Audio compression does not even give FM quality

Jim Lesurf wrote:
In article .com,
spiney
wrote:
DAB/ MPEG2 level2 audio coding (not actually "Musicam", although pretty
close!).

I can "throw a little light" on some of these things .....


Well, that's what I expect from Jim, sarcasm but no facts.

Jim, when you've replied about getting digital tv wrong, I might then
respond in turn about Hi Fi News.

On 29 Oct 2005 08:17:59 -0700, "spiney"
wrote:

Jim, when you've replied about getting digital tv wrong, I might then
respond in turn about Hi Fi News.

Well, don't wait for Jim, reply to me instead.

I repeat my earlier post, please cite some references.

--
Alan White
Twenty-eight miles NW of Glasgow, overlooking Loch Goil and Loch Long in Argyll, Scotland.
Web cam and weather:- http://www.windycroft.gt-britain.co....her/kabcam.htm
Some walks and treks:- http://www.windycroft.gt-britain.co.uk/walks/

On 29 Oct 2005 04:52:43 -0700, spiney wrote:

Oh dear, the usual rubbish replies

The only rubbish is that emanating from your direction, Spineless.

(why do I even bother saying anything?).

Then don't. We would all love it if you just crawled away and died
somewhere.

You have proved beyond all doubt that you are a completely ignorant
clueless moron, as people who post using made up names from throw away
junk accounts such as Hotmail mostly seem to be.

In article .com,
spiney wrote:
Dolby surround info is of course already in the existing film analogue
stereo soundtrack (existing A noise reduction is decoded prior to
re-transmission, but B remains locked into the surround info, and forms
part of any domestic surround decoding!).

Good grief.

--
*The most wasted day of all is one in which we have not laughed.*

Dave Plowman London SW
To e-mail, change noise into sound.

In article om,
spiney
wrote:

Jim Lesurf wrote:
In article .com,
spiney wrote:


Jim, when you've replied about getting digital tv wrong, I might then
respond in turn about Hi Fi News.

Erm. First you'd have to quote/reference so we can see what on earth you
are talking about. What *specific* "getting digital tv wrong" are talking
about? Give the dates/times of the postings you wish to claim as the
basis of your assertions, and quote the relevant statements. Without
such details you are making vacuous assertions that you are failing
to substantiate.

So far, it looks like you are simply making assertions based upon your
dreams... Afraid that no-one can help you with that.

You decide for yourself if you want to give references to the Hi Fi News
articles by Angus. I can't see that anyone takes your assertions seriously,
so it is no loss to anyone else if you fail.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

In article , Dave Plowman (News)
wrote:
In article .com,
spiney wrote:
Dolby surround info is of course already in the existing film analogue
stereo soundtrack (existing A noise reduction is decoded prior to
re-transmission, but B remains locked into the surround info, and
forms part of any domestic surround decoding!).

Good grief.

Classic 'spiney' I'm afraid.

Must admit that in all my years of uni teaching and using usenet I've never
met anyone before who was as impervious to learning or understanding as
spiney. Combined with his ability to go on spouting the same mistakes,
regardless of anyone trying to explain things to him.

Slainte,

Jim

--
Electronics http://www.st-and.ac.uk/~www_pa/Scot...o/electron.htm
Audio Misc http://www.st-and.demon.co.uk/AudioMisc/index.html
Armstrong Audio http://www.st-and.demon.co.uk/Audio/armstrong.html
Barbirolli Soc. http://www.st-and.demon.co.uk/JBSoc/JBSoc.html

spiney wrote:




Spiney, is your email address supposed to read:

sparki stuf

or

spark is tuf

Just interested, like.


--
Steve - www.digitalradiotech.co.uk - Digital Radio News & Info

Please sign the petition asking the BBC to provide better audio quality
on its radio stations on DAB, Freeview, satellite and cable:
http://tinyurl.com/a68e4

DAB sounds worse than FM wrote:
spiney wrote:




Spiney, is your email address supposed to read:

sparki stuf

or

spark is tuf

Just interested, like.


--
My addr? A curious question, DAB, since there are no spaces in email
addresses!

If you're on about the possible meaning - which seems implied - then it
doesn't have one, it's just an alliterative sound!

I chose this "nonsense" address ad hoc, as it seemed unlikely to be
already in use, an annoying situation if you're trying to set up an
email account.

Otherwise, who cares? If anyone emails me, it's probably going to be
pointless abuse!

Regarding DAB audio rates, yes I agree entirely, but i don't think a
petition will change anything (but good luck anyway!).

Agamemnon wrote:
wrote in message
oups.com...
Agamemnon wrote:
[snip]

Ogg isn't designed to target bitrates, but quality levels. -q5 and
above are worth using. However, you appear to be listing the CBR
bitrates you used. This suggests you don't know how to use the encoder
properly.

Wrong. I was using ABR.

Ah, so you neither got the exact bitrates you quoted, nor any
particular quality level.

