ntom;462237 Wrote:
Interesting thread
Have you reached any further conclusion Phil?
I'd be interested in knowing whether you are now putting everything
through DeClip?
Actually...no. What happened was it started me on a quest to hunt down
down better masterings of my favourite
Phil Leigh;462635 Wrote:
I think the software is clever and it works. But it doesn't turn
rubbish into gold.
Good point. I guess it also means that applying such a process to
Squeezebox could be good, with improvements ... but don't expect a
miracle.
--
cfuttrup
Interesting thread
Have you reached any further conclusion Phil?
I'd be interested in knowing whether you are now putting everything
through DeClip?
--
ntom
SB3,Perpetual Technologies P1-A/P3-A,Naim82 + 4xNaim135, Martin Logan
SL3
This is a very interested thread - I read it all. What happened to Slim
Devices considering adding a plug-in? I cross my fingers it will be
added to SqueezeCenter some day.
On my wishlist I have:
1. OK if the plugin ignores files that are not FLAC
2. Let it be an on-the-fly (streamed) process.
Phil Leigh;421459 Wrote:
yeah - I get all of that. But replaygain can't possibly know what the
gain of my amplifier is or what the efficiency of my speakers are so it
can't actually know what the absolute replay SPL is going to be...
The thing I don't get is how it could conclude that it is
tot;381284 Wrote:
I have been using DeClip Pro since last summer. I use it on Linux to
batch process clipped recordings, which means pretty much everything
released last ten+ years.
Even though the restored waveforms look impressive, I have tried to ABX
the result and I can't tell the
Ok. So following the advice from CliveB and Radish, I selected my test
track:
My God by Jethro Tull from Aqualung. Track peak is 1.0 but the album as
a whole needs a positive replaygain value - always a good sign for me
:-)
This track appears to be well mastered, sounds good and has a very good
cliveb;421287 Wrote:
Quite so - plenty of (unclipped) classical CDs have an album peak of 1.
Peak=1 plus a high replaygain value is probably a warning sign as you
say, but it's not quite so simple. There are plenty of modern pop/rock
albums that are clipped to buggery but have a peak
Been looking at my Replay gain distributions...
2.5k out of 30k tracks have a peak of 1.0
The most badly compressed album I have is Stop The Clocks by Oasis
(hmmm...) -12.5dB of correction!
If my maths is right (?) then this is a bit baffling since the max 16
bit dynamic range/snr is 98.08dB
cliveb;421451 Wrote:
It doesn't work quite like that.
First of all, the Replaygain algorithm doesn't use peak levels. It
determines the perceived loudness based on RMS levels and the ear's
varying sensitivity at different frequencies (using the Fletcher-Munson
curves, I believe).
cliveb;421451 Wrote:
Secondly, the 89dB reference level is derived from an SMPTE standard
which states that pink noise at -20dB RMS should be played at 83dB SPL.
(Replaygain started out using 83dB as its target loudness, but that's
been adjusted up to 89dB for reasons that are not entirely
Phil Leigh;420829 Wrote:
can you give us the names of some of your albums that are made
listenable?
Red hot Chilli Peppers: Stadium Arcadium
AC-DC: Black Ice
Daughtry: Daugtry
Runrig: Proterra, Everything you see, In search of angels
Tom Petty: Highway Companion
Eric Clapton: Me and Mr.
callesoroe;421041 Wrote:
Red hot Chilli Peppers: Stadium Arcadium
AC-DC: Black Ice
Daughtry: Daugtry
Runrig: Proterra, Everything you see, In search of angels
Tom Petty: Highway Companion
Eric Clapton: Me and Mr. Johnson, Back home
+ a lot of compilation albums
just to mention a few...
Phil Leigh;421074 Wrote:
I foresee a horrendous workflow... find all albums with clipping (ie
large replaygain adjustments) [...]
A large replaygain value does not necessarily imply there is clipping.
It is possible to hypercompress without actually clipping the signal.
And conversely, a low
Phil Leigh;421074 Wrote:
a-ha ... I have the EC one...
This is definitely clipped - album replaygain of -8.3 (that's hot!).
I tried the demo version of the software... obviously the gain change
is quite dramatic, so I stuck the before and after versions into Audio
DiffMaker - fascinating.
cliveb;421105 Wrote:
A large replaygain value does not necessarily imply there is clipping.
