What bitrate is needed to sound like analog FM?

[MLXXX]

If anyone can supply me with a link to an official file considered to represent "FM quality", or an official definition, I'd be obliged. Over the years I've seen many references to "FM quality" but no actual definition, or sample file.

My own experience is that using an older horizontally polarised external TV antenna connected to a medium quality AVR in a high signal strength metroplolitan reception area provides me with reception with a subjectively very good audio quality. For my ears, for a range of program material, it is markedly better than 48kbps DAB+, somewhat better than 64kbps DAB+. For my ears for classical music, it is noticeably better than 80kbps DAB+. I've found a modern car FM radio can also perform well, though with greater noise and subject to occasional glitches from multipath reception as the car moves along the road.
 

The poor audio quality of real life DAB+ and real life DRM relative to FM radio in a high signal strength reception zone has been mentioned a number of times on this forum over the years. I do not however recall the following research paper having been mentioned or discussed. I happened across it today and thought some forum members beyond myself might find it to be of interest, despite it having been published back in 2013!

_____________

 

Research published in 2013 concluded that a very high bitrate was required to achieve FM quality. This was certainly much more than the 48 or 64kbps commonly used in Australia for DAB+, or the even lower audio bitrates commonly used internationally for DRM30 broadcasting. Access to the full paper requires payment of a fee or possible free access at the discretion of the authors. For the purpose of this thread I will merely rely on:

1. The freely available abstract of the paper.

2. Comments from two sources, being observations by people who have read the paper and prepared relatively formal remarks.

Off we go then!

1. The abstract, accessible at http://www.aes.org/e-lib/browse.cfm?elib=16969:-

AES E-LIBRARY

Perceived Audio Quality of Realistic FM and DAB+ Radio Broadcasting Systems

The perceived audio quality of a digital broadcasting system (such as DAB+) is dependent on the type of coding and bit rates selected. Because of bandwidth constraints, the required number of channels, and conflicting auxiliary services, audio quality is sometimes degraded. In designing a broadcast system, it is necessary to have well-defined criteria for minimally acceptable quality. Two studies explored quality criteria and how quality degrades for various bit rates. For DAB+ the subchannel rate should not be less than the currently available maximum of 192 kbits/s for a stereo signal, which would be comparable to the quality of a modern FM system. Rates below 160 kbit/s can significantly degrade certain types of program material. To be truly perceptually transparent, bits rates of close to 300 kbits/s may be needed when using the current generation of coders.

Authors:

Berg, Jan; Bustad, Christofer; Jonsson, Lars; Mossberg, Lars; Nyberg, Dan

 

Affiliations:

Luleå University of Technology, Piteå, Sweden; Swedish Radio, Stockholm, Sweden; Swedish National Laboratory of Forensic Science - SKL, Linköping, Sweden

(See document for exact affiliation information.)

  

2. Comments at http://www.radiojackie.com/im/Digital%20radio%20AES%20research.doc

 

DIGITAL RADIO – AES RESEARCH

 

An important new study has been published in the Journal of the Audio Engineering Society: Perceived Audio Quality of Realistic FM and DAB+ Radio Broadcasting Systems.

 

Many different music and speech samples were used in listening tests with the conclusion that a bit rate "close to 300 kbit/s" is required to avoid falling below FM quality.

Neither UK-type DAB (from 1985) nor DAB+ (from 2007) are capable of offering this.

 

This research is significant for two reasons.

- For the first time, comparisons are made between digital radio and realistic FM radio.

- Processed audio was used, rather than unprocessed sounds.

   Processed audio (e.g. 'Optimod') is universal in radio.

 

Previous work, with unprocessed sounds, had suggested the necessity of 320 kbit/s

to pass the perceptual transparency and statistical undetectability point.

This was adopted by the BBC in 2010 for HD Sound in iPlayer.

 

These results with processed audio dispel the notion that the sound quality presently possible via digital radio is found inferior by only a tiny minority of audiophiles.

 

This new evidence, specifically examining everyday radio audio, supports the view that DAB has a role as a supplementary platform but could never prove satisfactory as the sole outlet over the airwaves.

 

 

Trevor Brook - 11 November 2013

Surrey Electronics  - radiofax.org - Editor: radiojackie.com

 

3. Comments at http://www.engineeringradio.us/blog/2013/11/what-bitrate-is-needed-to-sound-like-analog-fm/

What bitrate is needed to sound like analog FM?

By Paul Thurst, on November 25th, 2013 7 comments 

As it turns out, 300 kbp/s or greater.  At least in critical listening environments according to the paper titled Perceived Audio Quality of Realistic FM and DAB+ Radio Broadcasting Systems (.pdf) published by the Journal of the Audio Engineering Society. This work was done by group in Sweden and made various observations with different program material and listening subjects. Each person was given a sample of analog FM audio to listen to, then they listened to various audio selections which were using bit reduction algorithms (AKA CODEC or Compression) and graded each one.  The methodology is very thorough and there is little left for subjective interpretation.

