FM switchover.....Coalition steamroller?

[Dave2002]

Radio 3 FM - excellent basic sound quality, equivalent in digital terms to ca. 832 kbps, but usually ruined by dynamic range compression.

R3 DAB - good dynamic range ruined by low bitrates - 192/160 kbps - well below the recommendations of BBC's own 1990s r&d, which said 256/224 kbps were necessary for high quality stereo.

R3 'HDs' 320 kbps AAC - if you can get it, the only show in town!

I don't see how you can make the equivalence between FM and 832kbps digital?

Why the accuracy down to the last 2 kbps? I would submit that there is no scientific basis for this, and probably not even a subjective one. Perhaps you took the frequency range for FM and allocated what you felt to be an appropriate number of bits/sample and multipled by 2 for stereo? When I do the same for a frequency range of 15 kHz with 16 bit samples I get at least 960 kbps - but that is the raw PCM rate, not the rate which would be needed in a communication system using digital compression.

I agree that 320 kbps AAC is rather good for broadcast quality sound.

[amateur51]

Supplying keys for them to wind them up with, it wouldn't surprise me.

Isn't that Mr Bayliss' baileywick?

[DracoM]

See my letter from DCMS in the FM/DAB thread.

[Frances_iom]

it's almost a done deal - FM won't be switched merely given over to low power local broadcasters (ie franchised from a national junk/pop station - judging from the wording possibly hard right 'christian' garbage station along lines of Faux news) thus the gov including the true blue Yellow party can claim theat the sets are still usable, DAB will remain a little better than the old AM station until the next change which will require payment for an encrypted digital service

[OldTechie]

I don't see how you can make the equivalence between FM and 832kbps digital?

Why the accuracy down to the last 2 kbps? I would submit that there is no scientific basis for this.

I think many FM transmitters are still fed by the NICAM multiplexed system.

13 bits X 32kHz sample rate x 2 channels = 832000 bits per second. The BBC report on the multiplexed data distribution system states it achieves quality that matches the performance of the previous digital system that ran at this rate. It is only ca. 832 kbps because it is an equivalence (and you might choose to divide by 1024 rather than 1000 for kbps.) I think it now uses a 14 bit sample compressed to a 10 bit value scaled in blocks before multiplexing. See http://www.bbc.co.uk/rd/pubs/reports/1973-42.pdf for the orginal work. Because there are 6 channels on a 2048kbps feed it really only has an average 341kbps per channel.

However, my local transmitter has an analogue line from London so instead of the reduced sample accuracy, the BBC inserts hum by routing the incoming circuit into the transmtter building over the incoming power feeds. It is probably an analogue feed to Wrotham as well so that may be really good.

If we could get them to agree not to abandon FM, I would like to build a Radio 3 processor that worked by digitising the FM feed, and then using the Freeview feed as a control signal to undo the Optimod DRC they use in the FM feed (and digitally remove the hum by grabbing a sample when the Freeview feed says it is silent and replaying it locked to mains.) That sounds like hard work so if FM is doomed I won't bother. In the mean time it's 320kbps AAC for me.

Unfortunately this particular consultation is not about whether FM is a desirable signal to retain, it is about whether the Cost Benefit Analysis has been done correctly. That is complicated to argue about. I am certainly having trouble getting my head around it all.

I think it is good to make a final decision in 2013 as proposed. We have had years of saying there will be a change over, but with conditions that delay it for a totally indeterminate time. The time has come to state publicly that FM will not be abandoned after all, or that it is definitely dead and nobody should buy another FM receiver of any kind. Maybe we can get the DCMS to decide to keep FM, but I think the officials have made up their mind because someone told them that the future is digital. Politicians of all parties have not realised how unpopular a decision to abandon FM will be, and they are easily convinced that digital must be better. I wonder whether they have realised that using Big Ben as a time signal with some precision will no longer be possible because all digital delivery methods involve too much delay.

[Resurrection Man]

.... it is about whether the Cost Benefit Analysis has been done correctly. ....

It hasn't. Based on PwC report figures that dreamed up a notional benefit (unquantified or redacted) to consumers of £781 million based on 'choice'.

[Gordon]

I think many FM transmitters are still fed by the NICAM multiplexed system.

See here for a history of NICAM and its variants:

http://www.audiomisc.co.uk/BBC/PCMandNICAM/History.html

NICAM is nominally still used everywhere in the "FM" network but how long the NICAM/FM equipment can be maintained for is debateable. BH is entirely digital and has not seen an analogue circuit in a very long time. There are no audio analogue lines left because no carrier [eg BT] has them any more; they used to be carried on telephony "phantom" circuits. NICAM systems originally used analogue video TV circuits to carry the multiplex. Since such circuits also went a while ago it's difficult to see how the NICAM stream is carried now unless it is purely digital carriage. The only place in the FM network where you will find "analogue" is at the input to the transmitter power amplfier and that will be buried in circuitry. Even the dreaded Optimods in BH are digital.

