22 Feb
2007
22 Feb
'07
3:48 p.m.
I've got a sound card capable of 96 bits, ( M-Audio audiofile 24/96 )
but my sound editor, Audacity for Linux, only goes up to 32 bits.
Is there a sound editor for Linux that can do higher than 32 bits?
4:05 p.m.
On Thu, 22 Feb 2007 13:48:04 -0600
millward <millward(a)ms.umanitoba.ca> wrote:
I've got a sound card capable of 96 bits, (
M-Audio audiofile 24/96 )
but my sound editor, Audacity for Linux, only goes up to 32 bits.
Is there a sound editor for Linux that can do higher than 32 bits?
Maybe I'm missing something here, but what on earth does anyone need
such high resolution for?
ISTR The humble CD is recorded at 18bit. which is over 200,000:1 and
more than 100dB. The orchestral dynamic range is I believe quoted at
90dB. I accept some extra headroom is nice, and the calculations aren't
quite so simple, but even 32bit comes out at mind boggling 200dB.
--
Will Godfrey
www.musically.me.uk
4:10 p.m.
New subject: [linux-audio-user] Linux music editor, less than 16-bit ?
i'm thinking of picking up the new korg pocket DSD recorder. and im wondering - does
anything support editing of 1-bit audio recordings in the OSS world? i know theress 3
dozen wave editors, but theyre all pretty mediocre and can usually not even connect to
jack and play a flac without crashing. i guess this is what happens when wave editor
developers have NIH syndrome?
so is the best bet using one of the DSD editors from windows in WINE?
5:55 p.m.
New subject: [linux-audio-user] Linux music editor, less than 16-bit ?
Hallo,
carmen hat gesagt: // carmen wrote:
i'm thinking of picking up the new korg pocket DSD recorder. and im wondering - does
anything support editing of 1-bit audio recordings in the OSS world? i know theress 3
dozen wave editors, but theyre all pretty mediocre and can usually not even connect to
jack and play a flac without crashing. i guess this is what happens when wave editor
developers have NIH syndrome?
Hey, that's actually the first time I read about 1-bit recording and
at first I thought you were joking or had a typo. But you weren't and
hadn't.
Nevertheless I found some interesting papers linked from the Wikipedia
SACD page. One is this:
Audio Engineering Society Convention Paper 5395: Why 1-Bit Sigma-Delta
Conversion is Unsuitable for High-Quality Applications
http://sjeng.org/ftp/SACD.pdf
and it states:
Single-stage, 1-bit sigma-delta converters are in principle
imperfectible. We prove this fact. The reason, simply stated, is that,
when properly dithered, they are in constant overload. Prevention of
overload allows only partial dithering to be performed. The
consequence is that distortion, limit cycles, instability, and noise
modulation can never be totally avoided. We demonstrate these effects,
and using coherent averaging techniques, are able to display the
consequent profusion of nonlinear artefacts which are usually hidden
in the noise floor. Recording, editing, storage, or conversion systems
using single-stage, 1-bit sigma-delta modulators, are thus inimical to
audio of the highest quality. In contrast, multi-bit sigma-delta
converters, which output linear PCM code, are in principle infinitely
perfectible. (Here, multi-bit refers to at least two bits in the
converter.) They can be properly dithered so as to guarantee the
absence of all distortion, limit cycles, and noise modulation. The
audio industry is misguided if it adopts 1-bit sigma-delta conversion
as the basis for any high-quality processing, archiving, or
distribution format to replace multi-bit, linear PCM.
The other is this by famous James A. Moorer et al.:
Audio Engineering Society Convention Paper 4719 on the 1st DSD
commercial editor
http://www.sonicstudio.com/pdf/papers/DSDStereoEdit.pdf
which abstracts to:
The editing in native form of 2.8224 Mhz 1-bit audio (Direct-Stream
Digital) is problematic since the required computation rate to
achieve level control, pan, and crossfade in real time is beyond the
capabilities of general-purpose digital signal processing chips. A
system has been developed which incorporates a special- purpose chip
designed specifically for computations on 1-bit audio. This is
embedded into the framework of a general-purpose audio editing and
premastering system. It behaves, looks and feels exactly like a
stereo, PCM editing system with full flexibility of sample-accurate
cross-fades, level control, metering, and non- linear random-access
editing, except that internally it is operating directly on the 1-
bit signal. This paper describes the data flow and architecture of
the system.