In less critical listening environments, bit rates of 160-192 kbp/s will work.

...

[Note: The article immediately above goes on to provide a detailed table.]

I personally am not at all surprised at the reported findings!

[cwt]

56 minutes ago, MLXXX said:

I personally am not at all surprised at the reported findings!

Concur as well MLXXX; the well known Stereophile magazine did a test once along similar lines on the audibility of different codecs at different bit rates ; signal to noise ratio's slumped as the bitrate dropped . In fact depending on the codec something can easily be sacrificed like an extended frequency response [you would get with analog fm] in order to prop up another important specification like the aforementioned s/n ratio

http://www.stereophile.com/content/mp3-vs-aac-vs-flac-vs-cd-page-2#0PahPl6OQtYI9mZ1.97

All these codecs have tradeoffs due to their lossy nature ; a good fm tuner is unfortunately a dying breed in component hifi these days.. With hard drives so reasonably priced these days its an easy choice to ignore these lossy codecs and just convert to flac ; otherwise you would still have famous names like Naim ;Meridian ;Cambridge audio and many others still producing state of the art tuners  imo

[Guest Malich]

Fulltext of the article is available by following the link on this page through a couple of steps (just links, no registration). I think it's public access - at least, it doesn't appear to be using any of my access credentiaIs - but if it doesn't work for you, I can post or PM the article.

(Note: I haven't read it thoroughly, but it seems to be the basis of / related to the slides from an industry presentation I posted here a couple of years ago. Good catch, MLXXX!)

I don't understand why it's even still up for dispute though. The codec developers (e.g. Fraunhofer) themselves give rough equivalents that suggest it takes ~192kbps+ AAC to match good FM. Every bit of independent 3rd-party research, including this one involving radio / audio industry engineers, says the same thing. Everyone who looks at the research &/or does their own tests comes to the same basic conclusion - bitrates typically used in DAB/DAB+ & DRM are simply too low to match "FM quality", let alone the "near CD-quality" sometimes claimed.

It's only the DAB/DRM/etc marketing departments - and the fools who believe them - who claim otherwise.

[MLXXX]

3 hours ago, Malich said:

 

Fulltext of the article is available by following the link on this page through a couple of steps (just links, no registration). I think it's public access

 

Thx very much, Malich. I've been able to download it using that link and without using any credentials. I'll read through it in detail when I have the chance.

 

3 hours ago, Malich said:

I don't understand why it's even still up for dispute though.

It shouldn't be. I think some of the people who promote digital radio have fallen into the trap of believing propaganda rather than performing their own listening tests. If they only listened critically to some good quality FM radio it wouldn't take long for them to realise that nominal 48kbps or 64kbps DAB+ doesn't measure up. I'm assuming average hearing ability, not exceptionally good hearing ability.

But a key factor I feel is the lack of definiteness as to what is actually meant by "FM quality". If you use a portable radio at the outer edge of an FM reception zone, there may be a good deal of hiss, a loss of stereo, and if you are in a moving vehicle, serious dropouts and glitches from multipath. That however is not the type of sound I get from my FM car radio about town!

 

11 hours ago, cwt said:

a good fm tuner is unfortunately a dying breed in component hifi these days

I do certainly find variations in subjective audio quality between FM tuners/radios. I find the FM sound quality from my portable Roberts DAB+ radio is not as good as the FM sound quality of one of my old AVRs which in turn is not as good as the FM quality from my current mid-priced AVR. I haven't attempted measurements but I hear a slight edge (it sounds like distortion) to the sound on some FM radios that I don't hear on others. However even with such an edge, I find the FM sound for classical music noticeably better than DAB+ sound at a nominal 80kbps.  The DAB+ version has a blandness that takes away much of the enjoyment I hear with the FM version. I find a small amount of background hiss, and a small amount of distortion, to be minor detriments in comparison with the effects of a psychoacoustic codec operating at a modest bitrate.

[alanh]

Sound quality should not be compared with FM but to the origninal source.

FM has 15dB high frequency boost prior to modulation. So broadcasters use processors  which measure the signal level in each octave band and reduce the gain at high levels. So for example a symbol crash will have lots of its high frequency level reduced prevent overmodulation thus dulling the sound.

This is not required in audio systems or any digital radio system.

Where are the graphs of the MUSHRA Scores to show the sound quality against the bit rates under standardised double blind testing?

So I don't care what MLXXX says, he does not do any quality assessments without knowing the system therefore introduces bias. Why don't you complain about the compression on TV video broadcasters are not transmiting the HD video at 829 Mbit/s!

 

Alanh

[hrh]

1 hour ago, alanh said:

Sound quality should not be compared with FM but to the origninal source

If the comparison(s) were done as such then the differences would be even more marked.
The FM versus DAB+ comparison is relatively easy to do with one receiver, swapping between the FM version and the digital one and can be reasonably obvious.

