Audio WAV rip engineer in the house?
B. Peg
I've been experimenting with most all rippers and have found something
interesting. Sometimes when the CD extraction to a WAV file is done (with
CDex and EAC), there is noticeable clipping as observed in SoundForge 7.0
and confirmed listening closely on some "good" phones (forget speakers
here). If I alter and save the WAV at a lower volume to where it doesn't
show clipping, then the new waveform shows lower volume but with "flat-top"
portions still in same place as it is in the original. Clipping (with
audible distortion), although lower level, is still there. Puzzling?
If a rip is done into a Mp3 file, the flat-topped clips remain and will be
noticeable as distortion. If the WAV file is poor, so will anything else
that comes from it. Add to that, if the WAV file is at 100% volume, then
the new rip from that always seems to gain volume (with Lame v3.96) and
SoundForge will show the new rip clipped beyond belief. Even shows up in
Mp3Gain as being clipped and lowering Mp3Gain to around 93 will alleviate
the volume clips, but the distortion clip is still present.
Lowering the percentage of the original WAV rip will also lower the peaks
overall so the flat-top's remain. It's almost like you need to draw in a
smooth curve
with SoundForge's pencil tool to correct the primary WAV rip which can get
rather tedious.
Haven't got a hold onto as why this is happening but it may have something
to do with the original CD being extracted at 100% which might mean it was
originally mastered or digitally "expanded" to a point where the master is
clipped too.
I am really getting sick of drawing in waveforms. Anyone got an easier way
of smoothing out a clipping master's WAV?
Hopefully, I explained this well. Tia.
B~
Particle Salad
first hand. Methinks that's just the way the CD was mastered because the
record company wanted it that way.
Ben Bradley
>(excuse me but I posted this elsewhere but need to know....)
>
>I've been experimenting with most all rippers and have found something
>interesting. Sometimes when the CD extraction to a WAV file is done (with
>CDex and EAC), there is noticeable clipping as observed in SoundForge 7.0
>and confirmed listening closely on some "good" phones (forget speakers
>here). If I alter and save the WAV at a lower volume to where it doesn't
>show clipping, then the new waveform shows lower volume but with "flat-top"
>portions still in same place as it is in the original. Clipping (with
>audible distortion), although lower level, is still there. Puzzling?
Do the waveforms look anything like those in this article?
[dropping a bombshell]
http://www.prorec.com/prorec/articles.nsf/articles/8A133F52D0FD71AB86256C2E005DAF1C
>...
>Haven't got a hold onto as why this is happening but it may have something
>to do with the original CD being extracted at 100% which might mean it was
>originally mastered or digitally "expanded" to a point where the master is
>clipped too.
Say it ain't so...
If you want to learn more (if the artice above didn't spell out
everything), google for hypercompression.
>I am really getting sick of drawing in waveforms. Anyone got an easier way
>of smoothing out a clipping master's WAV?
Have you looked into buying CD's that are over ten years old (I
mean PHYSICALLY over 10 years old, and not a recent "remastered"
release of an older recording)? That's about the only thing I can
think of that can help.
>Hopefully, I explained this well. Tia.
>
>B~
>
Arny Krueger
news:GxRsd.1039051$Gx4.3...@bgtnsc04-news.ops.worldnet.att.net
> (excuse me but I posted this elsewhere but need to know....)
>
> I've been experimenting with most all rippers and have found something
> interesting. Sometimes when the CD extraction to a WAV file is done
> (with CDex and EAC), there is noticeable clipping as observed in
> SoundForge 7.0 and confirmed listening closely on some "good" phones
> (forget speakers here). If I alter and save the WAV at a lower
> volume to where it doesn't show clipping, then the new waveform shows
> lower volume but with "flat-top" portions still in same place as it
> is in the original. Clipping (with audible distortion), although
> lower level, is still there. Puzzling?
No puzzle. I've found a number of commercial CD that were clipped. The
Prorec article cited by another writer is, as that writer suggests, the tip
of the iceburg.
http://www.prorec.com/prorec/articles.nsf/articles/8A133F52D0FD71AB86256C2E005DAF1C
is not all that new, and its not all that unique. But it is highly relevant.
>I am really getting sick of drawing in waveforms. Anyone got an easier way
>of smoothing out a clipping master's WAV?
If you just want to smooth it, low pass filter it. If you want to draw in
the missing parts of clipped waves, products like Adobe audition have
features that essentially do this.
G Glass
"Particle Salad" <ma...@particlesalad.com> wrote in message
news:IhSsd.39123$6q2....@newssvr14.news.prodigy.com...
http://news.harmony-central.com/Newp/2002/DeClipper-20.html
http://www.silksound.com/geniesys/
http://www.cube-tec.com/DeClipper.html
B. Peg
whole bunch of kick-drum transient wave corrections. Looks like distortion
doesn't matter anymore at the expense of LOUDNESS.
