Dynamic Range Compression/Fader Fiddling on iPlayer/Sounds

[Another Howard]

Hi Jayne. Thankyou for your response, and the tip. I shall get some squidgy plugs together and audition that AAC feed on the next proper concert. I do wonder if an iPhone 10 has adequate audio out?
Today DAB is playing here from a Cambridge tuner, for first time in years: startlingly dynamic just as FM could be in pre-Optibodge days on the wonderful Wrotham transmitter.
But... however... I do hear something untidy in DAB texture: slight hard edges to vibratoing soprano, as though the virtual dissonances with hall reverberation confuse the encoding! Used to notice something similar on FM when BBC first generation digital distribution links went in - clogged dissonances.
Anyway, want to enjoy the music and somehow reduce habitual vigilance for artefacts! Will report back on AAC in a while.

[Ein Heldenleben]

Hi Jayne. Thankyou for your response, and the tip. I shall get some squidgy plugs together and audition that AAC feed on the next proper concert. I do wonder if an iPhone 10 has adequate audio out?
Today DAB is playing here from a Cambridge tuner, for first time in years: startlingly dynamic just as FM could be in pre-Optibodge days on the wonderful Wrotham transmitter.
But... however... I do hear something untidy in DAB texture: slight hard edges to vibratoing soprano, as though the virtual dissonances with hall reverberation confuse the encoding! Used to notice something similar on FM when BBC first generation digital distribution links went in - clogged dissonances.
Anyway, want to enjoy the music and somehow reduce habitual vigilance for artefacts! Will report back on AAC in a while.

Blimey you’ve got a good pair of ears. On the iPhone front I use the mini dac that plugs into a lightning socket and has a mini Jack at the other end - then copper cable to the amp. There is no headphone socket on my iPhone . I’ve done comparisons on my iPad which does but it’s very difficult to make comparisons unless they are instantaneous because the ear/brain compensates so quickly . Something that hi - fi shops have exploited throughout the ages..
Any way the upshot of my comparison is that the mini dac sounds better out of the iPad than the headphone socket.
If it’s anything like the video quality the audio output is likely to be excellent . iPhone 10’s are regularly used now on broadcast tv and , on the screen, very experienced editors tell me they can’t tell it from pics gathered on cameras costing ten times as much,

[jayne lee wilson]

Hi Jayne. Thankyou for your response, and the tip. I shall get some squidgy plugs together and audition that AAC feed on the next proper concert. I do wonder if an iPhone 10 has adequate audio out?
Today DAB is playing here from a Cambridge tuner, for first time in years: startlingly dynamic just as FM could be in pre-Optibodge days on the wonderful Wrotham transmitter.
But... however... I do hear something untidy in DAB texture: slight hard edges to vibratoing soprano, as though the virtual dissonances with hall reverberation confuse the encoding! Used to notice something similar on FM when BBC first generation digital distribution links went in - clogged dissonances.
Anyway, want to enjoy the music and somehow reduce habitual vigilance for artefacts! Will report back on AAC in a while.

I'm afraid this DAB edginess is inescapable, because the max 192 kbps mpeg2 codec for DAB (whether you have it off satellite, Freeview etc) was never enough for classical, especially strings and any HF generally.

Oh, I tried my best with it. I bought the first model available, the Arcam Alpha 10. Roof aerial, viterbi error rate: 0 or 1....
The vast dynamic range (compared to Optimod-FM) could give you a thrill with something like Janacek's Glagolitic Mass, but a Bruckner slow movement with its string counterpoints..... all but unbearable; distortion, as I put it to a friend with the Technics DAB/FM Model "where the edges rub".
(It was OK for him - he'd won his in a BBC prize draw....)...

I once compared the Arcam Alpha 10 with the Quad FM4.... on a Hindemith Symphony, these problems were starkly clear. The DAB, huge impact but often rough-edged and hard on the ears; the Quad, tonally lovely, 3D stereo, but with scarcely a climax in sight....

Even a solo violin sonata, in direct compare with FM, revealed those edgy colorations. Some relays from the RFH etc could be appallingly edgy and thin. ON TTN, the "cascading" effect (one codec from Euro Radio stations played through the DAB here) could sound truly painful (I spoke to BBC Engineering about it and they admitted the problem and said "we're working on it"...)

