Jump to content

What bitrate is needed to sound like analog FM?


MLXXX

Recommended Posts

48 minutes ago, alanh said:

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.

Unfortunately, alanh, what was meant by "FM quality" stated above was not defined.  I have to conclude it was FM reception quality under very poor reception conditions. As established in the testing reported in the 2013 paper the subject of this thread, much higher HE-AAC bitrates are actually required for the sound to be of comparable subjective quality to a good broadcast quality FM (FM2 in the paper).   I repeat what I said in the first paragraph of this thread:-

On 11/4/2016 at 0:45 AM, MLXXX said:

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.

The 2013 paper actually explains how the FM sources it tested were created. To my mind the FM2 source described in the paper would be likely to be broadly comparable to the FM quality delivered in my home, in a metropolitan reception zone, with a good antenna, and using a mid-price AVR FM tuner.  (Background hiss is extremely low. There is no impulse noise. And there are no apparent multipath effects.)

Link to comment
Share on other sites



  • Replies 56
  • Created
  • Last Reply

Top Posters In This Topic

3 hours ago, hrh said:

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.

Yes, the built-in stereo speakers of even a modest DAB+ radio reveal a lot. They readily reveal the much more extended treble than typical AM radios provide. And they readily reveal the use of psychoacoustic artefacts when a lower bitrate is in use, such as 48kbps nominal for stereo DAB+. Connecting the headphone output of such a radio to a hi-fi system can be expected to reveal the very extended bass response typical of DAB+ radio. It may also highlight a blandness in the sound compared with an FM version of the same station.

In short I would agree that just using the built-in speakers is enough to reveal considerable differences.

Link to comment
Share on other sites

There is a few problems with this report

1. Even now there is no DAB+ broadcasts in Sweden https://www.worlddab.org/system/news/documents/000/005/880/original/Digital_radio_in_Sweden_12.02.2016.pdf?1455287990

2. http://downloads.bbc.co.uk/rd/pubs/whp/whp-pdf-files/WHP006.pdf shows that DRC enables a wide dynamic range, which the FM multiplex does not. This was not demonstrated.

3. The DAB+ signals should not use any audio processing where as FM must have it. This was not tested. As a result DAB+ with DRC removing the compression would allow a wide dynamic range which is greater than for FM particularly near the edge of the coverage area. Multiband processing should not be used for DAB+ because the AAC encoder also contains a similar multiband filter set so if they are not exactly the same it will cause weird effects and wrecks the psychoacoustic system to minimise the effect of leaving lots of data outl

4. They did not transmit the FM signal, which is important because Sweden like Norway is a mountainous country thus is more likely to downgrade the quality of the received signal.

5. Similarly they did not compare a DAB+ and and FM multiplexed signal in real world RF conditions which the listeners must endure.

6. Considering that http://www.iis.fraunhofer.de/en/ff/amm/prod/digirundfunk/digirundf/xheaac.html comes from

"The Fraunhofer-Gesellschaft is the leading organization for applied research in Europe. Its research activities are conducted by 67 institutes and research units at locations throughout Germany. The Fraunhofer-Gesellschaft employs a staff of 24,000, who work with an annual research budget totaling more than 2.1 billion euros.

The Fraunhofer Institute for Integrated Circuits IIS is one of the world’s leading application-oriented research institutions for microelectronic and IT system solutions and services. It ranks first among all Fraunhofer Institutes in size. With the creation of mp3 and the co-development of AAC, Fraunhofer IIS has reached worldwide recognition.

In close cooperation with partners and clients the Institute provides research and development services in the following areas: Audio & Multimedia, Imaging Systems, Energy Management,

IC Design and Design Automation, Communication Systems, Positioning, Medical Technology, Sensor Systems, Safety and Security Technology, Supply Chain Management and Non-destructive Testing.

About 950 employees conduct contract research for industry, the service sector and public authorities. Founded in 1985 in Erlangen, Fraunhofer IIS has now 13 locations
in 10 cities: in Erlangen (headquaters), Nuremberg, Fürth, Dresden, further in Bamberg, Waischenfeld, Coburg, Würzburg, Ilmenau and Deggendorf. The budget of 130 million euros is mainly financed by projects. 22 percent of the budget is subsidized by federal and state funds."

So Fraumhofter has the ability to design including application specific integrated circuits) and optimise performance of compression systems when compared to a single paper from a single university using equipment from 2007.

MLXXX,

It is still about you. Your tirade about listening to AM from ABC radio from a very close transmitter using unavailable AM radios except they old one you have. So now you admit that there is more high frequencies from DAB+. Still no admission about the improvement made by stereo sound over AM's mono only.

So when are you going to a radio station which transmits FM and DAB+ and comparing the uncompressed unprocessed original sound and the off air received sound, where you don't know what the source is. Just like the study you quoted.

You rave on about artifacts that you are supposed to hear but cannot describe them other than blandness of the sound. What is that supposed to mean?

Lastly there is no actual definition of FM quality audio, the broadcasters were deregulated when the very old Australian Broadcast Control Board used to inspect radio stations. A engineer friend of mine used to have to do the inspections. In those days the ABC had the only the only 3 FM transmitters.The ACMA predecessor had http://www.acma.gov.au/~/media/Licence Issue and Allocation/Publication/pdf/TVRadio_Handbook_geninfo pdf.pdf. The current broadcast law makes no reference to any analog technical standards https://www.legislation.gov.au/Details/C2016C00882 

UK https://www.ofcom.org.uk/__data/assets/pdf_file/0017/37133/code2013.pdf you will notice that it is all about coverage and lack of interference, and transmitter performance. No mention of studio equipment including processing!

