Re: [Linphone-developers] Why does Opus sound so bad on mobile

[Tom Hartnett]

Thanks Steve,

I had actually just determined this is the case under Android. The encoder is sending around 12 Kb/s and it sure does sound like it. Of course under Android there's no way to control this so it needs to be fixed in software.

Under iOS the encoder is indeed sending out 64Kb/s and the problem seems to be some kind of high-pass filter applied (other than the lack of lows it actually sounds fine). I know there are some config options for Opus and my theory is one of these is mis-set under iOS.

Date: Thu, 5 Jun 2014 09:42:31 -0600
From: Steve Strobel <address@hidden>
To: address@hidden
Subject: Re: [Linphone-developers] Why does Opus sound so bad on
        mobile
Message-ID:
        <CAGrmWbxcKCXtOOdbf=cVdSo=address@hidden>
Content-Type: text/plain; charset="utf-8"

Opus supports a very wide range of bit rates.  Maybe the Linphone
implementations are using a much lower bit rate than your lab device.

Steve


On Fri, May 30, 2014 at 3:24 PM, Tom Hartnett <address@hidden>
wrote:

> Greetings,
> I've run some audio quality experiments and I'd appreciate some input.
>  These cuts are best listened via headphones. All cuts are encoded at
> 64Kb/s mono. Audio was fed into phones via headset port.
> Here's a reference of the source I used:
> http://54.201.208.1/opustests/reference.mp3
>
> I then ran a test on Android via a G.722 (non Linphone) app:( The Linphone
> G.722 strangely samples at 8KHz so was unusable)
> http://54.201.208.1/opustests/G.722android.mp3
>
> as you can hear, not perfect, but normal bandwidth for G.722
>
> I then ran LinPhone Opus on the same phone:
> http://54.201.208.1/opustests/opusandroid.mp3
>
> Many codec artifacts are present, inferior to G.722. I should note that
> echo canceler was off.
>
> I moved to iPhone and ran a G.722 app
> http://54.201.208.1/opustests/G.722iphone.mp3
>
> not bad
>
> Then to iPhone/LinPhone/Opus
> http://54.201.208.1/opustests/iphoneopus.mp3
>
> Not artifact heavy, but all low frequencies have been cut. Very tinny
> sound.
>
> Finally, for reference, I did an encode/decode pass on one of our lab
> devices that uses Opus (no Linphone)
> http://54.201.208.1/opustests/comrexopuslb.mp3
>
> Summary: Neither mobile implementation of Opus is superior to old
> fashioned G.722. Other implementations sound *much* better.
> We're trying to recommend Linphone mobile for broadcast purposes but with
> an inferior Opus implementation that's difficult. Any thoughts?
>
>
> _______________________________________________
> Linphone-developers mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/linphone-developers
>
>


--
Steve Strobel
Link Communications, Inc.
1035 Cerise Rd
Billings, MT 59101-7378
(406) 245-5002 ext 102
(406) 245-4889 (fax)
WWW: http://www.link-comm.com
MailTo:address@hidden
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.nongnu.org/archive/html/linphone-developers/attachments/20140605/7598aea5/attachment.html>

------------------------------