13 bits X 32kHz sample rate x 2 channels = 832000 bits per second. The BBC report on the multiplexed data distribution system states it achieves quality that matches the performance of the previous digital system that ran at this rate. It is only ca. 832 kbps because it is an equivalence (and you might choose to divide by 1024 rather than 1000 for kbps.) I think it now uses a 14 bit sample compressed to a 10 bit value scaled in blocks before multiplexing. See http://www.bbc.co.uk/rd/pubs/reports/1973-42.pdf for the orginal work. Because there are 6 channels on a 2048kbps feed it really only has an average 341kbps per channel.

However, my local transmitter has an analogue line from London

Really!! How do you know? See link above.

so instead of the reduced sample accuracy, the BBC inserts hum by routing the incoming circuit into the transmtter building over the incoming power feeds. It is probably an analogue feed to Wrotham as well so that may be really good.

Analogue stereo lines were hell to keep in alignment especially group delay which is one reason why NICAM was designed in the first place. NICAM audio performance is close to that of 14 bit PCM and is bandlimited in a similar way to FM [ie nominally around 15 kHz as per ITU specs but in practice it may be a little more - significantly more and it affects modulation index and hence S/N and distortion performance - but also falls foul of the pilot tone at 19kHz. I fail to see how hum from power lines gets into digitally carried audio and then becomes audible.

If we could get them to agree not to abandon FM, I would like to build a Radio 3 processor that worked by digitising the FM feed, and then using the Freeview feed as a control signal to undo the Optimod DRC they use in the FM feed (and digitally remove the hum by grabbing a sample when the Freeview feed says it is silent and replaying it locked to mains.) That sounds like hard work so if FM is doomed I won't bother. In the mean time it's 320kbps AAC for me.

That would be hard work!!

Unfortunately this particular consultation is not about whether FM is a desirable signal to retain, it is about whether the Cost Benefit Analysis has been done correctly. That is complicated to argue about. I am certainly having trouble getting my head around it all.

That is correct - this is not a consultation about the CBA it is about the methodology by means of which it will be carried out.

I think it is good to make a final decision in 2013 as proposed. We have had years of saying there will be a change over, but with conditions that delay it for a totally indeterminate time. The time has come to state publicly that FM will not be abandoned after all, or that it is definitely dead and nobody should buy another FM receiver of any kind. Maybe we can get the DCMS to decide to keep FM, but I think the officials have made up their mind because someone told them that the future is digital. Politicians of all parties have not realised how unpopular a decision to abandon FM will be, and they are easily convinced that digital must be better. I wonder whether they have realised that using Big Ben as a time signal with some precision will no longer be possible because all digital delivery methods involce too much delay.

Elements in the radio industry are behind the push to DAB - but government is certainly energised by the Digital Britain idea despite the fact that there is littl;e spectrumn to be regained and auctioned as has been the case with 800 MHz UHF TV spectrum currently being offered. Some radio manufacturers are dead against it but not for the reasons you might think. Some of these manufacturers are also dead against DAB+ [ie HE AAC with SR] being mandated in the switchover DAB receiver specification. One reason among several is that they don't want to pay the patent royalties.

By the way, David Meares, who wrote that report about 256/224 kBit/s DAB being needed for FM equivalence, was long ago convinced that this was based on old lab designs of coder and has been reconciled to the current 192 of R3 [on a good day, when it isn't 160].

[OldTechie]

See here for a history of NICAM and its variants:
I fail to see how hum from power lines gets into digitally carried audio and then becomes audible.

So do I. You asked what made me think we have an analogue circuit...

It is an odd situation, the site RBRs TV from Crystal Palace and radio from Wrotham or CP. However, they could not resolve dreadful interference on the Radio 3 reception at the site caused by other (non broadcast) telecomms gear around them. So in 2001 they put in a circuit to feed Radio 3 only. There is no hum on Radio 4 from the site. I could not tell you about Radio 1 and 2 - they never go quiet enough to hear anything. See http://personal.ee.surrey.ac.uk/Pers...ansmitter.html for the tale of this original issue as posted by the persistent guy who may have persuaded them to install the circuit.