DAB+ has a transmission standard in Australia https://infostore.saiglobal.com/en-au/Standards/AS-4943-1-2009-1094294/ AS 4943.1—2009 Digital radio—Terrestrial broadcasting
Part 1:  Characteristics of terrestrial digital audio broadcasting (T-DAB+) transmissions  Again this does not cover studios.

Lastly the AAC enocoder feeding the transmitter must be the only lossy digitiser between the microphone and the transmitter input. The interaction between compression systems in series is disasterous to the sound quality.
 

Alanh

Link to comment
Share on other sites

56 minutes ago, alanh said:

shows that DRC enables a wide dynamic range, which the FM multiplex does not. This was not demonstrated.

I don't think DRC is used in practice in Australia. At least it wasn't in the first few months of transmissions when I experimented a great deal with different settings. Is it used now?  It would be of greater benefit for classical music. Pop music and talk shows typically use very little dynamic range.
 

56 minutes ago, alanh said:

They did not transmit the FM signal, which is important because Sweden like Norway is a mountainous country thus is more likely to downgrade the quality of the received signal.

Well the report is about broadcast quality FM in good conditions. It does not explore poor quality FM reception (e.g. loss of stereo) or poor quality DAB+ reception (e.g. echo effects at the digital cliff). In metropolitan areas people can expect a good FM signal strength and can use an external antenna if they want the best reception. I suggest that that these are the type of listening circumstances to which FM2 source quality would apply.

 

56 minutes ago, alanh said:

So Fraumhofter has the ability to design including application specific integrated circuits) and optimise performance of compression systems

 Fraunhofer may be able to optimise performance of compression systems but they cannot perform miracles. Low bitrate codecs are a compromise. The lower the bitrate, the greater the compromise, all other things being equal.

 

56 minutes ago, alanh said:

about listening to AM from ABC radio from a very close transmitter

The ABC transmitter mast is over 12km away from my residence. I can't run a loudspeaker but I can just run headphones from a loop antenna and diode!  I'm looking forward to the Titus II radio -- discussed in another thread -- which may well be able to provide wideband AM demodulation capability, and at a modest price.

 

56 minutes ago, alanh said:

So now you admit that there is more high frequencies from DAB+. Still no admission about the improvement made by stereo sound over AM's mono only.  ... You rave on about artifacts that you are supposed to hear but cannot describe them other than blandness of the sound. What is that supposed to mean?

Alanh, you memory must be poor. I have never hesitated to concede on this forum that typical AM radios have a very limited treble performance. I have never hesitated to concede that stereo is better than mono (all other things being equal). As for describing the subjective sound of the codec artefacts I have done that in detail in the past on this forum (as have others). I note that you could not even recall my mentioning I had a Roberts ecologic 4 radio. [See search links in a post of mine above for the many occasions on which I have referred to the radio on this forum.]

 

56 minutes ago, alanh said:

Lastly there is no actual definition of FM quality audio, the broadcasters were deregulated when the very old Australian Broadcast Control Board used to inspect radio stations. A engineer friend of mine used to have to do the inspections. In those days the ABC had the only the only 3 FM transmitters.

What I was after was something that could be used for international comparisons of reception circumstances, such as a reference to the quieting level achieved, and the THD, in order to provide a broadly defined audio quality standard against which to compare a psychoacoustic codec.

Here are very broad outlines of the type of categories of "FM quality" that I have in mind. 

1. At a fixed location in a metropolitan area with a high signal strength, negligible interference, negligible impairment from multipath effects, and a high quality demodulator contributing no more than x% THD in stereo with one channel driven at 1kHz at -6dB and the other channel silent.

2.  ..

3. ...

4. At the outer limit of the reception zone, with a reversion to mono, and a weighted audio signal to noise ratio of 15dB. 

Edited by MLXXX
Link to comment
Share on other sites

Whether Brisbane broadcasters use DRC is not the fault of the designer of HE AAC V2 as is the selection of the bit rate, any prior volume compression and sound quality.

As I said you need to listen to source material, many pop recordings have dynamic range which the broadcasters remove. Expanders have been available for expanding the limited dynamic range, however without knowing what characteristics were used in compressing the origninal signal the results are quesionable particularly witn multichannel processing is used. The broadcasters including ABCFM all use compression, to ensure that the program can be heard they are just less aggressive about it.!.

Sure it's about optimum transmission, they didn't even use a transmitter. Their setup is wrong as I have already pointed out.AAC signals should not be processed prior to encoding. The multichannel processor settings were not outlined, and every broadcaster uses different values changing the sound and the effect on DAB+ vs FM.

The encoder software used was from 2007. http://www.dolby.com/us/en/professional/pc-and-mobile/products/dolby-media-generator.html talk of recent updates.I wonder how many updates they have had in 9 years including the use of faster processing in newer microprocessors. For example faster computers can use more samples in the comb filters which will improve the sound.

FM transmissions are not like digital. FM will degrade particularly for the left-right channel the hiss increases before the mute is applied by the receiver to prevent the loud hissing when the signal drops below limiting, where as DAB+ when decoded is always of the design quality including signal to noise ratio.