The original DAB specs were supposed to run at 256 kbps for FM-or-better stereo sound (I did hear this off a TV feed and it was far smoother), but..... Ofcom insisted on lowering it to 192 to allow for "greater diversity" i.e more stations. So sadly, the better your system, the more you hear the inherent HF distortions....

***
So do please make the effort with the AAC 320. If you can process it through an external DAC (USB or coax cable etc) you and your ears will be well rewarded....

[Ein Heldenleben]

I'm afraid this DAB edginess is inescapable, because the max 192 kbps mpeg2 codec for DAB (whether you have it off satellite, Freeview etc) was never enough for classical, especially strings and any HF generally.

Oh, I tried my best with it. I bought the first model available, the Arcam Alpha 10. Roof aerial, viterbi error rate: 0 or 1....
The vast dynamic range (compared to Optimod-FM) could give you a thrill with something like Janacek's Glagolitic Mass, but a Bruckner slow movement with its string counterpoints..... all but unbearable; distortion, as I put it to a friend with the Technics DAB/FM Model "where the edges rub".
(It was OK for him - he'd won his in a BBC prize draw....)...

I once compared the Arcam Alpha 10 with the Quad FM4.... on a Hindemith Symphony, these problems were starkly clear. The DAB, huge impact but often rough-edged and hard on the ears; the Quad, tonally lovely, 3D stereo, but with scarcely a climax in sight....

Even a solo violin sonata, in direct compare with FM, revealed those edgy colorations. Some relays from the RFH etc could be appallingly edgy and thin. ON TTN, the "cascading" effect (one codec from Euro Radio stations played through the DAB here) could sound truly painful (I spoke to BBC Engineering about it and they admitted the problem and said "we're working on it"...)

The original DAB specs were supposed to run at 256 kbps for FM-or-better stereo sound (I did hear this off a TV feed and it was far smoother), but..... Ofcom insisted on lowering it to 192 to allow for "greater diversity" i.e more stations. So sadly, the better your system, the more you hear the inherent HF distortions....

***
So do please make the effort with the AAC 320. If you can process it through an external DAC (USB or coax cable etc) you and your ears will be well rewarded....

Never understood why at least in the sticks , where there are fewer competing radio stations OFCOM couldn’t give R3 more bandwidth.
Even worse on telly where BBC 4 is only available as SD (rather than HD ) on DTT* where I live

* Standard definition, High Def, Digital Terrestrial TV

[Dave2002]

Interesting to see this discussion, which I'll join in more fully later. I always thought that FM was around 30dB not great - but significantly better than the average pop music station. I have not really been able to understand why there's a problem with digital sound and transmission. It should be possible to send the digital audio together with a steering signal - so that users who don't want to have any compression can get the full range, while others who may want to listen as background, or be concerned about a mix of music and speech can use other means to control the dynamic effects. What I'm basically suggesting is why not put the compression controls in the receivers rather than have the data pre-compressed at the broadcast end. After all - with digital broadcasting unlike analogue broadcasting - the signals should broadly speaking be the same output power irrespective of the content.

I also wonder whether some conductors - maybe many - have no idea about how quiet an orchestra can go - or alternatively how loud. Mostly they seem to produce a sound which varies between mf and ff - perhaps because they are scared to take sound levels too low. I do recall hearing Solti in a Prom conducting some Berlioz - and that really was quiet. Conductors who exploit that will generally get the attention of the audience, and they will go quiet and listen. On the other hand I also remember a concert with Jochum in which he must have instructed the horns to play flat out - as they were so loud it was harsh - very daring - brilliant concert. Many conductors just go for a sort of middle of the road dynamics - which is perhaps OK for some music, but not all.

Getting the balance right between instruments is another aspect of performance which may be not all performers or conductors attempt. A few instruments playing a little bit loud can spoil the effect. Maybe there's too much reliance on sound engineers to "sort these things out" - but if the musicians get things right, then the engineers should mostly be able to leave the controls alone - no need for close microphones etc., or microphones on bass drums or timpani to "give more impact" - well maybe!