Now I'm wiser than before.
Ciao
--
Frank Barknecht _ ______footils.org_ __goto10.org__
8:26 p.m.
New subject: [linux-audio-user] Linux music editor, less than 16-bit ?
carmen hat gesagt: // carmen wrote:
> i'm thinking of picking up the new korg pocket DSD recorder
I dunno, looks like it uses a lot of proprietary formats. I've made some
great recordings with the M-Audio MicroTrack 24/94 digital recorder.
http://www.m-audio.com/products/en_us/MicroTrack2496-main.html
It sounds great, and mounts like any USB storage device, making file
transfer trivial. It will record in MP3 format, but I've always used
WAV, usually 44.1K/16bit so I don't have to convert before burning a CD,
but have also done 96K/24bit. Both worked well, though I didn't think
there was enough difference to justify the extra space used. I got a
2GB microdrive for it, which will record 3h22m @ 16/44 and 1h33m at 24/96,
so there's plenty of time available. For shows I've recorded where I've
needed more time, I just fire up the thinkpad, download the files so
far, wipe the disk and begin again.
There's a review with more technical diagrams than you can shake a stick
at here: http://www.sonicstudios.com/mt2496rv.htm
--
======================================================================
Joe Hartley - UNIX/network Consultant - jh(a)brainiac.com
Without deviation from the norm, "progress" is not possible. - FZappa
26 Feb
26 Feb
2:43 p.m.
New subject: [linux-audio-user] Linux music editor, less than 16-bit ?
Frank Barknecht wrote:
Hey, that's actually the first time I read about 1-bit recording and
at first I thought you were joking or had a typo. But you weren't and
hadn't.
It works pretty well and is easy to implement in silicon.
Back in college, using a few square inches of printed circuit board, a
soldering iron and a handful of components, I made two little devices
whose only purpose in life was to:
1. Grab an analog input and transform it into a 1-bit digital signal
2. Grab the 1-bit digital signal and convert it back to analog
Not very useful in practice, but a fun project for a boring week-end
afternoon. The toys were so simple, they didn't even have a clock, the
transitions between 1 and 0 happened whenever the proper conditions were
met (there was a capacitor somewhere that was either charging or
discharging very quickly, with the average voltage trying to follow the
analog input, and it was sending to the digital output the information
of either "I'm charging" or "I'm discharging"). I guess the
digital
"carrier" was somewhat self-stabilizing (somewhere in the MHz range I
guess).
Those of you familiar with electronics can probably figure out the rough
design of the scheme already.
The sound quality was pretty damn good too, which is somewhat odd given
the rudimentary technology.
With minimal effort, more complex converters can be made that can offer
much better quality. Mine was a hack.
Single-stage, 1-bit sigma-delta converters are in
principle
imperfectible. We prove this fact. The reason, simply stated, is that,
when properly dithered, they are in constant overload. Prevention of
overload allows only partial dithering to be performed. The
consequence is that distortion, limit cycles, instability, and noise
modulation can never be totally avoided.
Interesting.
--
Florin Andrei
http://florin.myip.org/
22 Feb
22 Feb
4:25 p.m.
Folderol wrote:
On Thu, 22 Feb 2007 13:48:04 -0600
millward <millward(a)ms.umanitoba.ca> wrote:
I don't know about such things in the audio world, but in the world of
color photography, most professional digital equipment uses 48-bit
color. This is way outside the reproduction range of any photo
printing/display technology. But even though a particular 48-bit color
might not be printable or displayable, it is still there. It can be
taken into consideration when doing color adjustments and image filters.
The end result is that when color depth is reduced to the 24-bit color
range that JPG uses - you get better and more accurate color reductions.
So I would think that working in higher bit-depths for audio would
similarly result in better sounding audio when it's reduced to CD format.
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
I've got a sound card capable of 96 bits, (
M-Audio audiofile 24/96 )
but my sound editor, Audacity for Linux, only goes up to 32 bits.
Is there a sound editor for Linux that can do higher than 32 bits?
Maybe I'm missing something here, but what on earth does anyone need
such high resolution for?