[MLXXX]

2 hours ago, alanh said:

So broadcasters use processors  which measure the signal level in each octave band and reduce the gain at high levels. So for example a symbol crash will have lots of its high frequency level reduced prevent overmodulation thus dulling the sound.

If you read the report alanh you would see that the FM2 source involved standard processing, use of an FM exciter, and then demodulation with an FM receiver. Despite that, the digital methods required very high bitrates for comparable MUSHRA scores, much higher than digital bitrates in common use for DAB+.

[danTheMan1503560646]

48 minutes ago, hrh said:

If the comparison(s) were done as such then the differences would be even more marked.
The FM versus DAB+ comparison is relatively easy to do with one receiver, swapping between the FM version and the digital one and can be reasonably obvious.

Here's another experiment you could try:

1. Rip a CD with nicely mastered audio (assuming this is the gold standard and you believe in Nyquist/Shannon sampling theorem) 

2. Encode the CD to AAC+ at different bitrates. This is essentially what DAB+ gives you.

3. Hook your sound card output to an FM transmitter you can get on ebay, granted these are probably not as good as industrial ones.

4. Compare the uncompressed rip to the various AAC+ files and the FM reception

Just from casual listening, I notice that DAB+ can actually be silent when it needs to be, but can become messy with difficult to encode sounds (such as applause) where FM shines.

[alanh]

This has to be a joke, comparing the quality of a compression system using a radio with a mono speaker like this https://www.robertsradio.com/uk/products/heritage/revivals How about getting a real DAB+ and a real FM tuner fed into a good quality amplifier and decent speakers? such as http://klappav.com.au/products/rt-1570  or http://www.intl.onkyo.com/products/hi-fi_components/tuners/t-4070/index.html adding some good speakers. DAB+ in some cars would be a better device for comparison. Headphones have their own problems with bass response and the direction of the sound which appears to be from within the head.

After in designing the AAC system they used a professional recording studio to produce lossless recordings at a high data rate, then encoded them at the selected rate, decoded them fed then into broadcast quality audio amplifer and monitoring speakers, in an acoustically treated room.

Also the type of program has a great effect, speech is different to music. Rock musicians have a device called a fuzz box which distorts the sound on purpose.

The settings on an audio processor will greatly colour the result, because the settings and type of processing for FM is different to digital and for that matter AM. As I said you need to compare both back to the origiinal at least to a lossless format such as a CD or .WAV file.

You should also use the ABX tester or similar http://www.mcelhearn.com/can-you-really-tell-the-difference-between-music-at-different-bit-rates/ which will shuffle the samples and keep track of which is which. When this is done a WAV from the recording studio should then be compared to a .wav file produced from an FM receiver along with the xHE-AAC samples at different bit rates and with different types of program.

 

Just keep in mind that 3.5 million Australians are listening to Digital Radio and the number is rising despite what  MLXXX says.

 

Alanh

[Guest Malich]

21 minutes ago, alanh said:

This has to be a joke, comparing the quality of a compression system using a radio with a mono speaker like this https://www.robertsradio.com/uk/products/heritage/revivals How about getting a real DAB+ and a real FM tuner fed into a good quality amplifier and decent speakers? such as http://klappav.com.au/products/rt-1570  or http://www.intl.onkyo.com/products/hi-fi_components/tuners/t-4070/index.html adding some good speakers. DAB+ in some cars would be a better device for comparison. Headphones have their own problems with bass response and the direction of the sound which appears to be from within the head.

After in designing the AAC system they used a professional recording studio to produce lossless recordings at a high data rate, then encoded them at the selected rate, decoded them fed then into broadcast quality audio amplifer and monitoring speakers, in an acoustically treated room.

Also the type of program has a great effect, speech is different to music. Rock musicians have a device called a fuzz box which distorts the sound on purpose.

The settings on an audio processor will greatly colour the result, because the settings and type of processing for FM is different to digital and for that matter AM. As I said you need to compare both back to the origiinal at least to a lossless format such as a CD or .WAV file.

You should also use the ABX tester or similar http://www.mcelhearn.com/can-you-really-tell-the-difference-between-music-at-different-bit-rates/ which will shuffle the samples and keep track of which is which. When this is done a WAV from the recording studio should then be compared to a .wav file produced from an FM receiver along with the xHE-AAC samples at different bit rates and with different types of program.

Well, I guess we can all see who hasn't bothered to read or understand the paper, and is instead running on misinformed opinion and arrogance...

 

23 minutes ago, alanh said:

Just keep in mind that 3.5 million Australians are listening to Digital Radio and the number is rising despite what  MLXXX says.

Just keep in mind that the research presented in the paper was undertaken by broadcast & audio engineers, performed under the same requirements AlanH has specified, overseen by an expert in the field of comparative audio testing, and analysed by an expert in qualitative evaluation of perceptual audio codecs. Despite what AlanH says...

[alanh]

An even bigger joke,

FM radio transmits a mono signal for the single speaker radio and the when commercial FM radio started in Australia all FM stations added the 38 kHz Double Sideband Suppressed carrier carrying the difference signal between Left and Right microphones. This high frequency end of the modulated signal is much more suscepible to noise. When there is no radiated signal present the FM receiver hisses (which is mainly high frequencies) but receivers usually mute this.