[jayne lee wilson]

In fact, the Arcam Alpha 10 I mentioned above (a remarkable, technically advanced design, which inter alia showed you every transmitter it could find and their respective signal strengths) did indeed have user-adjustable dynamic range compression on a scale of 1-5, but not many listeners here would have wished to employ it! "5" was real steam-hammer, pop-radio stuff.

I think other designs implemented this too, but its too long ago to check now.... and as I said DAB simply wasn't good enough (too low a bitrate) for classical music to be a durable enjoyment, and as time went on, R3 was even reduced (quite often to accommodate Cricket Commentary etc) to a truly ruinous 160kbps.

"Spokespersons" tried to claim that the processors had been improved to allow for this; your ears told you otherwise. Having spent a small fortune on various tuners, I gave up on R3 concerts around 2006 because of this, and the 24/7 FM Compression, only returning when the first (much better) AAC streams began at 192kbps in 2009...

I recall my grand rapprochement with the Proms....plugging in a Cambridge DacMagic after midnight, feeding iplayer into it, and hearing the BBCSSO/Volkov playing Stravinsky's Orpheus from earlier that evening..... The Proms were back!

[Bryn]

Interesting to see this discussion, which I'll join in more fully later. I always thought that FM was around 30dB not great - but significantly better than the average pop music station. I have not really been able to understand why there's a problem with digital sound and transmission. It should be possible to send the digital audio together with a steering signal - so that users who don't want to have any compression can get the full range, while others who may want to listen as background, or be concerned about a mix of music and speech can use other means to control the dynamic effects. What I'm basically suggesting is why not put the compression controls in the receivers rather than have the data pre-compressed at the broadcast end. After all - with digital broadcasting unlike analogue broadcasting - the signals should broadly speaking be the same output power irrespective of the content.

I also wonder whether some conductors - maybe many - have no idea about how quiet an orchestra can go - or alternatively how loud. Mostly they seem to produce a sound which varies between mf and ff - perhaps because they are scared to take sound levels too low. I do recall hearing Solti in a Prom conducting some Berlioz - and that really was quiet. Conductors who exploit that will generally get the attention of the audience, and they will go quiet and listen. On the other hand I also remember a concert with Jochum in which he must have instructed the horns to play flat out - as they were so loud it was harsh - very daring - brilliant concert. Many conductors just go for a sort of middle of the road dynamics - which is perhaps OK for some music, but not all.

Getting the balance right between instruments is another aspect of performance which may be not all performers or conductors attempt. A few instruments playing a little bit loud can spoil the effect. Maybe there's too much reliance on sound engineers to "sort these things out" - but if the musicians get things right, then the engineers should mostly be able to leave the controls alone - no need for close microphones etc., or microphones on bass drums or timpani to "give more impact" - well maybe!

The DAB spec does indeed cater for the optional user-controlled variation in dynamic range compression you allude to in your first paragraph. However, most manufacturers either never implemented the necessary option to exploit the feature in the first place, or abandoned it for their mass-market DAB receivers.

[Ein Heldenleben]

In fact, the Arcam Alpha 10 I mentioned above (a remarkable, technically advanced design, which inter alia showed you every transmitter it could find and their respective signal strengths) did indeed have user-adjustable dynamic range compression on a scale of 1-5, but not many listeners here would have wished to employ it! "5" was real steam-hammer, pop-radio stuff.

I think other designs implemented this too, but its too long ago to check now.... and as I said DAB simply wasn't good enough (too low a bitrate) for classical music to be a durable enjoyment, and as time went on, R3 was even reduced (quite often to accommodate Cricket Commentary etc) to a truly ruinous 160kbps.

"Spokespersons" tried to claim that the processors had been improved to allow for this; your ears told you otherwise. Having spent a small fortune on various tuners, I gave up on R3 concerts around 2006 because of this, and the 24/7 FM Compression, only returning when the first (much better) AAC streams began at 192kbps in 2009...

I recall my grand rapprochement with the Proms....plugging in a Cambridge DacMagic after midnight, feeding iplayer into it, and hearing the BBCSSO/Volkov playing Stravinsky's Orpheus from earlier that evening..... The Proms were back!

