Sign in with
Sign up | Sign in
Your question

MP3 quality

Last response: in Home Audio
Share
Anonymous
January 8, 2005 1:27:10 AM

Archived from groups: rec.audio.tech (More info?)

I know 128kbps is "CD quality" when converting CD music to MP3. If I use a
higher value, will I get substantially better sounding MP3 files?

If so, what's the point of diminishing returns -- the point where the increase
in file size outweighs the quality imporvement?

TIA.
- Henry

More about : mp3 quality

Anonymous
January 8, 2005 1:27:11 AM

Archived from groups: rec.audio.tech (More info?)

On 1/7/2005 14:27, BandHPhoto wrote:

> I know 128kbps is "CD quality" when converting CD music to MP3. If I use a
> higher value, will I get substantially better sounding MP3 files?

No. The raw RIFF (wave) file is cd-quality. Any non-lossless compression
(as MP3 is) will result in something less than CD-quality.

Whether you can tell the difference is a matter of great
subjectivity. You can find out whether others can tell the
difference, but that won't really help you. To find out
for yourself, you really need to try it.

For me, I took various tracks from different CDs, and compressed
them using various bit rates (all MP3s, as that was the only
compression format I wanted to work with) and compared them in
blind tests, in my stereo system. Both my Wife and I were able
to tell when the track was played from the original RIFF file,
and when it was being played from the MP3 files.

Note, however, that the difference between the two was pretty
small.

>
> If so, what's the point of diminishing returns -- the point where the increase
> in file size outweighs the quality imporvement?

For me, space for the files was not an issue. I have all my
library of CDs scraped to a pretty large disk on my server
machine. If that disk ever runs out of space (hasn't come
close yet) I'll just buy a bigger disk (as they continue to
get cheaper...).

Because space was no issue, I opted to go with raw RIFF files
rather than compressing them at all.

Note that there are some lossless compression schemes out there,
like Org Vorbis (sp?), I believe. Once a compressed file is
converted back to it's original form (in order to play it) it
should match the original (RIFF) data exactly.

>
> TIA.
> - Henry


--
Mark Hansen
Anonymous
January 8, 2005 1:52:40 AM

Archived from groups: rec.audio.tech (More info?)

bandhphoto@aol.com (BandHPhoto) wrote in
news:20050107172710.23235.00002985@mb-m03.aol.com:

> I know 128kbps is "CD quality" when converting CD music to MP3. If I use
a
> higher value, will I get substantially better sounding MP3 files?
>
> If so, what's the point of diminishing returns -- the point where the
increase
> in file size outweighs the quality imporvement?
>
> TIA.
> - Henry
>

128k MP3 is not CD quality. 320k MP3 is not CD quality either. What
losses you can tolerate is a matter of the quality of your equipment, your
selection of program material, and what your ears tell you. Some say 128
is good enough and others say 320 is good enough but be assured that MP3
is not CD quality no matter what the encoding rate.

File size is not going to be an issue soon. Soon we will look back at
MP3s the same way we look at RLL encoding now.

r



--
Nothing beats the bandwidth of a station wagon filled with DLT tapes.
Related resources
Anonymous
January 8, 2005 1:52:41 AM

Archived from groups: rec.audio.tech (More info?)

"R"

> What
> losses you can tolerate is a matter of the quality of your equipment, your
> selection of program material, and what your ears tell you. Some say 128
> is good enough and others say 320 is good enough...

Isn't the compression algorithm you are using also significant? Windows
Media at 128 may or may not sound the same as i-tunes, for example. There
is nothing that states the efficiency or quality of the compression and
decompression of the MP3 track - correct?

So, what encoder/decoder you use is also significant.

What's disturbing about this thread is that there are likely 1000's of first
graders out there with new x-mas presents discussing the same thing...



- Nate
Anonymous
January 8, 2005 1:52:42 AM

Archived from groups: rec.audio.tech (More info?)

"Nate" wrote ...
> Isn't the compression algorithm you are using also significant?
> Windows Media at 128 may or may not sound the same as i-tunes,
> for example.

Huh? MP3 is MP3. WMA is WMA, and AAC is AAC.

WMA and/or AAC may be better (or worse) than MP3, at
any particular bitrate, but neither *is* MP3, which was the
original question (see the subject line).

> There is nothing that states the efficiency or quality of the
> compression and decompression of the MP3 track -
> correct?

Most of us use the bitrate as a simple indicator of quality.

> So, what encoder/decoder you use is also significant.

Yes, there is a selection of different encoders for MP3.
But not for WMA or AAC, etc. that I've ever heard of.

> What's disturbing about this thread is that there are likely
> 1000's of first graders out there with new x-mas presents
> discussing the same thing...

That's discouraging. So are you suggesting that we old-
timers (people in their 20s and up) just hang it up? :-)
Anonymous
January 8, 2005 2:00:32 AM

Archived from groups: rec.audio.tech (More info?)

BandHPhoto <bandhphoto@aol.com> wrote:
> I know 128kbps is "CD quality" when converting CD music to MP3.

er...no, it usually isn't.

> If I use a
> higher value, will I get substantially better sounding MP3 files?


Quite likely, yes.


> If so, what's the point of diminishing returns -- the point where the increase
> in file size outweighs the quality imporvement?

~190 kpbs variable bitrate (i.e., the --alt preset standard using LAME 3.90.3)
is an excellent compromise between size and quality, IMO



--

-S
If you're a nut and knock on enough doors, eventually someone will open one,
look at you and say, Messiah, we have waited for your arrival.
Anonymous
January 8, 2005 2:04:46 AM

Archived from groups: rec.audio.tech (More info?)

