Fraunhofer MP3 Codec

[mcp6453[_2_]]

Unfortunately encoding to MP3 is a fact of life. The best codec I've found is the Fraunhofer, which is the one used by
Adobe Audition and even iTunes. What other software uses the Fraunhofer (as opposed to the lame Lame)?

[JackA]

On Friday, September 25, 2015 at 5:46:59 PM UTC-4, mcp6453 wrote:
> Unfortunately encoding to MP3 is a fact of life. The best codec I've found is the Fraunhofer, which is the one used by
> Adobe Audition and even iTunes. What other software uses the Fraunhofer (as opposed to the lame Lame)?

FYI: "LAME has a much better proven track record, since a LOT of public testing has been done on it".


Jack

[Scott Dorsey]

In article >,
mcp6453 > wrote:
>Unfortunately encoding to MP3 is a fact of life. The best codec I've found is the Fraunhofer, which is the one used by
>Adobe Audition and even iTunes. What other software uses the Fraunhofer (as opposed to the lame Lame)?

The problem is that, unlike Lame, you can't get the Fraunhofer codec just
as a standalone application that you can call from the command line. About
as close as you can get is with one of the plugins that give you Fraunhofer
encoding from your existing DAW software, like the Sonnox plugin.

That said... I don't think that Lame is always a bad thing. Part of the
problem with Lame is that it has a million settings that you need to fiddle
around with, whereas Fraunhofer is pretty much set-and-forget, though.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[JackA]

On Friday, September 25, 2015 at 7:38:34 PM UTC-4, Scott Dorsey wrote:
> In article >,
> mcp6453 > wrote:
> >Unfortunately encoding to MP3 is a fact of life. The best codec I've found is the Fraunhofer, which is the one used by
> >Adobe Audition and even iTunes. What other software uses the Fraunhofer (as opposed to the lame Lame)?
>
> The problem is that, unlike Lame, you can't get the Fraunhofer codec just
> as a standalone application that you can call from the command line. About
> as close as you can get is with one of the plugins that give you Fraunhofer
> encoding from your existing DAW software, like the Sonnox plugin.
>
> That said... I don't think that Lame is always a bad thing. Part of the
> problem with Lame is that it has a million settings that you need to fiddle
> around with, whereas Fraunhofer is pretty much set-and-forget, though.

Side note:

I'm guessing, from internet searches, Neil Young's Pono player is pretty much a dead issue. If these (video) people (artists) did lie to other people, JUST, maybe, because they invested in Pono, it may have gained some acceptance...

https://www.youtube.com/watch?v=xH8I0LUjrqw

Good-bye Pono. Nice try, Neil.

Jack

> --scott
>
> --
> "C'est un Nagra. C'est suisse, et tres, tres precis."

[Les Cargill[_4_]]

Scott Dorsey wrote:
> In article >,
> mcp6453 > wrote:
>> Unfortunately encoding to MP3 is a fact of life. The best codec I've found is the Fraunhofer, which is the one used by
>> Adobe Audition and even iTunes. What other software uses the Fraunhofer (as opposed to the lame Lame)?
>
> The problem is that, unlike Lame, you can't get the Fraunhofer codec just
> as a standalone application that you can call from the command line. About
> as close as you can get is with one of the plugins that give you Fraunhofer
> encoding from your existing DAW software, like the Sonnox plugin.
>
> That said... I don't think that Lame is always a bad thing. Part of the
> problem with Lame is that it has a million settings that you need to fiddle
> around with, whereas Fraunhofer is pretty much set-and-forget, though.
> --scott
>


I use either the direct CBR settings ( -b xxx ) or -V 1
( next to highest variable bitrate setting ) with Lame
and have always been happy with the results. I nearly
always use -V 1 since it's quite close to -b 192 in
size, usually.

It's decidedly unfiddly.

--
Les Cargill

[JackA]

On Friday, September 25, 2015 at 9:05:48 PM UTC-4, Les Cargill wrote:
> Scott Dorsey wrote:
> > In article >,
> > mcp6453 > wrote:
> >> Unfortunately encoding to MP3 is a fact of life. The best codec I've found is the Fraunhofer, which is the one used by
> >> Adobe Audition and even iTunes. What other software uses the Fraunhofer (as opposed to the lame Lame)?
> >
> > The problem is that, unlike Lame, you can't get the Fraunhofer codec just
> > as a standalone application that you can call from the command line. About
> > as close as you can get is with one of the plugins that give you Fraunhofer
> > encoding from your existing DAW software, like the Sonnox plugin.
> >
> > That said... I don't think that Lame is always a bad thing. Part of the
> > problem with Lame is that it has a million settings that you need to fiddle
> > around with, whereas Fraunhofer is pretty much set-and-forget, though.
> > --scott
> >
>
>
> I use either the direct CBR settings ( -b xxx ) or -V 1
> ( next to highest variable bitrate setting ) with Lame
> and have always been happy with the results. I nearly
> always use -V 1 since it's quite close to -b 192 in
> size, usually.
>
> It's decidedly unfiddly.