Fraunhoffer are some of the designers of the system not the Swedish University who used a very old version of the encoder from Microsoft. All compression is a compromise.

I am not going to commit to memory which DAB+ radio you bought in 2009! What an ego!

You never mention any stereo characteristics and you mention old TRF. I wonder if you have poor hearing in at least one ear.

I know of other hifi gurus who have had their hearing checked, I think they wanted to set an equaliser to the inverse of their loss!

 

Alanh

 

Edited by alanh
complete post.
Link to comment
Share on other sites



2 minutes ago, alanh said:

A search for your radio on this site comes up empty.

Well I'll give you the search links again, alanh. You need to look down the page in the results section:

8 hours ago, MLXXX said:

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'll look at the rest of your most recent post later and may possibly comment.]

Link to comment
Share on other sites

The whole post is pretty irrelevant really.

There is no standard for audio performance for FM broadcasters here, UK or USA. The maximum deviation is +/- 75 kHz and the channel width is 200 kHz. . All we know is that the frequency range maximum is from 30 Hz to 15 kHz and that all but the Americas use 50 microseconds pre-emphasis. The Americas use 75 microseconds which requires more high frequency cut in the receiver. Studios are commonly digital from the audio desk to the transmitter site..

How do you propose to compare the total harmonic distortion and transient intermodulation distortion of an analog system to a digital system. Digital systems are not designed to carry constant pure tone required for testing and it is meaningless anyway. All lossy compression systems have to be subjectively assessed, so to be equal the analog systems have to be tested the same way. The only common measurement is the line up tone levels to ensure signals do not exceed the Analog to Digital and digital to analog converter voltage ranges. Also to try and make the volume between stations and between programs approximately the same.

You could measure the signal to noise ratio of both systems.

You probably could also measure left to right crosstalk

Frequency response of the main signal and the SBR is specified in the standards for each standard bit rate. SBR will not work on pure tones, because it needs a main channel frequency to make a harmonic from, so testing for 12 kHz will not work.

 

Alanh

 

Link to comment
Share on other sites

12 hours ago, alanh said:

I am not going to commit to memory which DAB+ radio you bought in 2009! What an ego!

I didn't expect you to recall the model number off-hand, but I did expect you to remember that the first DAB+ radio I purchased had stereo speakers. This was a matter mentioned more than once in my many discussions with you on this forum about the radio. 

 

12 hours ago, alanh said:

You never mention any stereo characteristics and you mention old TRF. I wonder if you have poor hearing in at least one ear.

Again alanh you have a poor memory. I have commented on stereo many times. I can specifically recall mentioning hearing stereo DAB+ artefacts. And no there was nothing wrong with my binaural hearing at the time when DAB+ sound quality was being earnestly discussed on this forum in the months following the start of transmissions. (Other readers may recall I made comparative DAB+ and FM radio stereo recordings and uploaded them with links for forum members.)

I still don't understand your preoccupation with my personal hearing.

 

12 hours ago, alanh said:

FM transmissions are not like digital. FM will degrade particularly for the left-right channel the hiss increases before the mute is applied by the receiver to prevent the loud hissing when the signal drops below limiting, where as DAB+ when decoded is always of the design quality including signal to noise ratio.

Well alanh, it appears you have at long last resiled from your former very firm position that FM does not gradually degrade! I refer to this statement of yours that started some quite intense discussion on that topic:

'alanh', on 15 Sept 2011 - 2:21 PM, said:

No phones are capable of AM reception, FM radio does not gradually degrade, it has a cliff like digital. As soon as the reception strength is below the limiting in the receiver, either the hiss makes reception unusable or a mute circuit operates.

[I note that the above original post is no longer visible, being a post lost in the move to the current platform.]

I would agree with your currently expressed view that FM will degrade by way of increasing levels of hiss.

 

12 hours ago, alanh said:

Whether Brisbane broadcasters use DRC is not the fault of the designer of HE AAC V2 as is the selection of the bit rate, any prior volume compression and sound quality.

DRC is neither here nor there. It has little impact with typical program material, which is why it is not typically used. As for the bit rate, that is the very factor that the paper was examining - the bitrate needed to bring DAB+ up to a similar level of minor subjective impairment to good quality FM.

12 hours ago, alanh said:

Their setup is wrong as I have already pointed out.AAC signals should not be processed prior to encoding.

But in practice they typically are. That is part of what was being examined: typical AAC processing used in broadcasting and typical FM processing. Of course what is "typical" is not an exact thing. Practices will vary to some extent.

 

12 hours ago, alanh said:

The encoder software used was from 2007. http://www.dolby.com/us/en/professional/pc-and-mobile/products/dolby-media-generator.html talk of recent updates.I wonder how many updates they have had in 9 years including the use of faster processing in newer microprocessors. For example faster computers can use more samples in the comb filters which will improve the sound.

I imagine if there had been material advances then later software would have been used. I'm not aware of any firmware updates for DAB+ decoders in radios. It's still the same specification. If you can point us to two examples alanh, one encoded at a particular bitrate with 2007 software and the other the same material encoded at the same bitrate and parameters but with later encoding sofware, then your point could be explored further. Are you aware of any claims as to the extent of subjective audible improvement (if any) resulting from encoder software updates? 