R <spmaway@ylhoo.com> wrote:
> bandhphoto@aol.com (BandHPhoto) wrote in
> news:20050107172710.23235.00002985@mb-m03.aol.com:

> > I know 128kbps is "CD quality" when converting CD music to MP3. If I use
> a
> > higher value, will I get substantially better sounding MP3 files?
> >
> > If so, what's the point of diminishing returns -- the point where the
> increase
> > in file size outweighs the quality imporvement?
> >
> > TIA.
> > - Henry
> >

> 128k MP3 is not CD quality. 320k MP3 is not CD quality either. What
> losses you can tolerate is a matter of the quality of your equipment, your
> selection of program material, and what your ears tell you. Some say 128
> is good enough and others say 320 is good enough but be assured that MP3
> is not CD quality no matter what the encoding rate.

In those cases where mp3 is audibly transparent compared to CD source in
an ABX test -- and it's not hard to find them at 320 kbps if you're willing
to do them -- it is effectively CD quality.


> File size is not going to be an issue soon. Soon we will look back at
> MP3s the same way we look at RLL encoding now.

Quite so. But until then it's certainly possible to make MP3s that
are audibly 'CD quality'...though verifying that that's true
for any given compressed track requires the listener to do
a proper comparison.




--

-S
If you're a nut and knock on enough doors, eventually someone will open one,
look at you and say, Messiah, we have waited for your arrival.
January 8, 2005 3:08:08 AM

Archived from groups: rec.audio.tech (More info?)

BandHPhoto wrote:
> I know 128kbps is "CD quality" when converting CD music to MP3.

Not to people that know how to use their ears.
128 sounds horrible on most material.

> If I use a
> higher value, will I get substantially better sounding MP3 files?

Yes.

> If so, what's the point of diminishing returns -- the point where the increase
> in file size outweighs the quality imporvement?

If you're encoding above 224kbps (It still doesn't sound the same as the
CD but getting close sometimes depending on the material) you're better
of using non-lossy compression like FLAC.


Sander
Anonymous
January 8, 2005 3:08:09 AM

Archived from groups: rec.audio.tech (More info?)

"Sander" wrote ...
> If you're encoding above 224kbps (It still doesn't sound
> the same as the CD but getting close sometimes depending
> on the material) you're better of using non-lossy compression
> like FLAC.

Makes sense for efficient storage (like archiving), but do any
portable players decode FLAC?
Anonymous
January 8, 2005 5:12:10 AM

Archived from groups: rec.audio.tech (More info?)

"Richard Crowley" <rcrowley7@xprt.net> a écrit dans le message de news:
10tuepqimbp6s89@corp.supernews.com...
> "Nate" wrote ...
>> Isn't the compression algorithm you are using also significant? Windows
>> Media at 128 may or may not sound the same as i-tunes, for example.
>
> Huh? MP3 is MP3. WMA is WMA, and AAC is AAC.
>

There ARE differences between compressions algorithms. Lame, Blade, etc...
Wether these differences are audible or not, would have to be tested in
DBT's, but there ARE differences between algorithms, differences in the code
that they consist of. Saying "MP3 is MP3" is just not precise enough.... If
you want to affirm something as if it was "The Truth", check your facts. If
you're just expressing an opinion, make it clear.
Anonymous
January 8, 2005 8:40:01 AM

Archived from groups: rec.audio.tech (More info?)

Steven Sullivan <ssully@panix.com> wrote in
news:crn4ie$if1$3@reader1.panix.com:

> R <spmaway@ylhoo.com> wrote:
>> bandhphoto@aol.com (BandHPhoto) wrote in
>> news:20050107172710.23235.00002985@mb-m03.aol.com:
>
>> > I know 128kbps is "CD quality" when converting CD music to MP3. If I
>> > use
>> a
>> > higher value, will I get substantially better sounding MP3 files?
>> >
>> > If so, what's the point of diminishing returns -- the point where the
>> increase
>> > in file size outweighs the quality imporvement?
>> >
>> > TIA.
>> > - Henry
>> >
>
>> 128k MP3 is not CD quality. 320k MP3 is not CD quality either. What
>> losses you can tolerate is a matter of the quality of your equipment,
>> your selection of program material, and what your ears tell you. Some
>> say 128 is good enough and others say 320 is good enough but be assured
>> that MP3 is not CD quality no matter what the encoding rate.
>
> In those cases where mp3 is audibly transparent compared to CD source in
> an ABX test -- ...it is effectively CD quality.

Yes, 320 MP3 can be the same as CD but those cases are rarer than one
might think.

The last guy to make the claim "320kc MP3 is as good as a CD" wound up
retracting his statement. He did admit the difference was slight, but
there was a difference and it was noticable enough to recognize rather
easily. It does require good program material and a good
reproduction system relatively flat from 20hz-20kc that is also low in
distortion. At 320kc one has to make the determination of "it is good
enough". What defines "good enough" for me may not be the same as someone
else's definition of "good enough" and vice versa.

r


--
Nothing beats the bandwidth of a station wagon filled with DLT tapes.
Anonymous
January 8, 2005 10:45:00 AM

Archived from groups: rec.audio.tech (More info?)

On 07 Jan 2005 22:27:10 GMT, bandhphoto@aol.com (BandHPhoto) wrote:

>I know 128kbps is "CD quality" when converting CD music to MP3. If I use a
>higher value, will I get substantially better sounding MP3 files?