What, no -q0 ?????

Jack

>
> --
> Les Cargill

[mcp6453[_2_]]

On 9/25/2015 7:38 PM, Scott Dorsey wrote:
> In article >,
> mcp6453 > wrote:
>> Unfortunately encoding to MP3 is a fact of life. The best codec I've found is the Fraunhofer, which is the one used by
>> Adobe Audition and even iTunes. What other software uses the Fraunhofer (as opposed to the lame Lame)?
>
> The problem is that, unlike Lame, you can't get the Fraunhofer codec just
> as a standalone application that you can call from the command line. About
> as close as you can get is with one of the plugins that give you Fraunhofer
> encoding from your existing DAW software, like the Sonnox plugin.
>
> That said... I don't think that Lame is always a bad thing. Part of the
> problem with Lame is that it has a million settings that you need to fiddle
> around with, whereas Fraunhofer is pretty much set-and-forget, though.
> --scott
>

This test is interesting. You can definitely hear the grunge in the LAME versions. (Two are LAME, one if Fraunhofer.)

http://www.richardfarrar.com/which-is-the-best-mp3-encoder-for-podcasts/

[JackA]

On Friday, September 25, 2015 at 10:13:46 PM UTC-4, mcp6453 wrote:
> On 9/25/2015 7:38 PM, Scott Dorsey wrote:
> > In article >,
> > mcp6453 > wrote:
> >> Unfortunately encoding to MP3 is a fact of life. The best codec I've found is the Fraunhofer, which is the one used by
> >> Adobe Audition and even iTunes. What other software uses the Fraunhofer (as opposed to the lame Lame)?
> >
> > The problem is that, unlike Lame, you can't get the Fraunhofer codec just
> > as a standalone application that you can call from the command line. About
> > as close as you can get is with one of the plugins that give you Fraunhofer
> > encoding from your existing DAW software, like the Sonnox plugin.
> >
> > That said... I don't think that Lame is always a bad thing. Part of the
> > problem with Lame is that it has a million settings that you need to fiddle
> > around with, whereas Fraunhofer is pretty much set-and-forget, though.
> > --scott
> >
>
> This test is interesting. You can definitely hear the grunge in the LAME versions. (Two are LAME, one if Fraunhofer.)
>
> http://www.richardfarrar.com/which-is-the-best-mp3-encoder-for-podcasts/

I'm curious, why are you even concerned about sound quality? The majority of the world really doesn't care. If you're thinking of taking over the world with HQ Poscasts, be my guest.

Jack

[Bill[_20_]]

In message >, Scott Dorsey
> writes
>The problem is that, unlike Lame, you can't get the Fraunhofer codec
>just as a standalone application that you can call from the command
>line. About as close as you can get is with one of the plugins that
>give you Fraunhofer encoding from your existing DAW software, like the
>Sonnox plugin.

I know I'm way out of date, but we used to use a command-line Fraunhofer
codec for file transfer in the broadcast software I was involved with in
the mid 90's.

In those early days, trying to get a smallish number of licences for the
codec was a nightmare. I rang Fraunhofer and managed to speak to what
sounded like the 'one man in an office' who promised to get back to me
when he had investigated. Of course, I'm still waiting.

I eventually managed to get a batch of licences via Xing (anyone
remember them?), but I've never been surprised that Fraunhofer messed up
their control of the legal side of things.

My complaint with the Fraunhofer codec in Audition is that it doesn't
comply with the mp3 design parameters, which say it should decode mp2 as
well.
--
Bill

[mcp6453[_2_]]

On 9/26/2015 8:03 AM, Bill wrote:

> My complaint with the Fraunhofer codec in Audition is that it doesn't comply with the mp3 design parameters, which say
> it should decode mp2 as well.

Does the version in iTunes comply? At this point, it looks like I may have to use iTunes, which is as bad as a virus.