I did not suffer from this issue because I had NTL cable TV and it had a successfully frequency-shifted version of Wrotham or CP on its analogue TV system. That stopped when they went to digital TV only. Speaker failure had sent me back to using my old Rogers Cadet II speakers that I was using in 1962. So, I could not hear the hum when I went back to using the local transmitter (not much output at 50Hz.)

When I bought some decent speakers a year or two ago (definitely OK at 50Hz) I instantly found hum on Radio 3 FM

I complained about the hum to the BBC. They despatched an engineer to check it, and reported back to me that they agreed it was poorly. They told me they were going to fix it. I knew they were on site when there was a short break in transmission after which it was perfect (except fot the Optimod) and without any hum, so I took a 10 minute walk up to the transmitter to say thank you to whoever was there.

I found an Arqiva man on site. He explained that the signal came into the BBC cabin from a BT cabin. Back in 2001 BT had fitted a cable termination on the outside of the BBC cabin, punched a hold through the wall and fed the cable into a convenient piece of trunking. This was, unfortunately just above the main switchear for the whole installation. The trunk then ran round the building almost 360 degrees to where the lines termination panel for the BBC kit existed. This would not be too bad if it had been a purely technical trunk, but is seems it was just a divided trunk with power in it.

Arqiva ran in a temporary cable from the BT cabin direct to the BBC panel and it worked fine. The hum was gone. But as that meant leaving the doors to both cabins open to the outside world it was not too practical as a long term solution.

They spent two weeks on site rewiring, but the entry point could not really move without digging up half the site. I suspect they installed a new trunk in basically the same route. After all this, the hum is a few dB less. Maybe a full turn coil of the audio cable round a set of racks full of TV transmitters is not healthy. I fear the new hum level now meets any SLA conditions agreed with Arqiva. Certainly the BBC does not seem inclined to have another go at it.

I did ask the Arqiva man whether the decoder from the digital circuit was in the BT cabin, in which case they could surely just have moved it to the BBC cabin, but he said it was an analogue circuit all the way from London. If that is the case I guess we have a pretty clean version of the output of the Optimod. The equalisation seems perfect to my ears - I can detect no frequency response difference between FM and any of the digital feeds. (Mind you, age has fitted a pair of 11 kHz low pass filters to my head so what happens above that I have no idea.)

There is an attractive quality to the sound on FM that makes it more CD-like than any of the digital feeds of Radio 3 provided I ignore the hum and the Optimod effects. When the source level is low enough that the Optimod is not changing the gain within the dynamics of the music but high enough to overcome the hum, both faults can be ignored. (Note that I do play the AAC radio 3 feed and CDs through the same decent quality DAC and without passing through the Windows 7 mixer.)

I'm inclined to believe the Arqiva man I met at the transmitter and that maybe we do have a remnant analogue circuit.

I think there may be justifiable reluctance from the BBC to spend money on the FM distribution system when they don't know whether it has a 25 year life or is to be cut off by regulation in 18 months time.

[Resurrection Man]

.......
I think there may be justifiable reluctance from the BBC to spend money on the FM distribution system when they don't know whether it has a 25 year life or is to be cut off by regulation in 18 months time.

I would dearly love to see the details of the contract between Arqiva and the BBC for the running costs etc for the Fm transmission chain. One could envisage the contract being drawn up on the assumption that DAb would have been here by now. That the existing FM transmitters would have been switched off. Instead, I suspect that there is a large replacement cost looming were FM to continue and that Arqiva are anxiously looking at the contractual terms. I can't see much support for the BBC on this given Hunt's stance on the BBC namely he is against its funding etc and seeks to destroy it.

This MOU makes an interesting read. Link

Nowhere does it say anything about the cost that the listening public will incur.

And comment from an industry insider here http://radiotoday.co.uk/2012/07/dab-...lete-shambles/

[Resurrection Man]

It is quite interesting (but not surprising) to see the vested interests at work.

While searching for any information on FM transmitters etc I came across this disingenuous rebuttal from a Matt Deegan. http://www.mattdeegan.com/2012/07/02...ttal/#comments

There are so many errors in his rebuttal to go into here but the key one is where he says

First of all – analogue radio isn’t being ‘turned off’. There will continue to be small commercial and community radio stations on FM.

Secondly – ‘we’ (whoever that is) aren’t, haven’t and won’t be forcing people to buy digital radios. As we haven’t been doing that, it can’t have failed.

So I emailed him to ask him what he understood by the phrase "isn't being turned off" when many millions of us listen to R3, R4 etc on FM on our transistor radios and that this service will no longer be available post any switch off of FM.