Your knowledge is defective - try 192 *minimum* for 'CD quality' - and
that's only on undemanding material.

>If so, what's the point of diminishing returns -- the point where the increase
>in file size outweighs the quality imporvement?

Most people find anything above 256kb/sec to be sonically transparent
in most cases, although some types of music will catch out almost any
compression algorithm at some point. More importantly, a 20GB machine
will hold at least 30 CDs (600 tracks, if you will) recorded at full
quality with no compression. Are you telling me that you *need* more
than 600 tunes stored in a portable device?
--

Stewart Pinkerton | Music is Art - Audio is Engineering
Anonymous
January 8, 2005 9:50:17 PM

Archived from groups: rec.audio.tech (More info?)

>> If I use a
>> higher value, will I get substantially better sounding MP3 files?
>
>Yes.
>
>> If so, what's the point of diminishing returns -- the point where the
>increase
>> in file size outweighs the quality imporvement?
>
>If you're encoding above 224kbps (It still doesn't sound the same as the
>CD but getting close sometimes depending on the material) you're better
>of using non-lossy compression

Thank you. This was very useful.
- Henry
Anonymous
January 9, 2005 3:54:55 AM

Archived from groups: rec.audio.tech (More info?)

Richard Crowley <rcrowley7@xprt.net> wrote:
> "Sander" wrote ...
> > If you're encoding above 224kbps (It still doesn't sound
> > the same as the CD but getting close sometimes depending
> > on the material) you're better of using non-lossy compression
> > like FLAC.

> Makes sense for efficient storage (like archiving), but do any
> portable players decode FLAC?


The Rio Karma and iAudio M3 do.


--

-S
If you're a nut and knock on enough doors, eventually someone will open one,
look at you and say, Messiah, we have waited for your arrival.
Anonymous
January 9, 2005 4:00:28 AM

Archived from groups: rec.audio.tech (More info?)

R <spmaway@ylhoo.com> wrote:
> Steven Sullivan <ssully@panix.com> wrote in
> news:crn4ie$if1$3@reader1.panix.com:

> > R <spmaway@ylhoo.com> wrote:
> >> bandhphoto@aol.com (BandHPhoto) wrote in
> >> news:20050107172710.23235.00002985@mb-m03.aol.com:
> >
> >> > I know 128kbps is "CD quality" when converting CD music to MP3. If I
> >> > use
> >> a
> >> > higher value, will I get substantially better sounding MP3 files?
> >> >
> >> > If so, what's the point of diminishing returns -- the point where the
> >> increase
> >> > in file size outweighs the quality imporvement?
> >> >
> >> > TIA.
> >> > - Henry
> >> >
> >
> >> 128k MP3 is not CD quality. 320k MP3 is not CD quality either. What
> >> losses you can tolerate is a matter of the quality of your equipment,
> >> your selection of program material, and what your ears tell you. Some
> >> say 128 is good enough and others say 320 is good enough but be assured
> >> that MP3 is not CD quality no matter what the encoding rate.
> >
> > In those cases where mp3 is audibly transparent compared to CD source in
> > an ABX test -- ...it is effectively CD quality.

> Yes, 320 MP3 can be the same as CD but those cases are rarer than one
> might think.

> The last guy to make the claim "320kc MP3 is as good as a CD" wound up
> retracting his statement. He did admit the difference was slight, but
> there was a difference and it was noticable enough to recognize rather
> easily. It does require good program material and a good
> reproduction system relatively flat from 20hz-20kc that is also low in
> distortion. At 320kc one has to make the determination of "it is good
> enough". What defines "good enough" for me may not be the same as someone
> else's definition of "good enough" and vice versa.


What is 'good material', in this context? Some very difficult samples are
documented to be ABX-able at 320 kbps (see the thread about this I started
over on hydrogenaudio.org a few days ago, as 'krabapple') but unless
you *define* 'good material' as that which is ABX-able at 320 kbps, I don;t
see the meaning. It isn't true, AFAIK, that recording quality or source
correlates to ABX-ability of mp3s.

It is true, though, that one would have to make the determination oneself,
using a controlled comparison, to verify 'transparency' subjectively.


--

-S
If you're a nut and knock on enough doors, eventually someone will open one,
look at you and say, Messiah, we have waited for your arrival.
Anonymous
January 9, 2005 8:35:29 AM

Archived from groups: rec.audio.tech (More info?)

Steven Sullivan <ssully@panix.com> wrote in
news:crpvnc$jcv$10@reader1.panix.com:

