio.org/faq/netjack.html > My focus is within a controlled > environment like a flat wired or optical 1g or 10g ethernet. That isn't what I would consider remote. Just use uncompressed audio, anything else is just adding latency and reducing audio quality for no reason. > Does the use of opus (and a sip stack/webrtc) add latency overall Yes > there efficiency gains (for example network bandwidth reduction) through > its use? Why do you care about network efficiency on a 10Gb or even 1Gb network in a flat? You can carry something like 500 channels of uncompressed audio on a 1Gb network. At 48k sampling that is 48k*4bytes =3D 192k bytes/second =3D 1.55-ish Mbits/second per channel. Even with overhead of packet header= s you can get hundreds of channels into 1000Mb/s Ethernet. > Could this difference, if any, be measured using jack_iodelay? Yes, as pointed out above you can just use jackd with netjack1 and switch between uncompressed and compressed audio. But you should only do that for the educational value, if you just want to get audio work done just load up zita-j2n and get going, messing around with opus on a LAN doesn't make sense. -- Chris Caudle _______________________________________________ Ardour-Users mailing list [email protected] To unsubscribe or change your mailing preferences please visit: http://lists.ardour.org/listinfo.cgi/ardour-users-ardour.org --_000_BN6PR19MB0114EBE8EFD4710A28830140AF029BN6PR19MB0114namp_ Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable <html> <head> <meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Diso-8859-= 1"> <style type=3D"text/css" style=3D"display:none;"> P {margin-top:0;margin-bo= ttom:0;} </style> </head> <body dir=3D"ltr"> <div style=3D"font-family: Calibri, Arial, Helvetica, sans-serif; font-size= : 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);"> Thank you, this is good info. </div> <div style=3D"font-family: Calibri, Arial, Helvetica, sans-serif; font-size= : 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);"> <br> </div> <div style=3D"font-family: Calibri, Arial, Helvetica, sans-serif; font-size= : 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);"> The intention behind the controlled environement was hopes of establishing = an accurate baseline for use use when factoring latency of uncompressed/com= pressed between physically remote clients. SIP/webrtc ride UDP (RTP) simila= r to jacktrip. Opus 48k sample 96kbps I presume is fairly similar to jacktrip 48k sample / 128 frames per sec, o= r perhaps adjustable to match on either side. They're also very close in te= rms of troubleshooting tools using things like iperf to gauge delay between= two sites.</div> <div> <div id=3D"appendonsend"></div> <div style=3D"font-family: Calibri, Arial, Helvetica, sans-serif; font-size= : 12pt; color: rgb(0, 0, 0);"> <br> </div> <hr tabindex=3D"-1" style=3D"display:inline-block; width:98%"> <div id=3D"divRplyFwdMsg" dir=3D"ltr"><font face=3D"Calibri, sans-serif" co= lor=3D"#000000" style=3D"font-size: 11pt;"><b>From:</b> Ardour-Users <ar= [email protected]> on behalf of Chris Caudle <6807.= [email protected]><br> <b>Sent:</b> Sunday, June 27, 2021 11:07 PM<br> <b>To:</b> [email protected] <[email protected]&= gt;<br> <b>Subject:</b> Re: [Ardour-Users] OT: jack_iodelay outside jack</font> <div> </div> </div> <div class=3D"BodyFragment"><font size=3D"2"><span style=3D"font-size:11pt"= > <div class=3D"PlainText">On Sun, June 27, 2021 8:09 pm, Matt Keys wrote:<br= > > I've asked in other forums without success.<br> <br> Have you tried Hydrogen Audio? Those guys have done a lot of listenin= g<br> tests.<br> Have you checked the opus information direct from the web site?<br> This page has a comparison that shows opus at 96k is high quality, but<br> still perceptible:<br> <a href=3D"https://opus-codec.org/comparison/">https://opus-codec.org/compa= rison/</a><br> <br> Latency seems to be in the 5ms-20ms range for high bit rate opus.<br> <br> > I'm trying to understand which is better for tx/rx of highest quality = and<br> > lowest latency audio to multple remote peers - uncompressed audio<br> <br> How remote? If on a LAN you control then multicast uncompressed is bo= th<br> highest quality and lowest latency.<br> <br> > example jacktrrip 2 channel 48k / 128 frames), or something like opus = at<br> > 48k 2 channel (for example webrtc using opus)?<br> <br> The network protocol built into jackd can use opus (well, celt, the<br> immediate precursor to opus); you could start with that, compare<br> uncompressed and compressed using netjack1.<br> <a href=3D"https://jackaudio.org/faq/netjack.html">https://jackaudio.org/fa= q/netjack.html</a><br> <br> > My focus is within a controlled<br> > environment like a flat wired or optical 1g or 10g ethernet.<br> <br> That isn't what I would consider remote. Just use uncompressed audio,= <br> anything else is just adding latency and reducing audio quality for no<br> reason.<br> <br> > Does the use of opus (and a sip stack/webrtc) add latency overall<br> <br> Yes<br> <br> > there efficiency gains (for example network bandwidth reduction) throu= gh<br> > its use?<br> <br> Why do you care about network efficiency on a 10Gb or even 1Gb network in<b= r> a flat? You can carry something like 500 channels of uncompressed aud= io<br> on a 1Gb network. At 48k sampling that is 48k*4bytes =3D 192k bytes/s= econd<br> =3D 1.55-ish Mbits/second per channel. Even with overhead of packet h= eaders<br> you can get hundreds of channels into 1000Mb/s Ethernet.<br> <br> > Could this difference, if any, be measured using jack_iodelay?<br> <br> Yes, as pointed out above you can just use jackd with netjack1 and switch<b= r> between uncompressed and compressed audio.<br> But you should only do that for the educational value, if you just want to<= br> get audio work done just load up zita-j2n and get going, messing around<br> with opus on a LAN doesn't make sense.<br> <br> -- <br> Chris Caudle<br> <br> <br> _______________________________________________<br> Ardour-Users mailing list<br> [email protected]<br> To unsubscribe or change your mailing preferences please visit:<br> <a href=3D"http://lists.ardour.org/listinfo.cgi/ardour-users-ardour.org">ht= tp://lists.ardour.org/listinfo.cgi/ardour-users-ardour.org</a><br> </div> </span></font></div> </div> </body> </html> --_000_BN6PR19MB0114EBE8EFD4710A28830140AF029BN6PR19MB0114namp_-- --===============7957778772144479537== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Ardour-Users mailing list [email protected] To unsubscribe or change your mailing preferences please visit: http://lists.ardour.org/listinfo.cgi/ardour-users-ardour.org --===============7957778772144479537==--