Edited by MLXXX
clarity
Link to comment
Share on other sites

53 minutes ago, alanh said:

The whole post is pretty irrelevant really.

There is no standard for audio performance for FM broadcasters here, UK or USA. The maximum deviation is +/- 75 kHz and the channel width is 200 kHz. . All we know is that the frequency range maximum is from 30 Hz to 15 kHz and that all but the Americas use 50 microseconds pre-emphasis. The Americas use 75 microseconds which requires more high frequency cut in the receiver. Studios are commonly digital from the audio desk to the transmitter site..

How do you propose to compare the total harmonic distortion and transient intermodulation distortion of an analog system to a digital system. Digital systems are not designed to carry constant pure tone required for testing and it is meaningless anyway. All lossy compression systems have to be subjectively assessed, so to be equal the analog systems have to be tested the same way. The only common measurement is the line up tone levels to ensure signals do not exceed the Analog to Digital and digital to analog converter voltage ranges. Also to try and make the volume between stations and between programs approximately the same.

You could measure the signal to noise ratio of both systems.

You probably could also measure left to right crosstalk

Frequency response of the main signal and the SBR is specified in the standards for each standard bit rate. SBR will not work on pure tones, because it needs a main channel frequency to make a harmonic from, so testing for 12 kHz will not work.

 

Alanh

 

I think you may have missed the point I was trying to make. I was not trying to "compare the total harmonic distortion and transient intermodulation distortion of an analog system to a digital system". I was trying to establish the quality level of the FM reception in broad terms. By your own admission, FM does degrade depending on reception circumstances. So when we try to compare the subjective (MUSHRA based) degradation of FM with the subjective (MUSHRA based) degradation of HE-AAC at lower bitrates, as assessed by a panel of human listeners, it becomes relevant to pin down what particular "FM quality" reception circumstances we are referring to.

Is it pristine "FM quality" reception, such as with a high quality tuner connected to a good antenna with high signal strength and negligible multipath or other interference? [That in my view is what the FM2 source in the paper approximated.] Or is it FM reception in a moving car at the outer edges of the reception zone, with distracting multipath artefacts, occasional dropouts in the stereo, and with noticeable hiss?  This latter lower grade of "FM quality" might indeed be accorded a quite poor MUSHRA rating by listeners on an assessment panel.

Link to comment
Share on other sites

17 hours ago, alanh said:

Irrelevant. But Alan, if you want to insist that's a valid criticism, I'm happy to insist the same applies to anything you might "evaluate". Deal?

 

17 hours ago, alanh said:

2. http://downloads.bbc.co.uk/rd/pubs/whp/whp-pdf-files/WHP006.pdf shows that DRC enables a wide dynamic range, which the FM multiplex does not. This was not demonstrated.

Alan, you might want to read that instead of just linking the first thing that comes up when you google "DAB DRC". There are better sources explaining DRC in MP3/AAC/etc - the one you chose simply reinforces the fact that few radios implement it properly!

Interestingly, as far as I can tell none of the DAB+ stations in Brisbane are exploiting DRC, Multiband DRC, or ETSI / Heavy DRC (at the very best they may have them set to fixed constants, so effectively no DRC - just fixed boost/cut). So it seems that's yet another reason the local implementation of DAB+ is failing to achieve its potential quality. Probably for the best, though, if receivers don't handle it properly...

 

17 hours ago, alanh said:

3. The DAB+ signals should not use any audio processing where as FM must have it. This was not tested. As a result DAB+ with DRC removing the compression would allow a wide dynamic range which is greater than for FM particularly near the edge of the coverage area. Multiband processing should not be used for DAB+ because the AAC encoder also contains a similar multiband filter set so if they are not exactly the same it will cause weird effects and wrecks the psychoacoustic system to minimise the effect of leaving lots of data outl

A valid criticism (and one which could've saved Alan from writing 1/2 a page of garbage if he'd only read the paper in the first place...). But, as I noted just above, it seems DAB/AAC DRC isn't being used properly anyway...

 

17 hours ago, alanh said:

4. They did not transmit the FM signal, which is important because Sweden like Norway is a mountainous country thus is more likely to downgrade the quality of the received signal.

5. Similarly they did not compare a DAB+ and and FM multiplexed signal in real world RF conditions which the listeners must endure.

Irrelevant. The point was to have a fixed reference  good-to-excellent FM reception, such as could be expected from a stationary receiver in a good-to-excellent reception area - not to test FM vs DAB+ under varying signal conditions.

 

17 hours ago, alanh said:

6. Considering ... Fraumhofter has the ability to design including application specific integrated circuits) and optimise performance of compression systems when compared to a single paper from a single university using equipment from 2007.

I'm somewhat inclined to agree - at least to a certain level - with Alan on this point.

However, if he's going to completely dismiss this study on that basis, in the absence of any other similar comparison you have to fall back on the fact that the MUSHRA testing process was specifically designed (by e.g. the use of hidden references & anchors, etc) to allow valid comparisons to be made between different tests, and rely on the information that Fraunhofer have published (i.e. papers & other studied comparisons - not press releases) where they compare uncompressed 44khz 16bit audio to HE-AACv2 & other codecs.