> R <spmaway@ylhoo.com> wrote:
>> Steven Sullivan <ssully@panix.com> wrote in
>> news:crn4ie$if1$3@reader1.panix.com:
>
>> > R <spmaway@ylhoo.com> wrote:
>> >> bandhphoto@aol.com (BandHPhoto) wrote in
>> >> news:20050107172710.23235.00002985@mb-m03.aol.com:
>> >
>> >> > I know 128kbps is "CD quality" when converting CD music to MP3. If
>> >> > I use
>> >> a
>> >> > higher value, will I get substantially better sounding MP3 files?
>> >> >
>> >> > If so, what's the point of diminishing returns -- the point where
>> >> > the
>> >> increase
>> >> > in file size outweighs the quality imporvement?
>> >> >
>> >> > TIA.
>> >> > - Henry
>> >> >
>> >
>> >> 128k MP3 is not CD quality. 320k MP3 is not CD quality either.
>> >> What losses you can tolerate is a matter of the quality of your
>> >> equipment, your selection of program material, and what your ears
>> >> tell you. Some say 128 is good enough and others say 320 is good
>> >> enough but be assured that MP3 is not CD quality no matter what the
>> >> encoding rate.
>> >
>> > In those cases where mp3 is audibly transparent compared to CD source
>> > in an ABX test -- ...it is effectively CD quality.
>
>> Yes, 320 MP3 can be the same as CD but those cases are rarer than one
>> might think.
>
>> The last guy to make the claim "320kc MP3 is as good as a CD" wound up
>> retracting his statement. He did admit the difference was slight, but
>> there was a difference and it was noticable enough to recognize rather
>> easily. It does require good program material and a good
>> reproduction system relatively flat from 20hz-20kc that is also low in
>> distortion. At 320kc one has to make the determination of "it is good
>> enough". What defines "good enough" for me may not be the same as
>> someone else's definition of "good enough" and vice versa.
>
>
> What is 'good material', in this context? Some very difficult samples
> are documented to be ABX-able at 320 kbps (see the thread about this I
> started over on hydrogenaudio.org a few days ago, as 'krabapple') but
> unless you *define* 'good material' as that which is ABX-able at 320
> kbps, I don;t see the meaning. It isn't true, AFAIK, that recording
> quality or source correlates to ABX-ability of mp3s.
>
> It is true, though, that one would have to make the determination
> oneself, using a controlled comparison, to verify 'transparency'
> subjectively.
>
>

Good material would be something that had some very high frequency
information in it as well as very low. I have heard some program material
that would be most inappropriate for a test. That is to say that it
contained little high frequency information. I have heard a fair amount
that contains little very low frequency information. If one never listens
to music that has neither highs nor lows, then the point is moot.

Not sure what you mean by ABX-able. One could ABX a violin and a piano
but it would be rather pointless.

r


--
Nothing beats the bandwidth of a station wagon filled with DLT tapes.
Anonymous
January 9, 2005 8:35:30 AM

Archived from groups: rec.audio.tech (More info?)

On Sun, 09 Jan 2005 05:35:29 GMT, R <spmaway@ylhoo.com> wrote:
>Good material would be something that had some very high frequency
>information in it as well as very low. I have heard some program material
>that would be most inappropriate for a test. That is to say that it
>contained little high frequency information. I have heard a fair amount
>that contains little very low frequency information. If one never listens
>to music that has neither highs nor lows, then the point is moot.

>Not sure what you mean by ABX-able. One could ABX a violin and a piano
>but it would be rather pointless.

For testing mp3 encoding/decoding, you need something more. It doesn't take
much bitrate to encode 20 to 20khz.

For critiqueing mp3s, listen for timbre. Accoustic guitar, or chamber
orchestra really can show the limitations of mp3 done at too low a bitrate.
Anonymous
January 9, 2005 10:08:11 AM

Archived from groups: rec.audio.tech (More info?)

"R" <spmaway@ylhoo.com> wrote in message
news:Xns95D8EFFCD1E4mc2500183316chgoill@10.232.1.1...
> Steven Sullivan <ssully@panix.com> wrote in
> news:crpvnc$jcv$10@reader1.panix.com:
>
>> R <spmaway@ylhoo.com> wrote:
>>> Steven Sullivan <ssully@panix.com> wrote in
>>> news:crn4ie$if1$3@reader1.panix.com:
>>
>>> > R <spmaway@ylhoo.com> wrote:
>>> >> bandhphoto@aol.com (BandHPhoto) wrote in
>>> >> news:20050107172710.23235.00002985@mb-m03.aol.com:
>>> >
>>> >> > I know 128kbps is "CD quality" when converting CD music to MP3. If
>>> >> > I use
>>> >> a
>>> >> > higher value, will I get substantially better sounding MP3 files?
>>> >> >
>>> >> > If so, what's the point of diminishing returns -- the point where
>>> >> > the
>>> >> increase
>>> >> > in file size outweighs the quality imporvement?
>>> >> >
>>> >> > TIA.
>>> >> > - Henry
>>> >> >
>>> >
>>> >> 128k MP3 is not CD quality. 320k MP3 is not CD quality either.
>>> >> What losses you can tolerate is a matter of the quality of your
>>> >> equipment, your selection of program material, and what your ears
>>> >> tell you. Some say 128 is good enough and others say 320 is good
>>> >> enough but be assured that MP3 is not CD quality no matter what the
>>> >> encoding rate.
>>> >
>>> > In those cases where mp3 is audibly transparent compared to CD source
>>> > in an ABX test -- ...it is effectively CD quality.
>>
>>> Yes, 320 MP3 can be the same as CD but those cases are rarer than one
>>> might think.
>>
>>> The last guy to make the claim "320kc MP3 is as good as a CD" wound up
>>> retracting his statement. He did admit the difference was slight, but
>>> there was a difference and it was noticable enough to recognize rather
>>> easily. It does require good program material and a good
>>> reproduction system relatively flat from 20hz-20kc that is also low in
>>> distortion. At 320kc one has to make the determination of "it is good
>>> enough". What defines "good enough" for me may not be the same as
>>> someone else's definition of "good enough" and vice versa.
>>
>>
>> What is 'good material', in this context? Some very difficult samples
>> are documented to be ABX-able at 320 kbps (see the thread about this I
>> started over on hydrogenaudio.org a few days ago, as 'krabapple') but
>> unless you *define* 'good material' as that which is ABX-able at 320
>> kbps, I don;t see the meaning. It isn't true, AFAIK, that recording
>> quality or source correlates to ABX-ability of mp3s.
>>
>> It is true, though, that one would have to make the determination
>> oneself, using a controlled comparison, to verify 'transparency'
>> subjectively.
>>
>>
>
> Good material would be something that had some very high frequency
> information in it as well as very low. I have heard some program material
> that would be most inappropriate for a test. That is to say that it
> contained little high frequency information. I have heard a fair amount
> that contains little very low frequency information. If one never listens
> to music that has neither highs nor lows, then the point is moot.
>
> Not sure what you mean by ABX-able. One could ABX a violin and a piano
> but it would be rather pointless.