ISTR The humble CD is recorded at 18bit. which is over 200,000:1 and
more than 100dB. The orchestral dynamic range is I believe quoted at
90dB. I accept some extra headroom is nice, and the calculations aren't
quite so simple, but even 32bit comes out at mind boggling 200dB. 
4:53 p.m.
david wrote:
Folderol wrote:
Yes, but this 48bit representation of color is just 16bit x3 colors
[channels]. In other words the 48bit representation is 3 unrelated
16bit [channel] representations concatenated, one each for the 3 primary
colors/CCD sensors. The equivilent for audio would be just 16bit as
there is only one channel.
On Thu, 22 Feb 2007 13:48:04 -0600
millward <millward(a)ms.umanitoba.ca> wrote:
I don't know about such things in the audio world, but in the world of
color photography, most professional digital equipment uses 48-bit color I've got a sound card capable of 96 bits, (
M-Audio audiofile 24/96 )
but my sound editor, Audacity for Linux, only goes up to 32 bits.
Is there a sound editor for Linux that can do higher than 32 bits?
Maybe I'm missing something here, but what on earth does anyone need
such high resolution for?
ISTR The humble CD is recorded at 18bit. which is over 200,000:1 and
more than 100dB. The orchestral dynamic range is I believe quoted at
90dB. I accept some extra headroom is nice, and the calculations aren't
quite so simple, but even 32bit comes out at mind boggling 200dB. This is way outside the reproduction range of any
photo
printing/display technology. But even though a particular 48-bit color
might not be printable or displayable, it is still there. It can be
taken into consideration when doing color adjustments and image
filters. The end result is that when color depth is reduced to the
24-bit color range that JPG uses - you get better and more accurate
color reductions.
Following from above, your final JPG color depth gets reduced to just
8bits per color which is why RGB uses values from 0-255. 8bits only has
informaiton for 256 values.
So I would think that working in higher bit-depths for
audio would
similarly result in better sounding audio when it's reduced to CD format.
This is true, but as Folderol wrote, 32 bits should be *plenty* of
dynamic range for audio. In fact, it has been argued that ~22bits is
sufficient as beyond this you get into the h/w noise floor, hearing
limits, etc.
(Any other experts out there, feel free to jump in if I'm wrong here...)
HTH,
Rick
5:06 p.m.
On Thu, 2007-02-22 at 15:53 -0500, Rick Wright wrote:
This is true, but as Folderol wrote, 32 bits should be
*plenty* of
dynamic range for audio. In fact, it has been argued that ~22bits is
sufficient as beyond this you get into the h/w noise floor, hearing
limits, etc.
32 bits of resolution gets you close to the *thermal* noise floor. i.e.
your recording contains noise contained by brownian motion. not very
useful.
32 bits of IEEE floating point gives you 24 bits of resolution and a
bunch of bits for an exponent to allow summing without clipping.
--p
5:56 p.m.
On Thu, Feb 22, 2007 at 04:04:34PM -0500, Paul Davis wrote:
On Thu, 2007-02-22 at 15:53 -0500, Rick Wright wrote:
Unfortunately digital processing has completely different set of
problems than analogue.
32 bit seems to be plenty for us mere mortals with more practical
problems, but I wouldnt dismiss the esoteric benefits of increased
resolution for complicated processing tasks so easily when quality
is the prime concern. The maths is very tricky.
Some nice products do use 64 bits (or other high resolution integer
formats), and I'm not so convinced that their nice sound is
completely unrelated to the sample size.
As an interesting, though slightly obsure example, this article claims
that 32bit float is not enough to make a dithered 24bit integer signal:
http://www.cadenzarecording.com/floatingdither.html
This old article also points out some mathematical problems:
http://www.jamminpower.com/PDF/48-bit%20Audio.htm
Perhaps someone more versed in dsp maths can point out the flaws
in these articles, and other hi-end audio misconceptions?
The primary use for 64 bit is the critical mastering stage, where it
arguably doesnt make sense not to make use of available processing power.
--
Tim Orford
This is true, but as Folderol wrote, 32 bits
should be *plenty* of
dynamic range for audio. In fact, it has been argued that ~22bits is
sufficient as beyond this you get into the h/w noise floor, hearing
limits, etc.
32 bits of resolution gets you close to the *thermal* noise floor. i.e.
your recording contains noise contained by brownian motion. not very
useful.
8:32 p.m.