Very interesting- I had no idea that had once been an option. An excellent idea. You could for example opt for compression in the kitchen or car .
There is some talk in the TV audio world of offering bespoke mixes so that for example music and FX could be kept down below dialogue if the viewer wanted. There are more complaints about this aspect of TV than any other (of a technical nature that is). There are now loudness meters which calibrate loudness of music and fx against speech and set limits . The problem is that young people want a noisy background and older people don’t …it is impossible to please everyone .

[Dave2002]

The DAB spec does indeed cater for the optional user-controlled variation in dynamic range compression you allude to in your first paragraph. However, most manufacturers either never implemented the necessary option to exploit the feature in the first place, or abandoned it for their mass-market DAB receivers.

I'm not sure if that's quite what I meant - but it is an option. Car systems often include some form of automatic control over dynamic range - to various user selectable degrees, and also boost the levels to (maybe) compensate for engine and road noise. Such systems do not have to be restricted to in-vehicle use.

Also by broadcast I now include digital streams sent over networks, so the audio could be much higher quality than current over the air programmes.

Apart from an automated user control, I was also thinking of a low bit rate steering stream which could be created by the broadcaster. That could be done manually at the transmitter end - though I guess since it would require human inputs and hence people to pay that's not going to happen, so automated systems will be what get done. At least it doesn't have to be at the broadcast end of the link, but I expect that user ignorance is most likely to force things to the "simplest" lowest common denominator.

Rather like people buying or using plug-in hybrid cars, but then only driving them with petrol - which saves the bother of reading the manual - thus ensuring that the most possible "damage" from such vehicles is achieved.

[mikealdren]

What we need is an option to turn down the volume of the Record Review presenters and turn up the music to get the relative levels correct.....

[Dave2002]

What we need is an option to turn down the volume of the Record Review presenters and turn up the music to get the relative levels correct.....

That would be a very simple thing to implement - if there was any will to do that. Unfortunately I don't think there is such a concern - either from broadcasters or equipment manufacturers.

I believe the BBC set the relative levels purely on the number of complaints they get. If too many people say they can't hear the dialogue, then they turn the voices up and the music down - for the next period, and if too many people say they can't hear the music then they turn that up - or maybe even worse - compress it to ***** (rude word) so that it doesn't offend too many people.

Arguably [very!! ] - this is a "democratic" way to provide acceptable sound. Ruin the sound first - then try (uselessly unsuccessfully) to put some control back so as to minimise end user complaints. Users who complain about sound quality have their views put straight in the bin - perhaps.

There are - allegedly - valid uses of various forms of compression for artistic purposes - which - again allegedly - can be used to retain clarity etc. - but in a radio context compression is often used just so as to boost the overall volume level and kill the dynamic range almost completely - usually - so they say - for commercial reasons such as advertising.

Sad, sad world!

[jayne lee wilson]

Fair enough D2K+2, but surely the main reason for FM dynamic compression was always in-car or portable or tabletop "transistor radio" listening; and this extends to the boosted audibiity of Presenters, Newsreaders and DJs on any of the digital platforms. So POp Radio was always going to be "all loud" to be heard everywhere, without ever being very loud and certainly never quiet. Which is why the technical presentation of such as Radio 3 was always going to be a very vexed question....

Long ago in an FM Galaxy far away, the Optimod Compressor was switched off before the evening concert....and if your roof aerial wasn't adequate, you noticed more white noise spoiling the pps.... hence the often large and complex roof arrays that came on the market...
Remember the Ron Smith Galaxies? I longed for the 17-element, whose prices suggested the art market for outdoor sculpture, but could never afford it.
THings of Beauty.....

https://www.google.com/search?q=ron+smith+galaxy+17&client=firefox-b-e&source=lnms&tbm=isch&sa=X&ved=2ahUKEwjFkMf39sz0AhVNa8AKHQ8aAfAQ_AUoAXoECAEQAw&biw=1200&bih=657&dpr=1.2

Angus McKenzie, doyen of HiFi writers on all things Radio, had a rotatable 23-element on a larger tower in his garden.... it might still be on Google Images somewhere...

But with BaL it is very annoying to have to keep adjusting volume oneself, and they could - and should - take another look at that.... but how many are listening on a HiFi anyway now?