[Mike Rivers[_2_]]

On 9/26/2015 8:03 AM, Bill wrote:
> In those early days, trying to get a smallish number of licences for the
> codec was a nightmare. I rang Fraunhofer and managed to speak to what
> sounded like the 'one man in an office' who promised to get back to me
> when he had investigated. Of course, I'm still waiting.
>
> I eventually managed to get a batch of licences via Xing (anyone
> remember them?), but I've never been surprised that Fraunhofer messed up
> their control of the legal side of things.

Fraunhofer is a developer. They aren't interested in selling licenses to
individual users, they want to license their technology to developers of
end products who, in turn, will work out a pricing scheme that,
hopefully, will allow them to make some money.


--
For a good time, visit http://mikeriversaudio.wordpress.com

[Bill[_20_]]

In message >, Mike Rivers >
writes
>Fraunhofer is a developer. They aren't interested in selling licenses
>to individual users, they want to license their technology to
>developers of end products who, in turn, will work out a pricing scheme
>that, hopefully, will allow them to make some money.

Well I was, sort of, a developer and was in the market for, say, 50 to
100 licences, and all I saw in the Fraunhofer literature was a request
for huge upfront payments that effectively ruled out selling the
technology to any small businesses.
I thought at the time that they were aiming for a per-encode rather than
per-encoder payment, which was never going to work.
--
Bill

[[email protected]]

And is it true that Fraunhofer *mandates* a
128kbps bitrate? Or are higher bitrates
acceptable at the risk of it not being gospel
Fraunhofer?

[John Williamson]

On 27/09/2015 19:04, wrote:
> And is it true that Fraunhofer *mandates* a
> 128kbps bitrate? Or are higher bitrates
> acceptable at the risk of it not being gospel
> Fraunhofer?
>
I don't know about mandating, as it's a while since I played with a
copy, but the earlier versions (And possibly the later versions also)
were apparently written to give the best sounding results on speech at
64kbps, to allow transmission in real time over a single ISDN line..
They were also initially implemented in hardware, with the computer
program coming later.

--
Tciao for Now!

John.

[Peter Larsen[_3_]]

On 27-09-2015 19:04, wrote:
> And is it true that Fraunhofer *mandates* a
> 128kbps bitrate? Or are higher bitrates
> acceptable at the risk of it not being gospel
> Fraunhofer?

Whatever you could possibly want is available when you use it from Cool
Edit or Audition.

Kind regards

Peter Larsen

[[email protected]]

On Monday, September 28, 2015 at 12:53:15 AM UTC-4, Peter Larsen wrote:
> On 27-09-2015 19:04, il.com wrote:
> > And is it true that Fraunhofer *mandates* a
> > 128kbps bitrate? Or are higher bitrates
> > acceptable at the risk of it not being gospel
> > Fraunhofer?
>
> Whatever you could possibly want is available when you use it from Cool
> Edit or Audition.
>
> Kind regards
>
> Peter Larsen
_____________

My point was I read somewhere that the 128kbps
bitrate was "most ideal" for the perceptual encoding
to be optimized.

[John Williamson]

On 28/09/2015 12:44, wrote:
> On Monday, September 28, 2015 at 12:53:15 AM UTC-4, Peter Larsen wrote:
>> On 27-09-2015 19:04, il.com wrote:
>>> And is it true that Fraunhofer *mandates* a
>>> 128kbps bitrate? Or are higher bitrates
>>> acceptable at the risk of it not being gospel
>>> Fraunhofer?
>>
>> Whatever you could possibly want is available when you use it from Cool
>> Edit or Audition.
>>
>> Kind regards
>>
>> Peter Larsen
> _____________
>
> My point was I read somewhere that the 128kbps
> bitrate was "most ideal" for the perceptual encoding
> to be optimized.
>
The standard lists acceptable bitrates from 32kbps to 320 kbps. The
Fraunhofer encoder used to sound best at 128kbps on music, with both
higher and lower bitrates sounding worse. LAME, which is *not* an mp3
encoder (The normal Linuxisti passion for "cleverness" claims that LAME
stands for LAME Ain't an Mp3 Encoder.), happens to produce mp3 player
compatible output files which sound better the higher the bitrate.

--
Tciao for Now!

John.

[[email protected]]

John Williamson wrote: "- show quoted text -
The standard lists acceptable bitrates from 32kbps to 320 kbps. The
Fraunhofer encoder used to sound best at 128kbps on music, with both
higher and lower bitrates sounding worse. LAME, which is *not* an mp3
encoder (The normal Linuxisti passion for "cleverness" claims that LAME
stands for LAME Ain't an Mp3 Encoder.), happens to produce mp3 player
compatible output files which sound better the higher the bitrate.
- show quoted text -"


THANK YOU John! :) And that is what I don't understand: How could
anything sound inferior at a higher bitrate? Was the original Faunhofer
designed to do its "mp3 thing" best at 128kbps?

