30 replies to this topic

[Robinson]

Very promising.

[Erik Slagter]

 

you'd need to "tap" the demuxer,

Or the manufacturer need to add something in their drivers so we can retrieve this information...

Agree, that would be even better. But we'd never get ALL of the manufacturers to come with the same api...

[littlesat]

Yep.... and using external tools is a work-a-round... brrrrr.... To get the provider ID from ECMs could also be done via DVBsnoop... but I choose a different way...  And with DVB snoop you do not always get the played stream. It shows the first recorded, streamed or played stream.... so no guarantee to get the proper data.

Edited by littlesat, 10 June 2017 - 09:41.

[athoik]

External tools is the unix way and it works fine.

Also what you are saying is not what I describe in my post earlier.

Just like we do with bitrate plugin... we can get the audio PID and pass the audio TS stream for analysis in a software like ffprobe.

What is the way you are proposing other than my work-a-round?

[jpuigs]

I've been playing with ffprobe, and I see it doesn't difference audio (mp2) in Stereo or Joint Stereo modes.

It always says "stereo".

Edited by jpuigs, 10 June 2017 - 23:59.

[Erik Slagter]

External tools is the unix way and it works fine.

Also what you are saying is not what I describe in my post earlier.

Just like we do with bitrate plugin... we can get the audio PID and pass the audio TS stream for analysis in a software like ffprobe.

What is the way you are proposing other than my work-a-round?

The neatest way would be to include the ffmpeg libraries in enigma and then you can probe withouth using external applications. BUT it is complex and difficult.

[Erik Slagter]

I've been playing with ffprobe, and I see it doesn't difference audio (mp2) in Stereo or Joint Stereo modes.

It always says "stereo".

That is because both ARE stereo

 

Why do you want to know? Most (all?) audio encoders can switch between "full stereo" and "joint stereo" on a frame-by-frame basis, e.g. audio consisting of only one voice "in the middle" will be encoded using joins stereo and as soon as the music starts, the encoder will switch to full stereo.

 

Why do you want to know, really? Do you think you can hear the difference?

[jpuigs]

Yes.

Some broadcasters (most of them radiostations) which use low bitrate values, 96 kb/s, 112 kb/s, 128 kb/s , are transmitted using joint stereo instead of stereo, because it gives better sound quality than real stereo (AT LOW BITRATES) , although it loses some stereo separation in certain frequencies.

Mpeg-1 layer 2 doesn't switch between Joint and stereo, at least I've never seen it.

[Erik Slagter]

I've seen it (in music clips).

 

I doesn't give better sound quality, it gives better compression rates, because only one channel needs to be encoded, plus the difference between left and right, which, when jount stereo is used, should be minimal. At low (CBR) bit rates, it will result in better quality. It will not result in less stereo separation, though. When both channels are 100% different (theoretical case), they will still be encoded as two 100% different channels, just through a matrix (left, right) -> (center, difference). The center  channel will yield very little information in this case and the difference channel will yield a huge amount of information. This will result in very inefficient compression, so when this is the case, it's really better to use two-independent-channel-stereo. But it has nothing to do with stereo separation.

 

Which is another situation to FM stereo, where also a center and a difference channel is broadcast, but where the difference channel is turned down deliberately, to minimise bandwidth, keeping the whole FM channel within the allotted 75 kHz boundaries. AFAIK digital audio encoders do not practise this.

 

I am still curious why you'd want to know this on a per-channel basis, anyway.

[jpuigs]

AFAIK, on MP2 (Mpeg 1 Layer 2) audio signals, the joint stereo mode used is the intensive coding mode.

 

The other joint mode, the Mid/Side (M/S) mode is used on MP3, AAC.....  and is similar to FM analog stereo. Mid = L+R = Mono  , Side= L-R

Edited by jpuigs, 11 June 2017 - 13:52.

[Erik Slagter]

Yes, you are right, apparently I was referring to mid/side stereo encoding.

 

I don't think even ffprobe will give this information, let alone drivers or hardware...