Compare and contrast (sound quality)

[Pianorak]

Compare and contrast the sound quality of these two broadcasts: Walkuere on Thursday and Otello last night. The former live and the latter recorded. There seemed to be general agreement that the sound quality of the live broadcast was outstanding. I felt that the sound quality of last night's broadcast wasn't nearly as clear and vivid as Thursday's broadcast. If correct, could someone please explain why this difference.

[ferneyhoughgeliebte]

Wagner a better scorer than Verdi?

[Serial_Apologist]

Wagner a better scorer than Verdi?

Verdigris, then...

[ferneyhoughgeliebte]

... apologies: Pianorak had a good, sensible comment, and I've turned it into something silly.

[Pianorak]

... apologies: Pianorak had a good, sensible comment, and I've turned it into something silly.

No problem, ferney! S_A's witty one-liners always make me smile - so all is forgiven!

[Serial_Apologist]

No problem, ferney! S_A's witty one-liners always make me smile - so all is forgiven!

[jayne lee wilson]

Going back to the great days of FM, I always heard some loss of SQ on recordings, "deferred relays" (!) as compared to a live relay. With RFH concerts on FM the difference was often all too obvious in losses of detail, freshness and ambience - evidently losses incurred in the recording process.

Listening to the Tippett Triple Concerto for a second time I didn't think it sounded as clearly 3-D and timbrally lustrous as the live webcast had been. One can think of various reasons why that could be the case, as the audio chain of reproduction is different. I'm always a bit bothered by the "AAC 128kbps" that comes up on listen again, is it still the case that it's really 320 kbps?

Jim Le Surf's Audio Miscellany website (www.audiomisc.co.uk) has gone into great analytical detail on much of this - at one time there was clearly a reduction in dynamic range on the iPlayer listen again service, as opposed to live webcast of the same programme.

[Pianorak]

. . . obvious in losses of detail, freshness and ambience - evidently losses incurred in the recording process. . . . One can think of various reasons why that could be the case, as the audio chain of reproduction is different. . .

Many thanks for the explanation, Jayne - it does seem to be the answer, although I must admit I know nothing of the technicalities of sound recording.

[David-G]

Going back to the great days of FM, I always heard some loss of SQ on recordings, "deferred relays" (!) as compared to a live relay. With RFH concerts on FM the difference was often all too obvious in losses of detail, freshness and ambience - evidently losses incurred in the recording process.

Listening to the Tippett Triple Concerto for a second time I didn't think it sounded as clearly 3-D and timbrally lustrous as the live webcast had been. One can think of various reasons why that could be the case, as the audio chain of reproduction is different.

I don't see why Otello broadcast as a digital stream subsequent to the performance should sound any different from the same digital stream broadcast live. Or am I missing something?

I'm always a bit bothered by the "AAC 128kbps" that comes up on listen again, is it still the case that it's really 320 kbps?

This is worrying. Can anyone tell us definitively if R3 on the iplayer (in "HD" mode) is actually 320 kbps?

[OldTechie]

This is worrying. Can anyone tell us definitively if R3 on the iplayer (in "HD" mode) is actually 320 kbps?

Yes it is definitely 320 kbps despite the fact that it says 128 kbps if you right-click on it. For those with Windows 7 you can look at the data rate easily. Run the task manager. Select the Performance tab and click the Resource Monitor button. Select the Networking tab and under Network Activity you can see the data stream. The source will be a LimeLight server with an llnw.net address.

This is the Otello playing on my machine:



320 k bits per second is 40 k bytes per second. The data rate is slightly higher because there is some overhead in the transmission. It will start with a much higher rate because it buffers around 30 seconds of data in advance.

[Bryn]

Yes it is definitely 320 kbps despite the fact that it says 128 kbps if you right-click on it. For those with Windows 7 you can look at the data rate easily. Run the task manager. Select the Performance tab and click the Resource Monitor button. Select the Networking tab and under Network Activity you can see the data stream. The source will be a LimeLight server with an llnw.net address.

This is the Otello playing on my machine:



320 k bits per second is 40 k bytes per second. The data rate is slightly higher because there is some overhead in the transmission. It will start with a much higher rate because it buffers around 30 seconds of data in advance.

Is 320kbps not in fact a target average aac data rate? I find the rate reported as varying from around 300kbps to around 340kbps from moment to moment.

[David-G]

Thank you OldTechie, that's very helpful. And a great relief! I find that I can display the same data in Vista.

Coming back to Pianorak's original question - I have not yet listened to Otello, but I will try to catch it on the iplayer, and I will be interested to hear what the sound quality is like.

[jayne lee wilson]

I've found the relevant pages (about live/listen again) in Jim LeSurf's analysis. They can be seen at:

(https://www.audiomisc.co.uk/BBC/iPla...sOK/Page2.html)
Scroll down to figs. 13 and 14 for the live/listen again iPlayer analysis itself.

It's worth noting that, while discussion of dynamic range naturally focusses on peak behaviour, even slight changes to the lowest levels (compressed to sound louder) can affect your perception of the detail not just of quiet passages, wind solos etc., but of the acoustic or ambient clues, noises from hall, stage and audience etc., all of which contribute to the sense of realism in the sound, the "living presence" of a musical event. It's often that sense of intimacy - low-level detail, or the lack of it - that distinguishes between average, good and great recordings and systems.

When I heard the live webcast of the Tippett Triple Concerto, I was really quite impressed by the setting of the soloists against the deeper backdrop of the orchestra, and the acoustic feel of the hall signature. On listen again, these characteristics seemed a little dried out and spatially flattened. It certainly sounds like the feed is different, encoded separately. But this is one case where a direct comparison is a little difficult!

[OldTechie]

The relevant pages in Jim LeSurf's analysis can be seen at

(www.audiomisc.co.uk/BBC/iPlayerRulesOK/Page2.html)

That page is from 2009. I think they have improved the Radio 3 iPlayer feeds a lot since then. Not only have they upped the data rate but also they no longer encode at a higher level than DTV so it is not clipped/limited any more.

There is another page of more recent tests he did at http://www.audiomisc.co.uk/BBC/Proms...ulesAgain.html

All very intersting - thanks for introducing me to his web site.

[OldTechie]

Is 320kbps not in fact a target average aac data rate? I find the rate reported as varying from around 300kbps to around 340kbps from moment to moment.

I think they use constant bit rate encoding (which means that the 320 kbps will be an absolute maximum, and if there is not enough data for the specified rate it will be padded out.)

However, the data delivery is not constant. It is fairly steady on the live feed.



For play-it-again it is in short burts every 4 seconds or so:



The orange line is the iPlayer data - the green is the total traffic on my machine.

With this you will get variable measurements if the averaging period is short - the Windows 7 Resource Monitor seems to average over a minute so the reading is pretty steady.