On Thu, 2007-02-22 at 21:45 +0000, Tim Orford wrote:
32 bit seems to be plenty for us mere mortals with
more practical
problems, but I wouldnt dismiss the esoteric benefits of increased
resolution for complicated processing tasks so easily when quality
is the prime concern. The maths is very tricky.
the issue is not complexity, its whether the operations produce results
outside of the range of a 24 bit integer. if they do not, then floating
point is 100% accurate, and greater resolution/bit depth is of no use.
if they do, then things get more interesting.
Some nice products do use 64 bits (or other high
resolution integer
formats), and I'm not so convinced that their nice sound is
completely unrelated to the sample size.
As an interesting, though slightly obsure example, this article claims
that 32bit float is not enough to make a dithered 24bit integer signal:
http://www.cadenzarecording.com/floatingdither.html
dither is a perennial subject on many pro-audio/DSP mailing lists.
having just survived the latest one on another list, i don't wish to
start another here. i will just note that Nika is smart, but his paper
skips over some basic issues in a way that causes his results to appear
more important than they actually are. adding two 24 bit numbers does
not create a 48 bit result. multiplying a 24 bit number by a relatively
small number does not create a 48 bit result. in fact, for the vast
majority of possble 24 bit sample values and the vast majority of gain
coefficients used in audio processing, the result is still a 24 bit
number. his results/conclusions are not incorrect, they just apply to
the majority of data flow through a typical 32 bit floating point
system.
This old article also points out some mathematical
problems:
http://www.jamminpower.com/PDF/48-bit%20Audio.htm
Perhaps someone more versed in dsp maths can point out the flaws
in these articles, and other hi-end audio misconceptions?
many of these articles seem, like Nika's, to assume that the full
representational range of the 32 bit floating point number will be in
use. it is certainly true that there are many inaccuracies when
representing computational results with floating point numbers *BUT*
these do not apply (or certainly do not apply in the same way) when the
range of values that you are representing remains within the bounds of
the 24 bit mantissa.
for most audio, most of the time, this condition is satisfied.
i will admit to a minor uncertainty in this when it comes to the
normalized form that is generally used: i.e. full scale (0 dBFS) samples
are represented not by 2^24-1 but by 1.0. i have a suspicion that this
still allows 100% accurate computation as long as the final answer is
less than 1.0, but on this i am slightly less sure than i would like.
The primary use for 64 bit is the critical mastering
stage, where it
arguably doesnt make sense not to make use of available processing power.
as i've explained repeatedly, the "64 bitness" of a processor is
irrelevant when discussing 32 bit floating point math except for an
increase in bus bandwidth, register width and a few other factors that
affect speed, not accuracy.
26 Feb
26 Feb
2:52 p.m.
Paul Davis wrote:
dither is a perennial subject on many pro-audio/DSP mailing lists.
having just survived the latest one on another list, i don't wish to
start another here.
Care to provide a link to a mail archive?
Thanks,
--
Florin Andrei
http://florin.myip.org/
27 Feb
27 Feb
10:12 a.m.
On 2/22/07, Tim Orford <orford(a)zonnet.nl> wrote:
On Thu, Feb 22, 2007 at 04:04:34PM -0500, Paul Davis
wrote:
Yes, in my experience 32-bit audio isn't necessary for listening so
much as it is for processing. Csound is available with 32- or 64-bit
calculations, and the usual recommendation is to use 32 for composing
and producing and making all decisions, and 64-bit for final
rendering, simply because, if you have a dozen or so filters all
making little round-off errors, sine wave tables that are accurate to
so many bits all being scaled and time-shifted, Fourier transforms
converting from time to frequency representations, and a thousand
other things being calculated and rounded off, having twice the bit
depth could change the final outcome, even if the eventual delivery
method will be 16 bits. I'm sure it's the same idea with graphics.
BTW, I thought 24-bit graphics were sometimes implemented as 32-bit,
with an alpha channel? What a brilliant idea.
I wouldn't make the comparison, though, that 48 bit graphics is like
16 bits with 3 channels, because I don't see red, green, and blue
separately. If you change the value in one channel, I don't say "hey
that looks bluer, but about the same red and green", I say "It's a
different color".
-Chuckk
--
http://www.badmuthahubbard.com
On Thu, 2007-02-22 at 15:53 -0500, Rick Wright
wrote:
Unfortunately digital processing has completely different set of
problems than analogue.