In addition on the website there is no mention of an antenna socket, so I can only assume they have a telescopic rod or a wire antenna particularly on battery operated radios. One of the characteristic of FM is that the distortion rises and the channel separation reduces (can't hear that on a mono speaker). Telescopic rod antennas are generally omnidirectional and when inside often receive reflected signals. Who can remember the multiple images on analog TV with an indoor antenna and trying to adjust it to get rid of them.

So using his Roberts radio speaker he listens to a mono sound even if he has one with twin speakers they are so close together the stereo effect is minimal.

As for DAB+ it has Spectral Bandwidth Replacement to replace the removed high frequencies. The speaker in this radio is not capable of these frequencies anyway, in addition the speaker is not big enough to produce low pitched bass to balance the high frequencies, so the radio sounds comparitively tinny.

Not only that you cannot hear the parametric stereo on a mono speaker.

Since MLXXX promotes AM radio but makes no lament of the lack of stereo, here is the reason, he doesn't listen to stereo on any other broadcasts either.

I am glad to say that listening to DAB+ on a decent sound system sounds great particularly with the programs not available on analog.

Alanh

Edited November 4, 2016 by alanh
twin tiny speakers

[Guest Malich]

5 minutes ago, alanh said:

An even bigger joke,

FM radio transmits a mono signal for the single speaker radio and the when commercial FM radio started in Australia all FM stations added the 38 kHz Double Sideband Suppressed carrier carrying the difference signal between Left and Right microphones. This high frequency end of the modulated signal is much more suscepible to noise. When there is no radiated signal present the FM receiver hisses (which is mainly high frequencies) but receivers usually mute this.

In addition on the website there is no mention of an antenna socket, so I can only assume they have a telescopic rod or a wire antenna particularly on battery operated radios. One of the characteristic of FM is that the distortion rises and the channel separation reduces (can't hear that on a mono speaker). Telescopic rod antennas are generally omnidirectional and when inside often receive reflected signals. Who can remember the multiple images on analog TV with an indoor antenna and trying to adjust it to get rid of them.

So using his Roberts radio speaker he listens to a mono sound even if he has one with twin speakers they are so close together the stereo effect is minimal.

As for DAB+ it has Spectral Bandwidth Replacement to replace the removed high frequencies. The speaker in this radio is not capable of these frequencies anyway, in addition the speaker is not big enough to produce low pitched bass to balance the high frequencies, so the radio sounds comparitively tinny.

Not only that you cannot hear the parametric stereo on a mono speaker.

Since MLXXX promotes AM radio but makes no lament of the lack of stereo, here is the reason, he doesn't listen to stereo on any other broadcasts either.

I am glad to say that listening to DAB+ on a decent sound system sounds great particularly with the programs not available on analog.

Alanh

So I take it you still haven't read the paper?

If you have particular problems with the paper under discussion, present them. At the moment, you're just ranting at imaginary clouds...

Edited November 4, 2016 by Malich
(include quote)

[MLXXX]

1 minute ago, alanh said:

So using his Roberts radio speaker he listens to a mono sound.

????

Alanh, my Roberts ecolologic 4  DAB+ radio has stereo speakers. Do you have no memory of that?  It's been mentioned many times over the years. Also my comparisons have not been limited to listening to that radio with its internal speakers. I have listened using the headphone output connected to a good quality hi-fi system, and I've listened with other DAB+ radios similarly connected, and with a DAB+/DVB-t USB stick with a direct interface. All of that has been mentioned before.

However this thread is not about me. It is about the paper. I suggest you address the contents of the paper if you wish to comment, rather than trying to question my listening methods.

[hrh]

7 hours ago, alanh said:

Just keep in mind that 3.5 million Australians are listening to Digital Radio and the number is rising

One of the reasons for digital is that it offers alternative programming to the analog besides the parallel - such as ABC local. When there is sport on the AM or local variant I often listen to what would normally be on there (the AM station) on the digital version. Other offerings not available on either FM or AM are those such as Smooth, Buddha, ABC Extra, etc. I would suggest that it is these that that are having a sizeable affect with regard to the digital market. And the automotive manufacturers slowly coming on board helps too.

7 hours ago, alanh said:

and the number is rising despite what  MLXXX says.

I give in, did MLXXX say that digital listener numbers were declining?

 

6 hours ago, alanh said:

when commercial FM radio started in Australia all FM stations added the 38 kHz Double Sideband Suppressed carrier carrying the difference signal between Left and Right microphones

So are you suggesting that up until the advent of commercial FM radio in Australia none of the already established community broadcasters and ABC FM were not broadcasting in stereo? Stereo FM was being broadcast in Australia on community stations long before commercial ones made the switch from AM or started new on FM .
I remember listening to 6UVS when it was the only FM station in Perth, in the mid 70s, and hearing it in stereo and being rather amazed.