Castanets are extremely difficult for any perceptual coder bit rate reducing
algorithm to deal with because the impulse-type sound has very little to
mask the noise due to the bit rate reduction. If you ABX the original and
the compressed sound, you should easily hear the difference. Start with low
bit rates. Once you know what to listen for, you'll be able to detect
artifacts at much higher bit rates than you might think, even with other
program material. You may never go back to compressed audio again :-)
Anonymous
January 9, 2005 10:58:03 AM

Archived from groups: rec.audio.tech (More info?)

"Karl Uppiano" <karl.uppiano@verizon.net> wrote in
news:vL4Ed.3818$hc7.1593@trnddc08:

>
> "R" <spmaway@ylhoo.com> wrote in message
> news:Xns95D8EFFCD1E4mc2500183316chgoill@10.232.1.1...
>> Steven Sullivan <ssully@panix.com> wrote in
>> news:crpvnc$jcv$10@reader1.panix.com:
>>
>>> R <spmaway@ylhoo.com> wrote:
>>>> Steven Sullivan <ssully@panix.com> wrote in
>>>> news:crn4ie$if1$3@reader1.panix.com:
>>>
>>>> > R <spmaway@ylhoo.com> wrote:
>>>> >> bandhphoto@aol.com (BandHPhoto) wrote in
>>>> >> news:20050107172710.23235.00002985@mb-m03.aol.com:
>>>> >
>>>> >> > I know 128kbps is "CD quality" when converting CD music to MP3.
>>>> >> > If I use
>>>> >> a
>>>> >> > higher value, will I get substantially better sounding MP3
>>>> >> > files?
>>>> >> >
>>>> >> > If so, what's the point of diminishing returns -- the point
>>>> >> > where the
>>>> >> increase
>>>> >> > in file size outweighs the quality imporvement?
>>>> >> >
>>>> >> > TIA.
>>>> >> > - Henry
>>>> >> >
>>>> >
>>>> >> 128k MP3 is not CD quality. 320k MP3 is not CD quality either.
>>>> >> What losses you can tolerate is a matter of the quality of your
>>>> >> equipment, your selection of program material, and what your ears
>>>> >> tell you. Some say 128 is good enough and others say 320 is good
>>>> >> enough but be assured that MP3 is not CD quality no matter what
>>>> >> the encoding rate.
>>>> >
>>>> > In those cases where mp3 is audibly transparent compared to CD
>>>> > source in an ABX test -- ...it is effectively CD quality.
>>>
>>>> Yes, 320 MP3 can be the same as CD but those cases are rarer than one
>>>> might think.
>>>
>>>> The last guy to make the claim "320kc MP3 is as good as a CD" wound
>>>> up retracting his statement. He did admit the difference was slight,
>>>> but there was a difference and it was noticable enough to recognize
>>>> rather easily. It does require good program material and a good
>>>> reproduction system relatively flat from 20hz-20kc that is also low
>>>> in distortion. At 320kc one has to make the determination of "it is
>>>> good enough". What defines "good enough" for me may not be the same
>>>> as someone else's definition of "good enough" and vice versa.
>>>
>>>
>>> What is 'good material', in this context? Some very difficult samples
>>> are documented to be ABX-able at 320 kbps (see the thread about this I
>>> started over on hydrogenaudio.org a few days ago, as 'krabapple') but
>>> unless you *define* 'good material' as that which is ABX-able at 320
>>> kbps, I don;t see the meaning. It isn't true, AFAIK, that recording
>>> quality or source correlates to ABX-ability of mp3s.
>>>
>>> It is true, though, that one would have to make the determination
>>> oneself, using a controlled comparison, to verify 'transparency'
>>> subjectively.
>>>
>>>
>>
>> Good material would be something that had some very high frequency
>> information in it as well as very low. I have heard some program
>> material that would be most inappropriate for a test. That is to say
>> that it contained little high frequency information. I have heard a
>> fair amount that contains little very low frequency information. If
>> one never listens to music that has neither highs nor lows, then the
>> point is moot.
>>
>> Not sure what you mean by ABX-able. One could ABX a violin and a piano
>> but it would be rather pointless.
>
> Castanets are extremely difficult for any perceptual coder bit rate
> reducing algorithm to deal with because the impulse-type sound has very
> little to mask the noise due to the bit rate reduction. If you ABX the
> original and the compressed sound, you should easily hear the
> difference. Start with low bit rates. Once you know what to listen for,
> you'll be able to detect artifacts at much higher bit rates than you
> might think, even with other program material. You may never go back to
> compressed audio again :-)
>
>
>

I knew those de Falla recordings would come in handy one day. (:>)

r

Music isn't bad either. (:>)

--
Nothing beats the bandwidth of a station wagon filled with DLT tapes.
Anonymous
January 9, 2005 11:09:42 AM

Archived from groups: rec.audio.tech (More info?)