[John Williamson]

On 28/09/2015 14:44, wrote:
> John Williamson wrote: "- show quoted text -
> The standard lists acceptable bitrates from 32kbps to 320 kbps. The
> Fraunhofer encoder used to sound best at 128kbps on music, with both
> higher and lower bitrates sounding worse. LAME, which is *not* an mp3
> encoder (The normal Linuxisti passion for "cleverness" claims that LAME
> stands for LAME Ain't an Mp3 Encoder.), happens to produce mp3 player
> compatible output files which sound better the higher the bitrate.
> - show quoted text -"
>
>
> THANK YOU John! :) And that is what I don't understand: How could
> anything sound inferior at a higher bitrate? Was the original Faunhofer
> designed to do its "mp3 thing" best at 128kbps?
>
The original mp3 hardware encoder was designed and built to produce the
best quality for speech at 64kbps, and was later rewritten to handle
music as well. There was a lot of research that went into designing the
perceptual encoding of the algorithm. It got at least one student a PhD,
and there are two different methods used at different bitrates and by
different encoders. They all produce a file that can be decoded by using
a lookup table for each possible frame content, which then is used to
produce an output bitstream to the DAC.

The principle of AAC and OGG is similar, but with different assumptions
as to what can be discarded by the encoder.

--
Tciao for Now!

John.

[Peter Larsen[_3_]]

On 28-09-2015 14:30, John Williamson wrote:

Someone using an alias wrote

>> My point was I read somewhere that the 128kbps
>> bitrate was "most ideal" for the perceptual encoding
>> to be optimized.

John supplemented with

> The standard lists acceptable bitrates from 32kbps to 320 kbps. The
> Fraunhofer encoder used to sound best at 128kbps on music, with both
> higher and lower bitrates sounding worse. LAME, which is *not* an mp3
> encoder (The normal Linuxisti passion for "cleverness" claims that LAME
> stands for LAME Ain't an Mp3 Encoder.), happens to produce mp3 player
> compatible output files which sound better the higher the bitrate.

There have been two versions of the Fraunhofer encoder. My experience is
that the only acceptable, albeit just barely, mp3 is variable wordlength
max quality, ms-stereo allowed, reduce separation not allowed.

I prefer to downsample to 32 kHz sample rate prior to encoding to reduce
the splattyness that comes from replacing treble with white noise, which
it appears to me that all perceptual coding does when I look at fft's of
the result.

Note that only decoding is standardized, encoding just has to fit it.

Kind regards

Peter Larsen

[[email protected]]

>
> I prefer to downsample to 32 kHz sample rate prior to encoding to reduce
> the splattyness that comes from replacing treble with white noise, which
> it appears to me that all perceptual coding does when I look at fft's of
> the result.
>
>

I post this link from time to time when this subject comes up.

http://www.avisoft.com/compression.htm

The audio samples are too fast for me to discern anything.. (it might be interesting to slow them down and listen) but the spectrographs are interesting to view.

Mark

[Peter Larsen[_3_]]

On 28-09-2015 19:59, wrote:

my statement

>> I prefer to downsample to 32 kHz sample rate prior to encoding to reduce
>> the splattyness that comes from replacing treble with white noise, which
>> it appears to me that all perceptual coding does when I look at fft's of
>> the result.

> I post this link from time to time when this subject comes up.

> http://www.avisoft.com/compression.htm

> The audio samples are too fast for me to discern anything.. (it might be interesting to slow them down and listen) but the spectrographs are interesting to view.

Extremely interesting, thank you!

> Mark

Kind regards

Peter Larsen

[JackA]

On Monday, September 28, 2015 at 2:59:14 PM UTC-4, wrote:
> >
> > I prefer to downsample to 32 kHz sample rate prior to encoding to reduce
> > the splattyness that comes from replacing treble with white noise, which
> > it appears to me that all perceptual coding does when I look at fft's of
> > the result.
> >
> >
>
> I post this link from time to time when this subject comes up.
>
> http://www.avisoft.com/compression.htm
>
> The audio samples are too fast for me to discern anything.. (it might be interesting to slow them down and listen) but the spectrographs are interesting to view.

Looks like script from the Mayan civilization!!! :-)

Jack

>
> Mark