Edited November 4, 2016 by hrh

[alanh]

hrh,

It is not true that most people are listening to the DAB+ only programs http://www.digitalradioplus.com.au/CMSPages/GetFile.aspx?guid=aa94cd80-03ee-44ac-afe3-8cb1a48b446a remember that the 3 hour time difference causes the ABC Local Perth to have more separate programming due to the cricket and AFL. For example Perth has half an hour on the end of the Drive program which is unique to DAB+. They now fill that time with repeat content with the presenter continuing their shift which finishes at 6:00 pm. I do however recommend on Saturday "It's Just Not Cricket" which is broadcast live on all Mainland Capitals on DAB+ but is also transmitted in some high powered  ABC Local FM radio stations in Queensland in winter.

The ABC metropolitan and Radio National AM radio transmits in stereo on DAB+. Their presenters are in mono, but the music and some of RN productions are in stereo, there is also high frequencies and sounds better particularly the half a million cars which contain DAB+ radios.

I was not trying to exclude the community stations but to exclude the ABC transmitted an FM program in Sydney, Melbourne and Adelaide for many years prior to the transfer of the FM to TV channels 3 - 5.  This forced this mono service to be shut down.This occurred in the 1960s when they discovered that TV was much more popular than the planners thought.

The radio industry had local manufacturing and the broadcasters did not wish to embrace new technology. There is no receivers and listeners. Then AM/FM receivers importing started to flood the market. So when the Government offered 14 MHz of free spectrum with an immunity from new commercial broadcasters they decided not to repeat history and started DAB+ broadcasting. This same thinking has been happening in TV and with DRM radio for country areas.

Alanh

 

Edited November 5, 2016 by alanh
Stereo in cars

[hrh]

1 hour ago, alanh said:

It is not true that most people are listening to the DAB+ only programs http://www.digitalradioplus.com.au/CMSPages/GetFile.aspx?guid=aa94cd80-03ee-44ac-afe3-8cb1a48b446a 

I didn't say most - just sizeable. We could argue the semantics along with statistics (as per "How To Lie With Statistics") to our own end.

1 hour ago, alanh said:

remember that the 3 hour time difference causes the ABC Local Perth to have more separate programming due to the cricket and AFL. For example Perth has half an hour on the end of the Drive program which is unique to DAB+. They now fill that time with repeat content with the presenter continuing their shift which finishes at 6:00 pm

That last half hour of the drive program to which you refer is typically during the footy (non-day light saving) season and only on the evening games and I can't say that I have heard it filled with repeat content as such, rather it just continues from where it was. However the current drive presenter drives me up the wall (pardon the pun) so I tend to not listen to my ABC during that period.

1 hour ago, alanh said:

I do however recommend on Saturday "It's Just Not Cricket"

Now this is one thing on which I whole heartedly concur!

[Guest Malich]

So... ignoring the irrelevant derail & incorrect assumptions about MLXXX's DAB+ receiver (both of which the mods seem disinclined to delete despite at least one request), let's consider the rest:

22 hours ago, alanh said:

Sound quality should not be compared with FM but to the origninal source

This is covered in the linked paper.

 

22 hours ago, alanh said:

FM has 15dB high frequency boost prior to modulation

It does - and a corresponding cut is made in the receiver. The end effect is a reduction in audible noise arising from the transmission path, which increases with increasing audio frequency (i.e. FM deviation rate, since to the demodulator random RF noise looks like fast deviations in signal).

 

22 hours ago, alanh said:

So broadcasters use processors  which measure the signal level in each octave band and reduce the gain at high levels. So for example a symbol crash will have lots of its high frequency level reduced prevent overmodulation thus dulling the sound.

Dynamic Range Compression (DRC; to distinguish it from the 'lossy' data compression used in digital radio) isn't really done to prevent overmodulation though (at least not in a well-lined-up broadcast chain) - it's done to (a) ensure any low-amplitude audio (e.g. soft parts of classical music) can be heard at the receiver in a typical listening environment, and (b) add 'punch' to rock/pop/etc so it sounds 'louder' (and, to many ears, 'better') than competing stations.

The amount of DRC used varies from station to station, generally in line with the programme content - classical stations tend to use little (just enough so that soft passages are audible & the loudest passages don't result in excessive deviation), while rock/pop/etc & commercial stations use a lot (so that both 'soft' & 'loud' sounds are closer in level to a specified average). The latter is the broadcast equivalent of the "CD loudness wars" (which itself pre-dates CDs by a long way!).

 

22 hours ago, alanh said:

This is not required in audio systems or any digital radio system.

Granted, it's not required in any digital audio systems - but it is still "required" in digital radio systems as much as it ever was in analogue, and for the same reasons: to ensure low-amplitude signals can be heard in a typical listening environment, and/or to add 'punch' to rock/pop/etc & commercial stations.

 

22 hours ago, alanh said:

Where are the graphs of the MUSHRA Scores to show the sound quality against the bit rates under standardised double blind testing?