"R" <spmaway@ylhoo.com> wrote in message
news:Xns95D8EFFCD1E4mc2500183316chgoill@10.232.1.1

> Not sure what you mean by ABX-able.

Seems pretty obvious. You round-trip a .wav file through a given MP3 bitrate
and ABX the before and after .wav files.

Please see http://www.pcabx.com/product/coder_decoder/index.htm for worked
out examples.

> One could ABX a violin and a piano but it would be rather pointless.

Pointless? So is this ludicrous comment!
Anonymous
January 10, 2005 12:44:23 AM

Archived from groups: rec.audio.tech (More info?)

"BandHPhoto" <bandhphoto@aol.com> wrote in message
news:20050107172710.23235.00002985@mb-m03.aol.com...
> I know 128kbps is "CD quality" when converting CD music to MP3. If I use a
> higher value, will I get substantially better sounding MP3 files?
>
> If so, what's the point of diminishing returns -- the point where the
increase
> in file size outweighs the quality imporvement?
>

128kb/s is 'near CD' quality. 192kb/s or higher is accepted to be 'true CD
quality'. If you want to be sure you are getting CD quality use 256kb/s or
VBR (variable bit rate) conversion.
Anonymous
January 10, 2005 9:33:38 AM

Archived from groups: rec.audio.tech (More info?)

R <spmaway@ylhoo.com> wrote:
> Steven Sullivan <ssully@panix.com> wrote in
> news:crpvnc$jcv$10@reader1.panix.com:

> > R <spmaway@ylhoo.com> wrote:
> >> Steven Sullivan <ssully@panix.com> wrote in
> >> news:crn4ie$if1$3@reader1.panix.com:
> >
> >> > R <spmaway@ylhoo.com> wrote:
> >> >> bandhphoto@aol.com (BandHPhoto) wrote in
> >> >> news:20050107172710.23235.00002985@mb-m03.aol.com:
> >> >
> >> >> > I know 128kbps is "CD quality" when converting CD music to MP3. If
> >> >> > I use
> >> >> a
> >> >> > higher value, will I get substantially better sounding MP3 files?
> >> >> >
> >> >> > If so, what's the point of diminishing returns -- the point where
> >> >> > the
> >> >> increase
> >> >> > in file size outweighs the quality imporvement?
> >> >> >
> >> >> > TIA.
> >> >> > - Henry
> >> >> >
> >> >
> >> >> 128k MP3 is not CD quality. 320k MP3 is not CD quality either.
> >> >> What losses you can tolerate is a matter of the quality of your
> >> >> equipment, your selection of program material, and what your ears
> >> >> tell you. Some say 128 is good enough and others say 320 is good
> >> >> enough but be assured that MP3 is not CD quality no matter what the
> >> >> encoding rate.
> >> >
> >> > In those cases where mp3 is audibly transparent compared to CD source
> >> > in an ABX test -- ...it is effectively CD quality.
> >
> >> Yes, 320 MP3 can be the same as CD but those cases are rarer than one
> >> might think.
> >
> >> The last guy to make the claim "320kc MP3 is as good as a CD" wound up
> >> retracting his statement. He did admit the difference was slight, but
> >> there was a difference and it was noticable enough to recognize rather
> >> easily. It does require good program material and a good
> >> reproduction system relatively flat from 20hz-20kc that is also low in
> >> distortion. At 320kc one has to make the determination of "it is good
> >> enough". What defines "good enough" for me may not be the same as
> >> someone else's definition of "good enough" and vice versa.
> >
> >
> > What is 'good material', in this context? Some very difficult samples
> > are documented to be ABX-able at 320 kbps (see the thread about this I
> > started over on hydrogenaudio.org a few days ago, as 'krabapple') but
> > unless you *define* 'good material' as that which is ABX-able at 320
> > kbps, I don;t see the meaning. It isn't true, AFAIK, that recording
> > quality or source correlates to ABX-ability of mp3s.
> >
> > It is true, though, that one would have to make the determination
> > oneself, using a controlled comparison, to verify 'transparency'
> > subjectively.
> >
> >

> Good material would be something that had some very high frequency
> information in it as well as very low. I have heard some program material
> that would be most inappropriate for a test. That is to say that it
> contained little high frequency information. I have heard a fair amount
> that contains little very low frequency information. If one never listens
> to music that has neither highs nor lows, then the point is moot.


Having 'very high' and 'very low' frequencies together, does not
seem to be what distinguishes the most difficult samples to encode.

What constitutes 'very low' and 'very high'? UNtil the advent of CD,
which could render sounds down to 20 Hz accurately, there wasn't much
in the way of recordings that had extended low frequency information.
Loud, low bass was especially tough to get into vinyl.

> Not sure what you mean by ABX-able. One could ABX a violin and a piano
> but it would be rather pointless.

It would not be pointless for mp3s though...certainly not for mp3s at 320 kbps.






--

-S
If you're a nut and knock on enough doors, eventually someone will open one,
look at you and say, Messiah, we have waited for your arrival.
Anonymous
January 10, 2005 9:37:37 AM

Archived from groups: rec.audio.tech (More info?)

Arny Krueger <arnyk@hotpop.com> wrote:
> "R" <spmaway@ylhoo.com> wrote in message
> news:Xns95D8EFFCD1E4mc2500183316chgoill@10.232.1.1

> > Not sure what you mean by ABX-able.