I'm going to go back and look at some early day CD's of The Corrs and see
what happened to their waveforms since their beginnings and current CD
"Borrowed Heaven" which also was an engineer's nightmare. I wonder to what
degree a poorly mastered CD has on sales overall?
Someone mentioned that Adobe Audition does some unclipping in it's program
where SoundForge doesn't. Dunno...
B~
G Glass
"B. Peg" <bent...@att.nett> wrote in message
was maximized to 0 dB, rather than 0.1 dB. With this in mind, you may be
hearing distortion not from what the meters are telling you, but rather
distortion caused by one of many factors during the mixing/mastering
process. I hear distortion every so often (bad distortion) from CDs in my
collection that don't show any clipping.
Also, the waveform may appear to have a square shape at the peaks only until
you zoom way in to have a closer look. After doing so, you may find that
the peaks look more normal.
B. Peg
http://www.silksound.com/geniesys/
I'll give it a try tomorrow and see what happens.
B~
Jim Gilliland
CD. It's a sad but true fact that many CDs are mastered this way.
There's not a lot that you can do to repair this. Adobe Audition (or
CoolEdit) has a tool for restoring clipped peaks which I've found
useful, but it's far from a perfect solution. It's under Effects, Noise
Reduction, Clip Restoration on the menu.
John
It would probably save a lot of time. Yes, it's quick to rip them, but it
seems like you are doing an excessive amount of editing afterwards. Just a
thought.
John
"B. Peg" <bent...@att.nett> wrote in message
news:GxRsd.1039051$Gx4.3...@bgtnsc04-news.ops.worldnet.att.net...
Arny Krueger
news:LrZsd.191223$Np3.7...@ursa-nb00s0.nbnet.nb.ca
> You know, maybe it would make sense to just sample the CD's in real
> time. It would probably save a lot of time. Yes, it's quick to rip
> them, but it seems like you are doing an excessive amount of editing
> afterwards. Just a thought.
Been there done that to circumvent some copy protection schemes.. Doesn't
help. If you do a clean job of recovering what started out as dirty audio,
you don't get clean audio, you get a very exact copy of the dirty audio.
The problem is that some newer CDs were mastered with clipping. If you
follow these things, managing hypercompression and clipping as EFX has
become a bit of an art. If you do your homework and arrange and play and
track and mix with the intent of hypercompressing and clipping, well the
results might be better than what happens if you don't do your homework.
Maybe a little...
Scott Dorsey
news:GxRsd.1039051$Gx4.3...@bgtnsc04-news.ops.worldnet.att.net
>
> I've been experimenting with most all rippers and have found something
> interesting. Sometimes when the CD extraction to a WAV file is done
> (with CDex and EAC), there is noticeable clipping as observed in
> SoundForge 7.0 and confirmed listening closely on some "good" phones
> (forget speakers here). If I alter and save the WAV at a lower
> volume to where it doesn't show clipping, then the new waveform shows
> lower volume but with "flat-top" portions still in same place as it
> is in the original. Clipping (with audible distortion), although
> lower level, is still there. Puzzling?
Yes, it's clipped. The original is clipped, the dubs will be clipped.
Clipping is the latest thing. Everybody's doing it. Customers come in
to me with horribly distorted major-label CDs and tell me that they need
to be just as loud as those guys. The current obsession with loudness at
all costs has resulted in flattopping being the normal state of affairs.
What is really horrible is to listen to some of the classic rock stuff from
the seventies out there... comparing the LP with an early CD issue and then
with the current CD issue makes it clear that something is really terribly
wrong with the new CDs. Not only distortion on peaks, but a total loss of
dynamics too. It's the NEW WAY.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
John
John
"Arny Krueger" <ar...@hotpop.com> wrote in message
news:RZWdnQ45kp6...@comcast.com...
DeserTBoB
>Yes, it's clipped. The original is clipped, the dubs will be clipped.
>Clipping is the latest thing. Everybody's doing it. Customers come in
>to me with horribly distorted major-label CDs and tell me that they need
>to be just as loud as those guys. The current obsession with loudness at
>all costs has resulted in flattopping being the normal state of affairs. <snip>
Ah, history repeats itself, yet again!
The craze in the '70s was horrid amounts of compression on just about
EVERYTHING. Looking (listening, actually) back, it was still more
tolerable than clipped digital, which completely obliterates the sound
on transients.
>What is really horrible is to listen to some of the classic rock stuff from
>the seventies out there... comparing the LP with an early CD issue and then
>with the current CD issue makes it clear that something is really terribly
>wrong with the new CDs. Not only distortion on peaks, but a total loss of
>dynamics too. It's the NEW WAY. <snip>
It SUCKS, and appeals to the lowest common denominator of
intelligence. The re-re-releases coming out now sound even worse than
the 45 singles of the '70s, and that's BAD. Even the LPs sound FAR
better, as do the original CD releases! Watch "classic rock" LPs zoom
in value soon, as well as cutout CDs from the '80s. This new stuff is
garbage, plain and simple.
dB
JD
change it. It just makes the track quieter but the distortion remains.