Not that there's anything wrong with using ABR for encoding, but the
results (quality wise, and bitrate wise) are a half-way house between
true CBR and free-ranging (quality based) VBR.


The implications of this experiment to DVB are clearly obvious. OGG
Vorbis
is equivalent in quality to AAC and maybe even AAC+.

Arguable. I can think of one test result that could "prove" your
point, but the AAC encoder in that test is a-typical.

OGG at 64kbps is
equivalent to mp2 at 256 kbps ie. 4 times the bit rate.

Arguable. Actually, I think you'd struggle to back this one up with any
published test.

Based on my comparison between this and the MUSICM codec for Cool Edit Pro I
have all the evidence I need.

You mean the buggy MPEG codec? The one that's even got a bug in the
pscyhoacoustic model? sarcasmwell that's a good basis for a
test!/sarcasm

btw, read here before you use the name MUSICAM again...
http://www.tnt.uni-hannover.de/proje.../mpeg1.html#1F


Thus for DVB to get
anywhere near to FM it needs to be encoded at the equivelent of 4 times
256
kbps ie. 1024 kbps

This is just silly reasoning. It doesn't work like that.

which is the same bit rate as sampling the 16 KHz FM
audio signal in PCM at 32 KHz in the first place.

Even freeformat mp3 stops at 640kbps. Try it - for some things its
"near lossless".

Why should I bother when FLAC is completely lossless.

At what fixed bitrate is FLAC universally lossless?

(Hint: for CD quality audio, the answer is approxiately 1.4Mbps, give
or take).


The fact the ogg can't even encode hiss properly at 256 kbps the
equivalent
of mp2 at 1024 kbps means that mp2 can not even approach FM quality PCM
even
if is sampled at the same bit rate so why was it ever devised when FLAC
can
give you completely lossless compression at 512 kbps !

That argument is based on your faulty reasoning. Plus FLAC often needs
over 800kbps (though probably not for Choral Evensong!)

In fact even the biased creators of mp2 and the EBU claimed that mp2 was
transparent at 256kbps per channel

No, 256kbps stereo - i.e. both channels total bitrate.

That's not what I read.

Where did you read it? The clearest document (from FhG) is no longer on
line.


Further more why was NICAM ever devised when FLAC uses less bandwidth
than
NAICM which only uses 12 bits.

Well, obviously because the inventors didn't have you knowledge and
genius! ;-)

(btw, the real answer is in the name: Near Instantaneous... FLAC isn't,
neither does it deliver a guaranteed bitrate)


On top of this if FLAC were used on DVB instead of mp2 using 64 QAM
modulation correct me if I'm wrong

I keep doing, but I doubt it'll help.

but wouldn't it use the same bandwidth as
mp2 at 128 kbps on DAB and be completely lossless and better than FM so
why
was DAB even chosen over FLAC in the first place ?

There's one major problem with lossless compression in broadcast: you
can't guarantee the bitrate. If you want near lossless, use NICAM, or
something like Wavepack hybrid mode. It's useful in some situations,
but a good lossy codec is a more elegant situation for one-time
encoding and a limited bitrate situation.

I'll turn the question around: what's wrong with AAC at 192kbps?

Same problem as with OGG at 192kbps I expect. Can't deal with moderate hiss,
electrical and digital interference noises (more or less square waves)
picked up by my FM tuner and speech echo.

Sounds like you're using an old Ogg encoder. It's never been great with
hiss at lower quality levels, but has improved.

Interesting that it was the faults in your FM reception that didn't
encode properly. How about the music itself?

Is AAC any good at encoding square waves. Didn't think so. It wouldn't be
very good at recording bird song and playing it back to the birds either.

So you're rejecting AAC after testing a _different_ codec; one that's
specifically designed to avoid every single one of the hundreds of
patents on which AAC is based?!

Your logic eludes me.

Cheers,
David.

wrote in message
ups.com...
Agamemnon wrote:
wrote in message
oups.com...
Agamemnon wrote:
[snip]

Ogg isn't designed to target bitrates, but quality levels. -q5 and
above are worth using. However, you appear to be listing the CBR
bitrates you used. This suggests you don't know how to use the encoder
properly.

Wrong. I was using ABR.

Ah, so you neither got the exact bitrates you quoted, nor any
particular quality level.

ABR is what is recommended by the encoder and what the Q ratings are based
on. Q6 = 192, Q8 = 256.


Not that there's anything wrong with using ABR for encoding, but the
results (quality wise, and bitrate wise) are a half-way house between
true CBR and free-ranging (quality based) VBR.

I tried VBR and for a range from 128 kbps to 500 kbps it was averaging out
at 410 kbps and even then I could still tell the compressed recording from
the live and uncompressed source.

Music might sound alright using OGG but I can now recognise distortion on
recorded speech and background noise instantly even at VBR ranging from 192
to 500 which averages at 466.



The implications of this experiment to DVB are clearly obvious. OGG
Vorbis
is equivalent in quality to AAC and maybe even AAC+.