In which case you find that the results are very similar - HE-AACv2 @ 48kbps mono / 96kbps stereo falls short of "CD quality", but achieves remarkably similar MUSHRA scores to what Fraunhofer & others consider - and this paper has agreed - to be "FM Quality"

So it's not "a single paper from a single university" - it's the general consensus of Fraunhofer, Coding Technologies / Dolby Labs, the AES, etc, etc.

And Alan, please stop with your ridiclous attempts to rewrite history and put words into MLXXX's (and my, and others) mouth. Our criticisms of DAB+ have always been that the bitrates and other coding configurations typically used in Australia are insufficient to even come close to FM quality. No amount of ridicule, denial, and other spurious & specious claims to the contrary from you will ever change that - it just wastes our time, your time, and makes you look stupid every time one of us has to refute it.

 

Link to comment
Share on other sites



20 hours ago, Malich said:

So it's not "a single paper from a single university" - it's the general consensus of Fraunhofer, Coding Technologies / Dolby Labs, the AES, etc, etc.

However semi-formal written claims do exist that state without qualification that FM quality is achieved at much lower bitrates.

For example, the following slide claims that DAB+ 48 (presumably DAB+ radio received under good conditions and broadcast with a nominal bitrate of 48kbps in stereo) provides better audio quality than FM.

an-audio-quality-evaluation-of-digital-r

It can be found as slide 31 in http://www.slideshare.net/rojiththomas5/new-microsoft-office-power-point-presentation-4-15210438

Preceding slides provide detail about the digital testing methodologies but are silent on the FM testing methodology.  As far as I can tell, this informal presentation simply relies on it as a given that FM has a MUSHRA rating of "72% good". This "fact" is incorporated into slide 6 as follows:

The Audio Quality for Analog Broadcasting Systems are;

>FM     :72% -Good

>SW2   :63% -Fair

>SW1   :61% -Fair 

>MF      :40% -Fair

Here is slide 6:

an-audio-quality-evaluation-of-digital-r


From where the above MUSHRA ratings have been obtained is not stated. Nor are the FM reception conditions stated.  

In passing I'd note that it is also not explained how the MF band has acquired a rating of 40%.

Fortunately, the paper referred to in the current thread used FM sources for the participants in the exercise to evaluate for themselves using MUSHRA scoring. The paper did not adopt an "off the shelf" rating for "FM quality".

Edited by MLXXX
Link to comment
Share on other sites

MLXXX,

The above link is unbelievable. It also conflicts with your first post which is just as unbelievable. Just Googleing for sources for any straw is not good enough!

Testing of the listeners only went to 7 kHz so SBR is not tested as is any noise and distortion between 7 and 15 kHz which will really affect FM signals.

Short wave broadcasting with better sound than MF! Both use the same modulation but MF band is a minimum of 9/10 kHz channel width, typically double these channel, where as short wave is 5 kHz perhaps this is due to the audio being limited to 3.5 kHz or telephone quality. Also Short Wave has high distortion due to multiple reflected signals.

Reference:F.HOFMANN,C.HANSEN AND W.SCHAFER,”DIGITAL RADIO MONDIALE (DRM) DIGITAL SOUND BROADCASTING IN THE AM BANDS , ”IEEEE TRANS. BROADCAST.,VOL.49,NO.,PP.319-328, SEP . 2003 Why is there no testing of DRM particularly in the HF bands. Your link was published in 2012!

RADIO BROADCASTING SYSTEMS: DIGITAL AUDIO BROADCASTING TO MOBILE , PORTABLE AND FIXED RECEIVERS , ETSI EN 300 401 VER 1.4.1.JAN 2006 

It does not cover HE AAC sound compression DAB+ standard ETSI TS 102 563 V1.2.1 (2010-05) which is before the slideshow publication date.

I wonder about this college Believers Church Caarmel Engineering College in India. I tried a few links to the actual college and none work. There are only a few student Utube posts.

It might explain the lack of DRM testing from AIR's Bengaluru 200 kW transmitter.

Lastly you are confusing my comments on FM noise and as you have shown in the past you don't understand how it works in detail. I am not going to argue it all over again. You can search my arguements yourself.

Alanh

Link to comment
Share on other sites

15 hours ago, alanh said:

MLXXX,

The above link is unbelievable.

Well I imagine the people who prepared the powerpoint presentation would have thought they were being reasonable; and would have expected users to accept the presentation as possibly simplified but essentially correct.

As you know, in India it has been claimed that DRM is of FM quality.  DRM is being marketed as a quality broadcasting solution in that country. For that to make sense it follows that analogue forms of radio broadcasting must be presented as being of poor quality.  (I have previously drawn attention on this forum to the -- to my mind -- outrageously generous claims that have been made in India about DRM broadcasting sound quality.)

It is not surprising to me that writers in India would readily latch on to low MUSHRA ratings for FM radio and particularly low ratings for MW AM broadcasting, in circumstances where their country was seriously considering or actually committed to a move to digital broadcasting using the MW band. However I would still like to know where the cited MUSHRA ratings for analogue broadcasting came from.

 

15 hours ago, alanh said:

Short wave broadcasting with better sound than MF! Both use the same modulation but MF band is a minimum of 9/10 kHz channel width, typically double these channel, where as short wave is 5 kHz perhaps this is due to the audio being limited to 3.5 kHz or telephone quality. Also Short Wave has high distortion due to multiple reflected signals.

I imagine the low MUSHRA score for MF could have been justified on the basis of extreme congestion in the MF broadcast band.