They are in the linked paper (along with the raw results, & explanation of where the significant differences were found).

 

19 hours ago, alanh said:

Also the type of program has a great effect, speech is different to music.

This is also covered in the linked paper. In fact, that's where the really interesting & surprising results are...

 

20 hours ago, alanh said:

As I said you need to compare both back to the origiinal at least to a lossless format such as a CD or .WAV file.

And, as I said near the top of this comment, "This is covered in the linked paper".

The upshot is they determined that, across a range of music types, 'average' broadcast-quality FM (and, admittedly, they set a high bar for 'average' quality) was slightly better than, but not significantly different statistically to, 96kbps DAB+ (using HE-AACv2). 'Good' quality FM was slightly worse than, but again not significantly different statistically to, 192kbps DAB+ (again, using HE-AACv2).

On non-music content, however ... well, I hinted earlier that was the interesting bit. Go read the paper for yourselves .

xHE-AAC is irrelevant - the extensions that come into play for that have no effect above ~96kbps (for music), and at lower bitrates still don't make it any better than ~96kbps HE-AAC.

The only other comment I'll make is that typical DAB+ bitrates in Australia (outside of a few anomalies) are well below what is required to match even 'average' quality FM, and in most cases are half (48kbps) or less.

[alanh]

This post sure is about your. Your first post in this strand continues your campaign against DAB+ using your ears and the never mentioned before your Roberts ecolologic 4 radio. The speakers are so close together there is no stereo effect. The distance between stereo speakers should form 60 degrees between the speakers and the listeners. Its amplifiers and speakers are not in the class of the types I mentioned above or my amplifier system and speakers, so in effect you are judging  DAB+ sysgtem on your radio as the performance of your radio in acoustics which are unknow along with the quality of your hearing.

Your comments on FM radio in the car does not compare with a DAB+ car radio under the same conditions. You do not say where you are driving. What about through the CBD of Brisbane where there are plenty of reflected signals? At least the audio amplifiers and speakers in built in DAB+ car radios are generally bigger and more powerful.

You cannot quote a summary without reading the whole article, to find out how they came to those conclusions. You can google all you like, but the quote you used came from a pressure group from Swedish Radio which is trying to prevent digital radio. They don't wish to loose money whilst the audience buys receivers. I know this because they were emailing me. The neighbouring Norway is switching off FM starting next year.

Your experience with horizontally polarised TV antennas for DAB+ is irrelevant because you have lots of signal. Obviously the telescopic antenna in your radio in side is enough, after all Mt Coot-tha is line of sight to the Bald Hills area. I know this because I have seen it on a number of occasions. It does matter when the signals are weak like near the edges of the coverage area and behind terrain.

Coding Technologies which has been taken over by Dolby, did extensive double blind testing when developing AAC compression to minimise the auditory effects of reducing the bit rates. The comments make no mention of the compression types used as MP-3 compression used in DAB is not the same as AAC used in DAB+ for a start.

So when are you going to compare an original CD, FM broadcast of the same pieces and a DAB+ broadcast of those peices on a broadcast quality receiver/amplifier speaker system in good acoustic surroundings instead of a portable low powered radio with small closely spaced speakers or even worse computer speakers. Put a request in to 4MBS.com.au for a CD you own. They have DAB+ and FM programs streams, or even better listen go to the station and listen to their off air receivers.

Alanh

[MLXXX]

3 hours ago, alanh said:

Your first post in this strand continues your campaign against DAB+ using your ears and the never mentioned before your Roberts ecolologic 4 radio.

Alanh, I suggest you peruse the list of forum references to my DAB+ radio using the search http://www.dtvforum.info/index.php?/search/&q="ecologic 4" and the further set of references to it using the search http://www.dtvforum.info/index.php?/search/&q=ecologic4  

I will not comment on the rest of your post as in my opinion it is mostly off point. My purpose in starting this thread was to draw attention to the research paper. If you disagree with any aspect of that paper (which can be downloaded as explained by Malich), please explain where you disagree.   

It really is quite simply the case that 64kbps (nominal) DAB+ is generally not as good to listen to as good quality FM radio. The paper provides evidence prepared by well qualified academics. [That also happens to be my own personal experience listening on good quality hi-fi equipment but let's not dwell on me.] Please refer to specifics in the paper, if you feel it is flawed. 

Edited November 13, 2016 by MLXXX

[alanh]

Malich,

The 50 microsecond preemphasis high frequency boost is compensted for by a similar high frequency cut a the receiver.

The sound signal is applied to a sound processor at the studios  mostly digitised, sent to the transmitter where sound signal is converted to analog, pre-emphasised then limited prior to modulation.

Processsors for FM broadcasting apply pre-emphasis to the signal prior to measuring of the level in each frequency band. This causes more compression of high frequencies than the lower frequencies. It is done this way to prevent the limiter in the transmitter from operating, because if it does, all frequences are reduced in level. The effect of this is the louder sounds with high frequency content become less sharp because the high frequencies have been reduced compared to the original sound.