But I know what you mean. I've seen some wave forms that looked like solid
bars top to bottom in SF when I've been putting together mixed CD's. If I
remember right Orgy's Blue Monday is really bad.
JD
Logan Shaw
> Do the waveforms look anything like those in this article?
> [dropping a bombshell]
>
> http://www.prorec.com/prorec/articles.nsf/articles/8A133F52D0FD71AB86256C2E005DAF1C
Interesting article, and it makes some very good points. However,
I do have one nit to pick: in the example of really bad clipping,
the guy said that 110 samples were clipped (i.e. were at the
absolute maximum level). But that is on a 10-second audio clip.
So only 11 clipped samples on average per second. Out of 44100
samples per second. Even if all 11 samples occur continuously
(which is possible since the article talks about kick drum, and
there might be about one kick drum beat every second), the damage
is not that egregious.
It's probably possible, but not super easy, to hear 11 clipped
samples per second. It's just not that horrible. HOWEVER, I'm
not saying that the music as a whole hasn't been screwed up.
But it seems to me that 95% of the problem is the hard limiting
that occurs before the clipping. If you look at that those
waveforms, you will see that there are giant continuous sections
where they aren't quite at the limit, but they hug the line
like Han Solo piloting the Millenium Falcon through the rebuilt
Death Star in Return of the Jedi. That's where the real damage
is being done, in that little area before clipping that the waveform
has been smashed into.
So my nit is that clipping is not the real problem. If the
hard limiter that was used had been set just slightly differently,
there might have been zero clipped samples instead of 110 of them.
But the damage to the music would have been almost as great.
(Or perhaps even greater, since even more compression would have
been required to get the levels that hot without actually clipping.)
However, I will say that clipping is a *sign* of the problem
with ridiculous compression. But it is not *the* problem.
- Logan
Geoff Wood
"B. Peg" <bent...@att.nett> wrote in message
> (excuse me but I posted this elsewhere but need to know....)
>
> I've been experimenting with most all rippers and have found something
> interesting. Sometimes when the CD extraction to a WAV file is done (with
> CDex and EAC), there is noticeable clipping as observed in SoundForge 7.0
> and confirmed listening closely on some "good" phones (forget speakers
> here).
I really can't see how this could be happening. What do you get when you
rip sttraight into SF ?
geoff
B. Peg
Just tried that too. Nope. Predominate clipping still occurs. All sorts
of flat tops so it's definitely in the CD.
The DeClipper 2.5 above is interesting though. It offers several outcomes
but looks like the best bet is "manually" auditing all the clips in Audition
and maybe penciling them in SoundForge. What a pita though.
B~
Ben Bradley
I agree with yout nit, and this is a result of defining clipping
with the usual digital "maxed-out ADC" definition, as consecutive
samples at the upper or lower limit. Under 'normal' circumstances (an
otherwise-unprocessed track that got "normalized" but with the level a
little too hot) that would be a fine definition.
In "limiting cases" (pun intended) such as this, it would be useful
to make up another definition of clipping: any sample point that is
within (coming up with a plausible figure) five percent of the rail
(or within five percent of the signal's peak) would be clipping. I
hypothesize that this definition would result in a measurement of MANY
TIMES more clipped samples per second (and NOT show significant
clipping in older, non-hypercompressed recordings), and would be a
much better metric of the damage to the recording.
> - Logan
Sean Conolly
news:gehbr0pabeme8qbtu...@4ax.com...
> In "limiting cases" (pun intended) such as this, it would be useful
> to make up another definition of clipping: any sample point that is
> within (coming up with a plausible figure) five percent of the rail
> (or within five percent of the signal's peak) would be clipping. I
> hypothesize that this definition would result in a measurement of MANY
> TIMES more clipped samples per second (and NOT show significant
> clipping in older, non-hypercompressed recordings), and would be a
> much better metric of the damage to the recording.
I think it's simpler to say that if the waveform out of a process looks like
a sawed-off version of the waveform going in, then the process has clipped
the waveform, regardless of the amplitude. Using a hard limiter to suck out
all of the transients is one thing (a bad thing also) but truncating the
waveform just sounds like crap. Having a processor that rounds off the
corners as it truncates just makes it a little less obvious to the ears.
Sean
M. MacDonald
Works pretty well (i.e. intuitivley) as it also has a "Clipping Correction"
function in it's array of four plugins.
Supposedly works well with most other editing programs too as it is DirectX
based. Never had too much luck with DeClipper 2.5 as it's manual (Help) is
not too comprehensive in its explanations of all it's sliders like
"Gate....duh?" Might work well if someone were to take the time to write a
better tutorial for it.
Mack