He replied "Post the switchover of large stations to DAB, small stations will continue to broadcast on FM. "

So I then asked him if by "small stations" he meant Radio3 and Radio 4...answer came there none.

Digging a little deeper, he works for Folder Media http://www.foldermedia.co.uk/ whose major line of business is in DAB multiplexes. So there's a surprise then.

[DracoM]

Fine research - many thanks.

Do we sense that the govt are getting very cold feet about the whole gruesome business? Could the BBC tell them privately that this isn't a goer?

[Resurrection Man]

Thanks, DracoM

A bit more digging. This from the BBC's own submission to OFCOM on the proposed DAB roll-out...

The BBC’s national DAB network is operated across the UK as a single-frequency network
(SFN). While this is a highly-efficient use of frequency resource, it does suffer from one
major limitation.
.........
The BBC network has already reached the stage where its coverage is, in places, being
limited by self-interference. This is particularly obvious in the patchy coverage visible in
Map A1 in the East Anglia area. As more transmitters are added to the network to improve
coverage, the interference regime increases, paradoxically making coverage improvements
harder.

This is repeated constantly in this submission http://stakeholders.ofcom.org.uk/bin...es/annex-c.pdf

Looks like DAB rollout has a bit of a technical problem...but one that seems not to be getting much coverage

[Gordon]

.............I did ask the Arqiva man whether the decoder from the digital circuit was in the BT cabin, in which case they could surely just have moved it to the BBC cabin, but he said it was an analogue circuit all the way from London. If that is the case I guess we have a pretty clean version of the output of the Optimod. The equalisation seems perfect to my ears - I can detect no frequency response difference between FM and any of the digital feeds. (Mind you, age has fitted a pair of 11 kHz low pass filters to my head so what happens above that I have no idea.)

There is an attractive quality to the sound on FM that makes it more CD-like than any of the digital feeds of Radio 3 provided I ignore the hum and the Optimod effects. When the source level is low enough that the Optimod is not changing the gain within the dynamics of the music but high enough to overcome the hum, both faults can be ignored. (Note that I do play the AAC radio 3 feed and CDs through the same decent quality DAC and without passing through the Windows 7 mixer.)

I'm inclined to believe the Arqiva man I met at the transmitter and that maybe we do have a remnant analogue circuit.

Well you do have an unusual situation down there don't you, pretty much a one off relic!! An RBR would just transpose but the interference on R3 is an issue. I'm still not clear as to what exactly is happening there - if the link is analogue stereo audio isn't it a baseband GE/Zenith signal already with RDS on top? Did that happen at BH after the Optimod? If so and it arrives by devious and peculiar means at your RBR site then I see how a bit of local bodgery has added hum on a cross site cabin-cabin link. How an analogue circuit can be provided all the way from London [from all digital BH??] these days is a bit of a puzzle though. I thought that Energis do the networking using fibre nowadays too, not BT, although Arqiva certainly own and operate the transmitters? Must tap Arqiva contacts at the local maintenance base around Croydon/CP and see what can be discovered.

See this from my old pal Nick Tanton:

http://www.bbc.co.uk/rd/pubs/papers/pdffiles/ibc98nt.pdf

it primarily describes the TV network for Freeview [the paper is from 1998 when Freeview as we know it didn't yet exist] but by implication there are references to the changeover to digital network technologies during the 90s which would also include capacity for the radio networks. I have not been able to find any definitive BBC ResD papers or publications that delve into radio network provision. The original NICAM papers are from the 70s and there are some articles from the 1980s and 90s in the BBC Eng Inf periodical that report on some stages in the roll out of the new system.

[Dave2002]

The BBC network has already reached the stage where its coverage is, in places, being
limited by self-interference. This is particularly obvious in the patchy coverage visible in
Map A1 in the East Anglia area. As more transmitters are added to the network to improve
coverage, the interference regime increases, paradoxically making coverage improvements
harder.

This is repeated constantly in this submission http://stakeholders.ofcom.org.uk/bin...es/annex-c.pdf

Looks like DAB rollout has a bit of a technical problem...but one that seems not to be getting much coverage

Perhaps that is only really a problem in areas where reception over 74km is realistically possible. That could explain why East Anglian rollout has encountered problems. Other areas, such as in Scotland, where there are many obstacles, may not experience these problems as distant signals will be stongly attenuated.

[Resurrection Man]

Perhaps that is only really a problem in areas where reception over 74km is realistically possible. That could explain why East Anglian rollout has encountered problems. Osther areas, such as in Scotland, where there are many obstacles, may not experience these problems as distant signals will be stongly attenuated.

I didn't get that impression on reading the publication, Dave (admittedly speed reading)