As regards your comments about audio bandwidth, it is usual in discussion of received AM sound quality to ignore (the now rare) wideband tuners. So it makes no difference whether the station occupies a nominal 18/20kHz bandwidth or a nominal 10kHz bandwidth. Both are able to provide audio up to 5kHz, but in practice the AM receiver will be designed not even to use that much of the audio bandwidth.

 

15 hours ago, alanh said:

I wonder about this college Believers Church Caarmel Engineering College in India. I tried a few links to the actual college and none work. There are only a few student Utube posts.

The Carmel College of Engineering & Technology seems to be a small operation (and occasional spelling mistakes in some of the webpages do not instill confidence). It is claimed the college started up in 2014/15, so on that basis has not been going for long. Their website overview (http://carmelcet.in/overview/) states in part:

The Carmel College of Engineering & Technology (CCET) is owned and managed by the St.Joseph’s Carmel Educational & Charitable Trust of CMI, registered on 13-09-2012, a Public Charitable Trust established this Institution in 2014-15 at Punnapra, Alappuzha, with the object of providing education to all in the field of Engineering & Technology. Our motto is moulding Engineers par Excellence with Integrity, commitment and human values, who are competent to meet the global standards in business, industry and research and would act as catalysts for the transformation of the society.

 

The point I was trying to make alanh is that there do exist claims in writing to the effect that very low bitrate DAB+ sounds as good as FM. However such claims do not appear to have a solid foundation.

 

15 hours ago, alanh said:

It also conflicts with your first post which is just as unbelievable

The paper I refer to in the first post, a copy of which I understand Malich has forwarded to you, is a properly structured academic paper. I find nothing about its conclusions difficult to believe.  

Edited by MLXXX
Slight elaborations
Link to comment
Share on other sites

16 hours ago, alanh said:

Reference:F.HOFMANN,C.HANSEN AND W.SCHAFER,”DIGITAL RADIO MONDIALE (DRM) DIGITAL SOUND BROADCASTING IN THE AM BANDS , ”IEEEE TRANS. BROADCAST.,VOL.49,NO.,PP.319-328, SEP . 2003 Why is there no testing of DRM particularly in the HF bands. Your link was published in 2012!

I can't help but see a certain irony in citing a paper published in 2003, then immediately complaining about one published in 2012... :blink:

(And why even bring DRM into the discussion? It's like bringing a knife to a gunfight - not even xHE-AAC is going to come anywhere near to FM quality, even at a practical absolute max bitrate of ~30kbps...)

 

15 hours ago, MLXXX said:

As you know, in India it has been claimed that DRM is of FM quality.

And, having earlier this year spent some time around Delhi in the company of a ham/radio enthusiast friend listening to the MW broadcasts using my homebrew SDR and his AV-DR-1401, no-one there believes it either...

Link to comment
Share on other sites



  • 1 month later...

Occasionally I get notified of a new topic (http://www.dtvforum.info/index.php?/forum/76-television-and-radio-off-topic/), such as today. I was going to post this there, but seeing as my response is more to do with bit-rate, it's probably better I post here.

I'm sure many others, just like myself, spend little to no time reading stuff here any more thanks to trolls such as.. you know who. Although he doesn't really abuse or threaten anyone, constantly going off-topic, especially when you have spent many hours providing links of research to backup your claims, is not what forums are or should be about.

Normally this behaviour is worthy of warnings, and then bans. Obviously that's never going to happen here.

Checking random days from the archive of this website, it seems I'm not alone, and many have lost interest. The numbers of users online reflect that...
2007: 419, 2008: 407, 2010: 544, 2011: 421, 2012: 500, 2013: 271, 2014: 273, 2016: 197, Today: 88 users online

Anyway...

I've always "appreciated" HE-AAC as a solution to low bit-rate situations, but never as an "alternative" to anything but that. If/when we're one day sending data between Earth and people on Mars, then sure, codecs such as HE-AAC and Opus will prove to be very valuable. But sending bit-rates as low as 24kbps across town for music and claiming it as an alternative to FM radio? I think CRA is trolling just as much as a certain someone here.

And in fact if you now check out their website, they're advertising/pushing their mobile phone app just as much as DAB+. Interesting how not so long ago they were suggesting the internet as an unviable option for radio stations. "Why stream 48kbps to each listener when you can use DAB+ to stream just 48kbps to every listener at the same time?" was what they were basically pushing, if anyone remembers that.

Personally I find a good 160-192kbps mp3 or AAC stream to be preferable to FM radio, perhaps even a 128kbps mp3 with the right LAME encoder is very competitive. But even a 128kbps HE-AAC is something I try to avoid if possible, due to the nature of how SBR works.

It's a big read and I'm not sure if the age group or average age was mentioned for the test, but that would definitively contribute to the test/s results.

But I guess if my ears were no longer able to hear much above 10-11kHz, I'd probably think HE-AAC sounded good too :D

 

Link to comment
Share on other sites

1 hour ago, lpy said:

"appreciated" HE-AAC as a solution to low bit-rate situations, but never as an "alternative" to anything but that. If/when we're one day sending data between Earth and people on Mars, then sure, codecs such as HE-AAC and Opus will prove to be very valuable. But sending bit-rates as low as 24kbps across town for music and claiming it as an alternative to FM radio? I think CRA is trolling just as much as a certain someone here.