It is possible to hypercompress without actually clipping the signal.
And conversely, a low replaygain value does not necessarily mean there
is no clipping.
Yes Clive - I just figured that out by
callesoroe;421114 Wrote:
I have written to the devolpers, and they in fact do consider writing a
version, that works with squeezecenter. It's is already possible to run
a streaming version with I-tunes. I have not tried that because I am not
using I-tunes. So if more people write to them,
callesoroe;421114 Wrote:
I have written to the devolpers, and they in fact do consider writing a
version, that works with squeezecenter. It's is already possible to run
a streaming version with I-tunes. I have not tried that because I am not
using I-tunes. So if more people write to them,
Phil Leigh;421115 Wrote:
I think it is safe to say that tracks with Peak of 1.000 are
clipped... (20% of my tracks!)
I'm not sure that follows - I'm no audio engineer but when I'm
mastering tracks I always set it to normalize so the peak is 1, but
obviously with no clipping. In fact, I
radish;421134 Wrote:
I'm not sure that follows - I'm no audio engineer but when I'm mastering
tracks I always set it to normalize so the peak is 1, but obviously with
no clipping. In fact, I can't think why you'd ever master a track with
the peak 1 unless it's a quiet track on an album and
That makes perfect sense, thanks Phil :)
--
radish
radish's Profile: http://forums.slimdevices.com/member.php?userid=77
View this thread: http://forums.slimdevices.com/showthread.php?t=57872
ntom;418890 Wrote:
Anyone got any further trying this software out?
Having read the blurb on their site, I think reference to compression
in discussions so far are not pertinent.the software is intended to
deal with clipping which is a different issue and not compression which
we
callesoroe;420816 Wrote:
I am using the software, and it's worth every penny. I have the pro
version where there is a rounding feautere. I first declip the file and
then use the rounding function to rebuild the life, that can be lost
while lowering the level.
It works very well and many
Phil Leigh;381808 Wrote:
Even with the rules, you can't go back! You would have to remix from the
raw multitrack.
If you stick a signal through a compressor followed by an expander, you
don't get the original signal out! Analogue or digital Compression is a
one-way street. It's EXACTLY
ntom;418890 Wrote:
As stated there is no way to reverse compression such that you get back
to the original signal.
If all you're interested in is reversing the effects of the final
smash-the-levels compression that was applied during mastering (and
will accept the compression applied during
cliveb;419126 Wrote:
Formally there is no way to know for sure that something is clipped. If
you happen to see a flat top waveform then you can deduce with fairly
high confidence that it *is* the result of clipping, but you can't know
for sure the original signal wasn't intended to be
Phil Leigh;419179 Wrote:
other than I know of no way create to a consecutive set of full range
values other than to clip (either in software or by overloading an ADC)
Plenty of synths use basic oscillators like square waves to generate
tones. A simple square wave at full range will have flat
Phil Leigh;419179 Wrote:
Clive - I don't want to be too picky, but a consecutive sequence of 16
1's contain as much musical information as a consecutive sequence of
16 0's - i.e. no information at all. In fact, one could argue a
consecutive sequence of any fixed value has no musical
Anyone got any further trying this software out?
Having read the blurb on their site, I think reference to compression
in discussions so far are not pertinent.the software is intended to
deal with clipping which is a different issue and not compression which
we are stuck with.
As stated
Having thought about this a bit more and chatted with my DSP buddies, I
can see how this might be approached for brief (handful of
samples)periods of flat top clipping.
--
Phil Leigh
You want to see the signal path BEFORE it gets onto a CD/vinyl...it
ain't what you'd call minimal...
SB3
seanadams;381271 Wrote:
There is no way to undo compression.
This doesn't make sense to me. As long as there is at least a small
variation in dynamics why can't a dynamic range expander work to expand
the dynamic range, which in effect is undoing some of the compression?
DBX range expanders
Rick B.;381632 Wrote:
This doesn't make sense to me. As long as there is at least a small
variation in dynamics why can't a dynamic range expander work to expand
the dynamic range, which in effect is undoing some of the compression?
DBX range expanders were around even 30 years ago or
Rick B.;381632 Wrote:
This doesn't make sense to me. As long as there is at least a small
variation in dynamics why can't a dynamic range expander work to expand
the dynamic range, which in effect is undoing some of the compression?