This is not compensated for at the receiver because it does not know how much compression was applied in the processor. At least in DAB+/DRM the amount of compression is transmitted to the receiver, and the listener can then restore the original dynamic range if they wish. FM radio does not give this option.

It is FM station policy as to how much compression, in each band is applied, the attack and decay times and trigger levels.

AAC transmits the audio with a small dynamic range and add the DRC signal, so it is up to the listener as to how much compression is heard rather than the station. The end result is that the available dynamic range available on DAB+/DRM is determined by the system design and is unaffected by received signal strength. As we know digital transmission is either error free or non existant, not noisy. The AAC encoder must not be fed with the output of an FM compressor because the sound quality has been reduced and it also affects any SBR which was designed to relate harmonics in the original sound not one which has been altered by the processors variable characteristics.

By comparision FM has a smaller signal to noise ratio, which gets worse in stereo under weak signal conditions. (The left-right signal is double sideband supressed carrier modulated in the noisiest part of the demodulated spectrum and DSBSC gives no noise protection) Many FM receivers measure the supersonic noise level and blend the left and right signals and under the worst conditions make the signal mono. This does not occur in digital radio.

So as I say any quality comparisons should not be made to FM which is altered even under the best reception conditions but to the original sound either live or at least recorded using AES lossless 48 kHz per channel sampling standard.

Alanh

[hrh]

22 minutes ago, alanh said:

any quality comparisons should not be made to FM which is altered even under the best reception conditions but to the original sound

 But then considering the title, yes it should!

 

1 hour ago, alanh said:

your Roberts ecolologic 4 radio. The speakers are so close together there is no stereo effect. The distance between stereo speakers should form 60 degrees between the speakers and the listeners. Its amplifiers and speakers are not in the class of the types I mentioned above or my amplifier system and speakers, so in effect you are judging  DAB+ sysgtem on your radio as the performance of your radio in acoustics which are unknow along with the quality of your hearing

Do we know what system MLXXX ACTUALLY listens to the product on? Other than saying what the receiver is, it could be connected to a broadcast quality system.
Seriously, how many people would have their HI-FI setup as per the ideal listening positions? You yourself haven't said what yours is other than "Its amplifiers and speakers are not in the class of the types I mentioned above or my amplifier system and speakers".

2 hours ago, alanh said:

in acoustics which are unknow along with the quality of your hearing

Isn't this how we mostly do it? Who has access to an acoustically treated room for specific audio comparisons? And when it comes to hearing, that is extremely subjective.
I know my hearing has deteriorated with age (tinnitus doesn't help either) as I haven't been able to hear the squeal from the flyback transformer in a CRT telly for a long time unless I got right up to it - I used to be able to hear one from another room away. But because I know some of the things to listen for when comparing compression rates etc I can still hear the differences between the digital and analog - ie DAB+ vs FM. That's even without plugging my DAB+, AM, FM radio into my HI-FI which isn't too shabby either.

[Guest Malich]

1 hour ago, alanh said:

Malich,

The 50 microsecond preemphasis high frequency boost is compensted for by a similar high frequency cut a the receiver.

The sound signal is applied to a sound processor at the studios  mostly digitised, sent to the transmitter where sound signal is converted to analog, pre-emphasised then limited prior to modulation.

Processsors for FM broadcasting apply pre-emphasis to the signal prior to measuring of the level in each frequency band. This causes more compression of high frequencies than the lower frequencies. It is done this way to prevent the limiter in the transmitter from operating, because if it does, all frequences are reduced in level. The effect of this is the louder sounds with high frequency content become less sharp because the high frequencies have been reduced compared to the original sound.

This is not compensated for at the receiver because it does not know how much compression was applied in the processor. At least in DAB+/DRM the amount of compression is transmitted to the receiver, and the listener can then restore the original dynamic range if they wish. FM radio does not give this option.

It is FM station policy as to how much compression, in each band is applied, the attack and decay times and trigger levels.

AAC transmits the audio with a small dynamic range and add the DRC signal, so it is up to the listener as to how much compression is heard rather than the station. The end result is that the available dynamic range available on DAB+/DRM is determined by the system design and is unaffected by received signal strength. As we know digital transmission is either error free or non existant, not noisy. The AAC encoder must not be fed with the output of an FM compressor because the sound quality has been reduced and it also affects any SBR which was designed to relate harmonics in the original sound not one which has been altered by the processors variable characteristics.

By comparision FM has a smaller signal to noise ratio, which gets worse in stereo under weak signal conditions. (The left-right signal is double sideband supressed carrier modulated in the noisiest part of the demodulated spectrum and DSBSC gives no noise protection) Many FM receivers measure the supersonic noise level and blend the left and right signals and under the worst conditions make the signal mono. This does not occur in digital radio.

So as I say any quality comparisons should not be made to FM which is altered even under the best reception conditions but to the original sound either live or at least recorded using AES lossless 48 kHz per channel sampling standard.