> Seems pretty obvious. You round-trip a .wav file through a given MP3 bitrate
> and ABX the before and after .wav files.

> Please see http://www.pcabx.com/product/coder_decoder/index.htm for worked
> out examples.

> > One could ABX a violin and a piano but it would be rather pointless.

> Pointless? So is this ludicrous comment!

Indeed.

A collection of hard-to encode samples, used by codec developers to
improve mp3 codecs, may be found at

http://lame.sourceforge.net/gpsycho/quality.html]http://lame.sourceforge.net/gpsycho/quality.html





--

-S
If you're a nut and knock on enough doors, eventually someone will open one,
look at you and say, Messiah, we have waited for your arrival.
Anonymous
January 10, 2005 12:27:07 PM

Archived from groups: rec.audio.tech (More info?)

"Steven Sullivan" <ssully@panix.com> wrote in message
news:crt7rh$pf5$2@reader1.panix.com
> Arny Krueger <arnyk@hotpop.com> wrote:
>> "R" <spmaway@ylhoo.com> wrote in message
>> news:Xns95D8EFFCD1E4mc2500183316chgoill@10.232.1.1
>
>>> Not sure what you mean by ABX-able.
>
>> Seems pretty obvious. You round-trip a .wav file through a given MP3
>> bitrate and ABX the before and after .wav files.
>
>> Please see http://www.pcabx.com/product/coder_decoder/index.htm for
>> worked out examples.
>
>>> One could ABX a violin and a piano but it would be rather pointless.
>
>> Pointless? So is this ludicrous comment!

> Indeed.

> A collection of hard-to encode samples, used by codec developers to
> improve mp3 codecs, may be found at

> http://lame.sourceforge.net/gpsycho/quality.html

Looks like a pretty good collection.

My, more generally-intended collection is at:

http://www.pcabx.com/technical/reference/index.htm

and

http://www.pcabx.com/technical/sample_rates/index.htm
Anonymous
January 10, 2005 12:31:44 PM

Archived from groups: rec.audio.tech (More info?)

"Steven Sullivan" <ssully@panix.com> wrote in message
news:crt7k2$pf5$1@reader1.panix.com

> R <spmaway@ylhoo.com> wrote:

>> Good material would be something that had some very high frequency
>> information in it as well as very low. I have heard some program
>> material that would be most inappropriate for a test. That is to
>> say that it contained little high frequency information. I have
>> heard a fair amount that contains little very low frequency
>> information. If one never listens to music that has neither highs
>> nor lows, then the point is moot.

> Having 'very high' and 'very low' frequencies together, does not
> seem to be what distinguishes the most difficult samples to encode.

Agreed.

> What constitutes 'very low' and 'very high'? UNtil the advent of CD,
> which could render sounds down to 20 Hz accurately, there wasn't much
> in the way of recordings that had extended low frequency information.
> Loud, low bass was especially tough to get into vinyl.

There are many collections of stuff that someone says is either good for
auditioning, or good for breaking certain kinds of audio technology. (two
different things)

I've reviewed a lot of it,and there are common threads in places, but its
hard to simply categorize it all. There are alot of angles to play.
Anonymous
January 10, 2005 1:51:16 PM

Archived from groups: rec.audio.tech (More info?)

On Fri, 7 Jan 2005 17:45:29 -0800, "Richard Crowley"
<rcrowley7@xprt.net> wrote:

>> Isn't the compression algorithm you are using also significant?
>> Windows Media at 128 may or may not sound the same as i-tunes,
>> for example.
>
>Huh? MP3 is MP3. WMA is WMA, and AAC is AAC.
>
>WMA and/or AAC may be better (or worse) than MP3, at
>any particular bitrate, but neither *is* MP3, which was the
>original question (see the subject line).

No. There are several MP3 codecs. They output different code. Some
say that they sound different.
Anonymous
January 10, 2005 1:51:17 PM

Archived from groups: rec.audio.tech (More info?)

"Laurence Payne" <l@laurenceDELETEpayne.freeserve.co.uk> wrote in
message news:ken4u05s13aruq4qv7gbv6b0eiqgu1a172@4ax.com...
> On Fri, 7 Jan 2005 17:45:29 -0800, "Richard Crowley"
> <rcrowley7@xprt.net> wrote:
>
>>> Isn't the compression algorithm you are using also significant?
>>> Windows Media at 128 may or may not sound the same as i-tunes,
>>> for example.
>>
>>Huh? MP3 is MP3. WMA is WMA, and AAC is AAC.
>>
>>WMA and/or AAC may be better (or worse) than MP3, at
>>any particular bitrate, but neither *is* MP3, which was the
>>original question (see the subject line).
>
> No. There are several MP3 codecs. They output different
> code. Some say that they sound different.

Exactly what I said. Perhaps you have an alternate definition
of the word "no".
Anonymous
January 10, 2005 4:35:00 PM

Archived from groups: rec.audio.tech (More info?)

"Henk Boonsma" <hboonsma@teranet.news> wrote in message
news:1105303438.d62a941d4e000e9b02ddd9dabd4edab2@teranews...
> If you want to be sure you are getting CD quality use 256kb/s or
> VBR (variable bit rate) conversion.

As others have pointed out, if you want to be sure of CD quality, use 1.4Mbs
rate, or a bit less if you use FLAC.

MrT.
January 10, 2005 6:22:40 PM

Archived from groups: rec.audio.tech (More info?)