DBX range expanders were around even 30 years ago or
Thank you, Dave, for your excellent reply. I see what you mean.
--
Rick B.
Rick B.'s Profile: http://forums.slimdevices.com/member.php?userid=8973
View this thread: http://forums.slimdevices.com/showthread.php?t=57872
Rick B.;381632 Wrote:
This doesn't make sense to me. As long as there is at least a small
variation in dynamics why can't a dynamic range expander work to expand
the dynamic range, which in effect is undoing some of the compression?
DBX range expanders were around even 30 years ago or
DaveWr;381685 Wrote:
DBX tried to use a linear compression to improve performance of tape
recordings in general. So they mapped say 75dB of incoming signal
range to 50dB of outgoing. Then applied inverse approach on playback.
There were still the problems of how fast to react to big
Hi,
two questions, the first of which I think has the potential of creating
some debate:)
1. Does anyone have experience with the SeeDeClip-software?
http://www.cutestudio.net/data/products/audio/seedeclip/index.php
They claim to be able to sometimes recover some of the information lost
by
Sounds cool, but using the Red Hot Chili Peppers -Californication- album
as an example - it's cited prominently on the software's home page as a
user comment - the clipped samples are -gone-.
Not pining for the fjords.
Ceased to be.
An ex-parrot, err, sample.
So how can they be recovered
Mark Lanctot;381127 Wrote:
The samples would have to be created somehow. I saw something about
FFT, and that's very nice, but they are still
guessed/interpolated/created from no real data. That data is just not
there anymore - it may hopefully exist on the masters but it sure
doesn't on
Definitely sounds cool, please keep us informed of your results!
--
Mark Lanctot
Current: SB2, Transporter, Boom (PQP3 - late beta)
Stored: Boom (PQP1 - early beta), SBC (beta - no battery)
Sold: SB3, Duet
Mark Lanctot's
Compression is NOT the same as clipping! In fact, the whole point of
compression is to maximize loudness without clipping.
You can't undo clipping, but you could mitigate the harsh sound of
digital clipping after the fact using DSP.
But this isn't going to fix a Red Hot Chili Peppers album.
I have been using DeClip Pro since last summer. I use it on Linux to
batch process clipped recordings, which means pretty much everything
released last ten+ years.
Even though the restored waveforms look impressive, I have tried to ABX
the result and I can't tell the difference on short tests.
tot;381284 Wrote:
I have been using DeClip Pro since last summer. I use it on Linux to
batch process clipped recordings, which means pretty much everything
released last ten+ years.
Cool, I then have a first sanity-check type question. Does the DeClip
software leave non-clipping parts of
bhaagensen;381294 Wrote:
Cool, I then have a first sanity-check type question. Does the DeClip
software leave non-clipping parts of the music 100% unchanged? Or does
it do its thing with parts that are e.g. just heavily compressed?
(Yes I will try as soon as time allows)
Bjørn
I'm
Phil Leigh;381299 Wrote:
By the way, heuristic algorithm is code for a guess, based on some
rules we invented
Yes, but that doesn't mean it doesn't work. Lots of algorithms utilise
heuristics with great results. It all depends on how well one can
structure the domain specific information
bhaagensen;381303 Wrote:
Yes, but that doesn't mean it doesn't work. Lots of algorithms utilise
heuristics with great results. It all depends on how well one can
structure the domain specific information needed to define one. As I
have no clue as to how one might guess what went on in the
I imagine uncompressing things (not many things are actually clipped)
would be very difficult so I'm not holding my breath.
Some processes aren't reversible. An example, I used to work at IBM
Hursley and they have massive, labyrinthine offices. On every corner
there are signs to direct you to
bhaagensen;381294 Wrote:
Cool, I then have a first sanity-check type question. Does the DeClip
software leave non-clipping parts of the music 100% unchanged? Or does
it do its thing with parts that are e.g. just heavily compressed?
It just drops the level of to have the room for restored
Phil Leigh;381299 Wrote:
I'm going to try it at the weekend but almost certainly it will have to
lower the overall level of the track in order to buy some headroom to
try and recover the clipped peaks.
By default it lowers the overall level to have room for the maximum
peak, but there are
49 matches
Mail list logo