Alanh

Alan, this thread is not about you or any special "knowledge" you may claim to have. It's about articles & a research paper comparing FM with DAB+.

(But allow me to indulge you a bit: Your seem to be claiming that, since CDs (digital) sound better than FM, then digital DAB+ must therefore also sound better than FM. Is that pretty much a correct summary of your position? You've already stated your reasoning, so a simple yes/no answer will do.)

I suggest once again you read the paper, because it seems that you haven't...

Edited November 13, 2016 by Malich

[alanh]

hrh, If the original sound is not the reference, as you say, otherwise DAB+ could be accused of having too much high frequencies in loud studio sound, because of the audio processors in FM broadcasts.

MLXXX is making a comparison between the systems of FM multiplex and DAB+ on the basis of one radio and a pair of headphones.

Portable radios inparticular modify the sound depending on the design of each brand and model of radio. As a result each person will have a different opinion. Speakers along with the case have resonances which will emphasise different frequencies from model to model .

A different result can be found in the 307,500 cars sold last year, as the speakers and amplifiers are generally a lot better than portable usually battery operated radios.

The following common new cars contain DAB+ radios https://www.holden.com.au/about/technology-and-design/mylink/insignia http://www.ford.com.au/sync/package, http://www.toyota.com.au/camry/features/audio-and-navigation Where is his test of this range of radios where the speakers are atleast well spaced.

Malich,

I am not claiming that FM is better or worse than DAB+. The article required bit rates higher than any broadcaster can ever aspire to use. Quality comparisons should use an independent source which is the original sound which is common to both systems.

Considering that all radio stations use live, AES standard digital audio and CDs as source material, that should be the reference. Both FM and DAB+ both downgrade the sound quality in different ways. After all the AES3 samples each channel at 48 kHz 20 bit/s = 960 kbit/s per channel (audio only) so lossy systems are required even between the studios and the FM transmitter.

I cannot read the complete text, it is not available, however I have read the graphs of degredation vs bit rate for various encoding systems used in the development of AAC. I have looked for a link, which has been removed. The curves flatten out greatly as imperceptible difference betwen the source and the compressed signal as well as quality assessments, particularly when large numbers of untrained listeners are used, with multiple samples in random order so that the results can be double checked against each other. http://www.iis.fraunhofer.de/content/dam/iis/de/doc/ame/wp/FraunhoferIIS_Technical-Paper_xHE-AAC.pdf page 6 https://www.ietf.org/lib/dt/documents/LIAISON/file1298.doc

http://www.iis.fraunhofer.de/content/dam/iis/de/doc/ame/conference/AES-122-Convention_AAC-ELD_LowBitrateHighQualityCommunication_AES6998.pdf

"2.3. MPEG-4 HE-AAC The next milestone in MPEG-4 towards low bit rate coding was the introduction of SBR, a generic parametric coding tool for high frequencies. The combination of SBR and AAC-LC was standardized in 2003 in the MPEG-4 High-Efficency (HE-AAC) and achieves FM quality at bitrates as low as 16 kbit/s per channel. In order to limit the perceptible coding artifacts of common audio coding systems to a subjectively acceptable level, the entropy of the source has to be limited and the coding gain has to be optimized." My bolding.

DAB+/DRM should not emulate FM but the origintal sound. What matters is an accurate reproduction of what happened at the microphone as all transmission systems will degrade the sound in different ways which becomes more complicated by the different types and models of receiver and the listening environments. All battery operated radios and car systems require compression due to the lack of power output of battery operated radios and for both situations the level of background noise in the listening environment.

Alanh

 

[MLXXX]

11 minutes ago, alanh said:

MLXXX is making a comparison between the systems of FM multiplex and DAB+ on the basis of one radio and a pair of headphones.

Incorrect, alanh. As I have already said in this thread:-

On 04/11/2016 at 8:03 PM, MLXXX said:

Also my comparisons have not been limited to listening to that radio with its internal speakers. I have listened using the headphone output connected to a good quality hi-fi system, and I've listened with other DAB+ radios similarly connected, and with a DAB+/DVB-t USB stick with a direct interface. All of that has been mentioned before.

 

11 minutes ago, alanh said:

I cannot read the complete text, it is not available,

Simply follow the steps Malich has outlined, alanh! I was able to download the paper that way, and read it in full.  As I have already stated in this thread:-

On 04/11/2016 at 6:45 PM, MLXXX said:

If you read the report alanh you would see that the FM2 source involved standard processing, use of an FM exciter, and then demodulation with an FM receiver. Despite that, the digital methods required very high bitrates for comparable MUSHRA scores, much higher than digital bitrates in common use for DAB+.

[Guest Malich]

31 minutes ago, alanh said:

The article required bit rates higher than any broadcaster can ever aspire to use.

...

I cannot read the complete text, it is not available, however ...

Alan, even though my instructions above are still valid, I've just sent a copy of the paper directly to your publicly-available email address. Please read it.