On 2005-01-07, BandHPhoto <bandhphoto@aol.com> wrote:
> I know 128kbps is "CD quality" when converting CD music to MP3. If I use a
> higher value, will I get substantially better sounding MP3 files?
>
> If so, what's the point of diminishing returns -- the point where the increase
> in file size outweighs the quality imporvement?
>
> TIA.
> - Henry

128mbps is not CD quality to my ears.
I suggest you carry out your own tests and find your own level of
tolerance.

--
Toby.
FT: 250 DK Bongo Stars
Anonymous
January 10, 2005 7:50:55 PM

Archived from groups: rec.audio.tech (More info?)

In article <20050107172710.23235.00002985@mb-m03.aol.com>, bandhphoto@aol.com
says...

>I know 128kbps is "CD quality" when converting CD music to MP3.

If you believe that I have a bridge to sell you.

>If I use a
>higher value, will I get substantially better sounding MP3 files?

Define substantially? It is noticeable if you listen carefully.

>If so, what's the point of diminishing returns -- the point where the increase
>in file size outweighs the quality imporvement?

I usually look at it the other way around. How bad can the the music be
before you can't stand to listen to it. I usually can't listen to anything
less than 128kbps. Even that I can only use as background music.
-------------
Alex
Anonymous
January 10, 2005 7:59:42 PM

Archived from groups: rec.audio.tech (More info?)

In article <slrncu57e0.csb.google@ID-171443.user.uni-berlin.de>,
google@asktoby.com says...
>
>
>On 2005-01-07, BandHPhoto <bandhphoto@aol.com> wrote:
>> I know 128kbps is "CD quality" when converting CD music to MP3. If I use a
>> higher value, will I get substantially better sounding MP3 files?
>>
>> If so, what's the point of diminishing returns -- the point where the
increase
>> in file size outweighs the quality imporvement?
>>
>> TIA.
>> - Henry
>
>128mbps is not CD quality to my ears.
>I suggest you carry out your own tests and find your own level of
>tolerance.

Best advice I have read, so far.
-----------------
Alex
Anonymous
January 10, 2005 8:40:15 PM

Archived from groups: rec.audio.tech (More info?)

On Mon, 10 Jan 2005 05:41:15 -0800, "Richard Crowley"
<rcrowley7@xprt.net> wrote:

>> No. There are several MP3 codecs. They output different
>> code. Some say that they sound different.
>
>Exactly what I said. Perhaps you have an alternate definition
>of the word "no".

The thread, as I read it went:

Nate: Isn't the compression algorithm you are using also significant?

RIchard: Huh? MP3 is MP3. WMA is WMA, and AAC is AAC.

Laurence: No. There are several MP3 codecs. They output different
code. Some say that they sound different.
Anonymous
January 12, 2005 3:05:25 AM

Archived from groups: rec.audio.tech (More info?)

BandHPhoto wrote:

> I know 128kbps is "CD quality" when converting CD music to MP3. If I
> use a higher value, will I get substantially better sounding MP3
files?
>
> If so, what's the point of diminishing returns -- the point where the
> increase in file size outweighs the quality imporvement?

Well, I and a colleague of mine who is a semipro musician each
independently arrived at the same conclusion: 192 kbps is the point
where it starts sounding pretty much the same as the original and most
listeners won't be able to tell any difference. 128 doesn't cut it.
Which is unfortunate because 128 is the rate most commonly used.
Whatever, 128 is not TOO bad, generally speaking. But if there's a
magic optimal point in the curve, I'd say it's 192.
Anonymous
January 12, 2005 2:07:34 PM

Archived from groups: rec.audio.tech (More info?)

paul-NOZPAM@paulkienitz.net writes:

> BandHPhoto wrote:
>
>> I know 128kbps is "CD quality" when converting CD music to MP3. If I
>> use a higher value, will I get substantially better sounding MP3
> files?
>>
>> If so, what's the point of diminishing returns -- the point where the
>> increase in file size outweighs the quality imporvement?
>
> Well, I and a colleague of mine who is a semipro musician each
> independently arrived at the same conclusion: 192 kbps is the point
> where it starts sounding pretty much the same as the original and most
> listeners won't be able to tell any difference. 128 doesn't cut it.
> Which is unfortunate because 128 is the rate most commonly used.
> Whatever, 128 is not TOO bad, generally speaking. But if there's a
> magic optimal point in the curve, I'd say it's 192.

I have never been able to distinguish 128 kpbs from an uncompressed
44.1 kHz wav file as tested by Arnie's ABX. But then my hearing
isn't what it used to be.
--
% Randy Yates % "Rollin' and riding and slippin' and
%% Fuquay-Varina, NC % sliding, it's magic."
%%% 919-577-9882 %
%%%% <yates@ieee.org> % 'Living' Thing', *A New World Record*, ELO
http://home.earthlink.net/~yatescr
Anonymous
January 12, 2005 4:12:56 PM

Archived from groups: rec.audio.tech (More info?)

Randy Yates wrote:
> I have never been able to distinguish 128 kpbs from an uncompressed
> 44.1 kHz wav file as tested by Arnie's ABX. But then my hearing
> isn't what it used to be.

Well, the bitrate is only one parameter in the codec, isn't
it? There is also *channel* and *quality* setting in my MP3
encoding software. So if choose *joint stereo* instead of
*stereo* the quality of the compression is impaired and
audible with earphones. For example, clear left-right
distinction is lost. However, I haven't tested it with
Arnie's ABX.

Cheers,

Franco
!