Yes 24kbps might make sense for interplanetary communication, and it does makes sense for international shortwave using DRM.  I've recently been listening to ABC News Radio on 48kbps (nominal) comparing the sound via an RTL-SDR USB stick and my old Roberts DAB+ radio. The artefacts are just as bad with both sources. I really find it a trial listening to DAB+ bitrates less than around 64kbps (nominal).    

 

1 hour ago, lpy said:

 But even a 128kbps HE-AAC is something I try to avoid if possible, due to the nature of how SBR works.

It isn't usual to encode with SBR activated if the available bitrate is as high as 128kbps.  You'd have to go out of your way to force an HE-AAC encoder to use SBR at 128kbps - it would default to not using SBR.

 

1 hour ago, lpy said:

But I guess if my ears were no longer able to hear much above 10-11kHz, I'd probably think HE-AAC sounded good too :D

Oh the artefacts are audible at much lower frequencies than 10kHz. Low bitrate AAC (whether or not bolstered by SBR) has numerous defects, which have variously been described as giving the following subjective impressions: a hollowness, a lack of bite, a phasey quality, a metallic quality. These defects extend across the low and mid frequencies for human hearing and aren't restricted to the top octave (10kHz to 20kHz).

I must say that at times I do find SBR very intrusive with its sprays of bland high frequency sibilants when rendering the spoken word. Speech can sound to me like it's been generated by a speech synthesiser rather than spoken by a human being!

Link to comment
Share on other sites

17 hours ago, lpy said:

I've always "appreciated" HE-AAC as a solution to low bit-rate situations, but never as an "alternative" to anything but that.

Yeah, that's the thing. Quite frankly, it's amazing what HE-AAC can do. Providing 'acceptable' (i.e. approximately FM-quality) at bitrates around 7% that of uncompressed audio is an impressive feat.

Unfortunately, in my opinion the average listener is being mislead by 2 things:

  • Firstly, the perception - supportable in some aspects, but over-sold by marketing to the point that it's popularly believed to be the case in all aspects - that digital is always better.
  • And secondly, the fact that - as well-documented in other posts here and elsewhere - broadcasters choose & prefer to use inappropriately low bitrates.

The latter seems to be a particularly hard thing for some to accept, mostly because of the former...

 

19 hours ago, lpy said:

Personally I find a good 160-192kbps mp3 or AAC stream to be preferable to FM radio, perhaps even a 128kbps mp3 with the right LAME encoder is very competitive. But even a 128kbps HE-AAC is something I try to avoid if possible, due to the nature of how SBR works.

I'd generally agree with that, with one small proviso: SBR shouldn't even begin to be operational in HE-AAC at 128kbps.

From memory the spec only characterises it up to 96kbps, and even there it's considered a 50:50 bet against HE-AAC without SBR (i.e. AAC-LC). Above that, it more commonly sounds subjectively worse!

Really, SBR should only be used up to ~80kbps and, as far as I'm aware, the standard profiles enforce that. Chances are, if you have 128kbps files that are reported as "HE-AAC" or "AAC-LC, SBR", they're in fact simply AAC-LC...

(I won't go in to the details of AAC/MP4 metadata, but if you're curious there's a short relevant run-down in the answer here. Note carefully the distinction between "MPEG-2 AAC LC" and "MPEG-4 Audio" object types, and the specific audio object types available under "MPEG-4 Audio".)

Link to comment
Share on other sites

3 hours ago, Malich said:

Really, SBR should only be used up to ~80kbps and, as far as I'm aware, the standard profiles enforce that. Chances are, if you have 128kbps files that are reported as "HE-AAC" or "AAC-LC, SBR", they're in fact simply AAC-LC...

Indeed yes.  For example in Brisbane at the moment we have two 104kbps (nominal) DAB+ services* and they are both AAC-LC.  The actual audio bitrate is approximately 85kbps, according to the DAB software** I'm using.

___________________

* Magic 882, also listed as *Magic 882*; and 1116 4BC, also listed as 4BC1116 NewsTalk.

** DAB Player by Andreas Gsinn. 

 

 

Link to comment
Share on other sites

MLXXX,

I though you had decided to give this topic a break. Obviously you cannot help yourself.

Pity when you were looking for a new car 3 years ago, you didn't listen to DAB+ radios in them at the car dealers.

All your opinions and that's what they are are well known so give it a rest as promised.

 

Alanh

Link to comment
Share on other sites



11 minutes ago, alanh said:

MLXXX,

I though you had decided to give this topic a break. Obviously you cannot help yourself.

...

All your opinions and that's what they are are well known so give it a rest as promised.

 

Alanh, you seem to have got the wrong end of the stick again. I did not say I had decided to give any topic a break. I said I had decided to cut back my responses to a certain forum member. For your information, here is an extract from my relevant post:

I'm seeing a repeated pattern of a certain forum member arguing the point endlessly, such as in the thread below: and as of now I've decided to cut back my responses. (I see no point in continually going around in circles or off on tangents, with little prospect of any resolution.)

Link to comment
Share on other sites

  • 2 months later...