[Dave2002]

Jayne - my point is that the world has moved on. Most productions are now done digitally - even if not all receiving equipment is. That applies even if the original source material was analogue as it's much easier for distributors to work with digital versions. I'm not absolutely sure about the BBC or other broadcasters, but I suspect that the days when you could hear a record skipping grooves in broadcasts have long gone. I'm not going to get into an argument about whether vinyl or shellac or digital are better as sources for broadcasting.

Given that the sources are 99.9% likely to be digital it would still be possible to have steering tracks which could be picked up by suitable end user devices and used to moderate/control the output volume. However, with FM radio - although that might still be a viable approach, the audio would have to be compressed for various reasons, so for FM radio compression would be an essential part of the distribution chain, and the broadcaster would most likely do that - though as we are discussing here - that might be far too severe.

With digital audio there is no good reason not to transmit the full dynamic range and then have the receiving/decoding device sort out the levels actually heard by the end user.

My reference to cars earlier was not restricted to FM radio - my previous car would do the dynamic compression even on other material such as CDs - and for reasons to do with adjusting the level to combat engine and road noise.

A more flexible - and hopefully "intelligent" approach to this could be adopted (but won't be ...) so that you don't have to keep adjusting the volume.

[Bryn]

Jayne - my point is that the world has moved on. Most productions are now done digitally - even if not all receiving equipment is. That applies even if the original source material was analogue as it's much easier for distributors to work with digital versions. I'm not absolutely sure about the BBC or other broadcasters, but I suspect that the days when you could hear a record skipping grooves in broadcasts have long gone. I'm not going to get into an argument about whether vinyl or shellac or digital are better as sources for broadcasting.

Given that the sources are 99.9% likely to be digital it would still be possible to have steering tracks which could be picked up by suitable end user devices and used to moderate/control the output volume. However, with FM radio - although that might still be a viable approach, the audio would have to be compressed for various reasons, so for FM radio compression would be an essential part of the distribution chain, and the broadcaster would most likely do that - though as we are discussing here - that might be far too severe.

With digital audio there is no good reason not to transmit the full dynamic range and then have the receiving/decoding device sort out the levels actually heard by the end user.

My reference to cars earlier was not restricted to FM radio - my previous car would do the dynamic compression even on other material such as CDs - and for reasons to do with adjusting the level to combat engine and road noise.

A more flexible - and hopefully "intelligent" approach to this could be adopted (but won't be ...) so that you don't have to keep adjusting the volume.

Regarding the playing 'old' vinyl discs on Radio 3, I recall when the Le Roux recording of Messiaen's Turangalîla-Symphonie was played on, I think CD Masters, they got one of the engineers to carefully transfer the (French) DECCA double LP re-issue to digital for the broadcast, all commercial CD issues of the recording being pretty dire in comparison to the DECCA LPs. I never did get a decent pressing of the original Vega LPs, despite returning them for exchange several times. Fortunately, I also got the DECCA LP double album when it came out. As far as I know, there is still no decent commercial CD issue of the recording, despite several releases on an individual disc or in a boxed set.

[Ein Heldenleben]

That would be a very simple thing to implement - if there was any will to do that. Unfortunately I don't think there is such a concern - either from broadcasters or equipment manufacturers.

I believe the BBC set the relative levels purely on the number of complaints they get. If too many people say they can't hear the dialogue, then they turn the voices up and the music down - for the next period, and if too many people say they can't hear the music then they turn that up - or maybe even worse - compress it to ***** (rude word) so that it doesn't offend too many people.

Arguably [very!! ] - this is a "democratic" way to provide acceptable sound. Ruin the sound first - then try (uselessly unsuccessfully) to put some control back so as to minimise end user complaints. Users who complain about sound quality have their views put straight in the bin - perhaps.

There are - allegedly - valid uses of various forms of compression for artistic purposes - which - again allegedly - can be used to retain clarity etc. - but in a radio context compression is often used just so as to boost the overall volume level and kill the dynamic range almost completely - usually - so they say - for commercial reasons such as advertising.

Sad, sad world!

In fact the BBC technical requirements for programme delivery are publicly available and the sound delivery requirements haven’t changed that much.

https://www.bbc.co.uk/delivery/technical-requirements