Arguable. I can think of one test result that could "prove" your
point, but the AAC encoder in that test is a-typical.

OGG at 64kbps is
equivalent to mp2 at 256 kbps ie. 4 times the bit rate.

Arguable. Actually, I think you'd struggle to back this one up with any
published test.

Based on my comparison between this and the MUSICM codec for Cool Edit
Pro I
have all the evidence I need.

You mean the buggy MPEG codec? The one that's even got a bug in the
pscyhoacoustic model? sarcasmwell that's a good basis for a
test!/sarcasm

btw, read here before you use the name MUSICAM again...
http://www.tnt.uni-hannover.de/proje.../mpeg1.html#1F


Thus for DVB to get
anywhere near to FM it needs to be encoded at the equivelent of 4
times
256
kbps ie. 1024 kbps

This is just silly reasoning. It doesn't work like that.

which is the same bit rate as sampling the 16 KHz FM
audio signal in PCM at 32 KHz in the first place.

Even freeformat mp3 stops at 640kbps. Try it - for some things its
"near lossless".


At that rate I might as well use FLAC which averages at 666 kbps since I've
got plenty of free disc space and DVDs don't cost that much.

Further more OGG VBR uses 18% of my system resources (3.2GHz) while FLAC
only uses 5% or less.

Why should I bother when FLAC is completely lossless.

At what fixed bitrate is FLAC universally lossless?

(Hint: for CD quality audio, the answer is approxiately 1.4Mbps, give
or take).

Do you know of any naturally produced sound in existence that forced FLAC to
encode as 1.4 Mbps ?



The fact the ogg can't even encode hiss properly at 256 kbps the
equivalent
of mp2 at 1024 kbps means that mp2 can not even approach FM quality
PCM
even
if is sampled at the same bit rate so why was it ever devised when
FLAC
can
give you completely lossless compression at 512 kbps !

That argument is based on your faulty reasoning. Plus FLAC often needs
over 800kbps (though probably not for Choral Evensong!)

In fact even the biased creators of mp2 and the EBU claimed that mp2
was
transparent at 256kbps per channel

No, 256kbps stereo - i.e. both channels total bitrate.

That's not what I read.

Where did you read it? The clearest document (from FhG) is no longer on
line.

Some EBU paper on testing mp2 which I read about 7 years ago.



Further more why was NICAM ever devised when FLAC uses less bandwidth
than
NAICM which only uses 12 bits.

Well, obviously because the inventors didn't have you knowledge and
genius! ;-)

(btw, the real answer is in the name: Near Instantaneous... FLAC isn't,
neither does it deliver a guaranteed bitrate)


On top of this if FLAC were used on DVB instead of mp2 using 64 QAM
modulation correct me if I'm wrong

I keep doing, but I doubt it'll help.

but wouldn't it use the same bandwidth as
mp2 at 128 kbps on DAB and be completely lossless and better than FM
so
why
was DAB even chosen over FLAC in the first place ?

There's one major problem with lossless compression in broadcast: you
can't guarantee the bitrate. If you want near lossless, use NICAM, or
something like Wavepack hybrid mode. It's useful in some situations,
but a good lossy codec is a more elegant situation for one-time
encoding and a limited bitrate situation.

I'll turn the question around: what's wrong with AAC at 192kbps?

Same problem as with OGG at 192kbps I expect. Can't deal with moderate
hiss,
electrical and digital interference noises (more or less square waves)
picked up by my FM tuner and speech echo.

Sounds like you're using an old Ogg encoder. It's never been great with
hiss at lower quality levels, but has improved.

Xiph.Org libVorbis I 20020717 (for Total Recorder)


Interesting that it was the faults in your FM reception that didn't
encode properly. How about the music itself?

It wasn't instrumental music. It was unaccompanied Choral singing and the
Radio 3 presenters, ie. human speech.


Is AAC any good at encoding square waves. Didn't think so. It wouldn't be
very good at recording bird song and playing it back to the birds either.

So you're rejecting AAC after testing a _different_ codec; one that's
specifically designed to avoid every single one of the hundreds of
patents on which AAC is based?!

AAC at 256 kbps using the FAAC (v.2.4) encoder for Cool Edit Pro is worse
than OGG at 256 kbps. In fact its even worse than OGG at 192. A muddy warbly
sound on the speech of the Radio 3 presenter is instantly noticeable on the
256 kbps ACC file but its not there on the 256 OGG file (using the Cool Edit
Pro encoder) and is much less noticeable on the OGG 192 kbps than the ACC
256 file.

That said I can still tell the 256 kbps OGG is compressed because the
ambience is all wrong compared to the original FLAC recoding I made.



Your logic eludes me.

Do the tests yourself. I can provide you with the one minute long 5mb FLAC
file I recorded. In fact all you need to do is to listen to the first 10
seconds of it and the artefacts are easy to pick up.


Cheers,
David.