Yes I have read the paper ;-)

well I have been following this situation for some time.  Since 'digital' first emerged on my scene back in 2006 when I acquired a digital TV card and found an ABC side audio radio transmission of undeniable CD quality, I was blown away.  On large speakers I had that going and was so happy.  Then came the ever increasing numbers of 'channels' and ever decreasing quality.  Similarly with digital radio.  When it first arrived here in Canberra, 96Kbit/s SBS Chill was tolerable.  Then it slowly fell back to 64Kbit/s which we have been assured is a step forward from FM.

I think it sounds hollow.  A subjective evaluation suggests 64Kbit/s is tolerable as background or low volume music, but not on large speakers or in the car.  Less than that starts to really become annoying after a while.

However, we keep being assured, 64Kbit/s is the future for radio and will usher in a new era of high quality audio and choice.

I'm not convince.  I think this is a step backward just like telcos used 3G as an excuse to maintain if not even further lower voice channel bandwidth to the point where mobile phone calls became intolerable for more than a few minutes at a time.  Perhaps intentionally as capped plans became the norm, just like cafe's which are known to intentionally design acoustically reflective surfaces to make the place noise and uncomfortable for extended visits (buy, eat, move one).

I believe DAB has taken on the same character as the above.  It's being rolled out for the benefit of the broadcasters cost savings and revenue generation - not for the comfort of listeners.

So I see a dim future for Australian radio as mobile bandwidth steadily improves.

Here in Canberra on this never-ending trial of five years or so, 64Kbit/s is the highest.  And it sucks.

That's my subjective non-scientific anecdotal assessment.

What have others found?

Edited by Audiofile
Link to comment
Share on other sites

5 hours ago, Audiofile1503564314 said:

Similarly with digital radio.  When it first arrived here in Canberra, 96Kbit/s SBS Chill was tolerable.  Then it slowly fell back to 64Kbit/s which we have been assured is a step forward from FM.

I think it sounds hollow.  A subjective evaluation suggests 64Kbit/s is tolerable as background or low volume music, but not on large speakers or in the car.  Less than that starts to really become annoying after a while.

However, we keep being assured, 64Kbit/s is the future for radio and will usher in a new era of high quality audio and choice.

Well there certainly is a lot of choice of "channel" to listen to.

As for 64kbps (nominal) DAB+ radio quality these features would be pluses:

  • much extended treble compared with standard AM radios
  • in practice an extended bass compared with typical AM broadcasts
  • less THD (Total Harmonic Distortion) than with AM reception
  • generally a much quieter noise floor compared with an AM radio (unless the AM radio is close to the transmitter)
  • much less susceptible to impulse interference than AM radio
  • usually in stereo whereas AM radio in Australia is almost always limited to mono
  • less susceptibility to multipath reception glitches than FM radio
  • noticeably quieter noise floor than FM radio reception towards the outer limits of the FM reception zone.

 And these factors would be minuses:

  • some people find the higher treble components created by SBR unnatural sounding and/or irritating
  • some people hear a hollowness or thinness in the sound 
  • some people hear a jangly or phasey quality
  • some people find the sound lacks "bite" and "immediacy".

 My personal opinion is that 64kbps (nominal) stereo DAB+ is very competitive with high signal strength AM reception using typical treble-impaired AM radios. It will for certain critical listeners noticeably lack verve and vigour and naturalness, but even for such critical listeners on an overall basis it may be perceived subjectively as being of comparable, or even superior, subjective quality.

48kbps (nominal) stereo DAB+ is in my opinion significantly more problematic. I personally find it a bit of a trial to listen to and given a choice would very possibly opt instead for treble-impaired monophonic AM with its noticeable THD.

32kbps (nominal) stereo DAB+ tends to be so poor that even non-critical listeners may find the artefacts of the psychoacoustic codec annoying. At the same time, many other people will report being happy with it!  Tolerance of low bitrate AAC+ is a highly personal and subjective matter.



Turning to consider FM, I am still in the dark as to what standard of FM reception is considered appropriate to use for subjective comparisons with DAB+. As I said in my opening post:

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.

  

Link to comment
Share on other sites

Yep, so an undisputed improvement from AM.  Great.  However FM at the low-bitrates used is open to debate.

I'd say no.  It sounds hollow and tinny - especially with acoustics and high vocals.  It's really noticeable.

It's not something most people would turn up loud on a set of speakers which FM was terrific for.

I would argue that in many respects, Australian digital radio is a step backward from FM and it is false advertising to claim otherwise.

How do I know?  Just by listening to it.

Link to comment
Share on other sites

Audiofile,

This debate has been on this site for years. A lot of the poor quality audio criticism came from the UK initially because they use the original version of DAB which uses a much less efficient compression system and they put on so many program streams that even now they have virtually no stereo transmissions on digital.

In Australia the highest data rates are from the music channels of the ABC including Double J which has the highest listenership of any pure digital station.

You don't say what you are listening to digital on. There is no restriction in bass from the transmission side so suggest you go to a new car dealer and listen to DAB+ in one of the many new models available. The other option is a DAB+ tuner in a hifi system with decent speakers. Computer decoders are not as good as the dedicated receiver chips in new radios. I have a Bush headphone radio which sounds great and has plenty of bass on good headphones.

FM is not perfect either. It reduces the sharpness of any loud sounds prior to transmissions, but most listeners are unaware of this because they have never heard the original and are used to this sound.

You need to compare the original sound prior to transmission with the received sound and not know whether the source is the original, FM, DAB+, AM.

Alanh

Link to comment
Share on other sites



  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...
To Top