32 bit seems to be plenty for us mere mortals with more practical
problems, but I wouldnt dismiss the esoteric benefits of increased
resolution for complicated processing tasks so easily when quality
is the prime concern. The maths is very tricky. This is true, but as Folderol wrote, 32 bits
should be *plenty* of
dynamic range for audio. In fact, it has been argued that ~22bits is
sufficient as beyond this you get into the h/w noise floor, hearing
limits, etc.
32 bits of resolution gets you close to the *thermal* noise floor. i.e.
your recording contains noise contained by brownian motion. not very
useful.
22 Feb
22 Feb
6:25 p.m.
Rick Wright wrote:
That's right. But the more information you start with, the smarter the
reduction can be - particularly in areas where the 48-bit color value is
outside the 24-bit dynamic range. In photography, this means the
highlight and shadow areas. In the audio world, would mean this more
accurate renditions in the lowest and highest volume sounds? I don't know.
Depth reduction in color photography doesn't have to be a mindless
simple mathematical calculation, good software like Photoshop
(Windows/Mac) Bibble (Windows/Linux), Raw Therapee (Windows/Linux),
DCRaw (Linux) can do things more intelligently because they have more
information available to them when working with 48-bit color.
I think you can go even higher than 48-bit color in the professional
color world - I think high-end drum scanners scan at 64-bit color.
Yes, experts? I'm no audio engineer!
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
david wrote:
Hmmm, wouldn't there actually be TWO channels for audio - stereo?
Folderol wrote:
Yes, but this 48bit representation of color is just 16bit x3 colors
[channels]. In other words the 48bit representation is 3 unrelated
16bit [channel] representations concatenated, one each for the 3 primary
colors/CCD sensors. The equivilent for audio would be just 16bit as
there is only one channel. On Thu, 22 Feb 2007 13:48:04 -0600
millward <millward(a)ms.umanitoba.ca> wrote:
I don't know about such things in the audio world, but in the world of
color photography, most professional digital equipment uses 48-bit color I've got a sound card capable of 96 bits, (
M-Audio audiofile 24/96 )
but my sound editor, Audacity for Linux, only goes up to 32 bits.
Is there a sound editor for Linux that can do higher than 32 bits?
Maybe I'm missing something here, but what on earth does anyone need
such high resolution for?
ISTR The humble CD is recorded at 18bit. which is over 200,000:1 and
more than 100dB. The orchestral dynamic range is I believe quoted at
90dB. I accept some extra headroom is nice, and the calculations aren't
quite so simple, but even 32bit comes out at mind boggling 200dB. This is way
outside the reproduction range of any photo
printing/display technology. But even though a particular 48-bit color
might not be printable or displayable, it is still there. It can be
taken into consideration when doing color adjustments and image
filters. The end result is that when color depth is reduced to the
24-bit color range that JPG uses - you get better and more accurate
color reductions.
Following from above, your final JPG color depth gets reduced to just
8bits per color which is why RGB uses values from 0-255. 8bits only has
informaiton for 256 values. So I would
think that working in higher bit-depths for audio would
similarly result in better sounding audio when it's reduced to CD format.
This is true, but as Folderol wrote, 32 bits should be *plenty* of
dynamic range for audio. In fact, it has been argued that ~22bits is
sufficient as beyond this you get into the h/w noise floor, hearing
limits, etc.
(Any other experts out there, feel free to jump in if I'm wrong here...) 
6:48 p.m.
On Thu, Feb 22, 2007 at 12:24:47PM -1000, david wrote:
Rick Wright wrote:
No :-) Sure, two channels is still the most common delivery format
(eg. CD). But mono files will still be important for a long time
(eg. as individual regions in a multi-channel editor session), and for
delivery, the six-channel "5.1" format is becoming pretty common.
So in the audio world, by convention we talk about bit depth per
channel, not total. If the graphics people used our conventions,
they'd talk about a "16-bit, 3 channel" image, rather than saying "48
bit" and leaving the number of channels implicit. (Which is
weird. What about when you switch color spaces, eg. from RGB to CMYK?
Is it still 16 bits per channel? Do you call the CMYK file a "64-bit"
file or what?)
Conversely, if the audio world were to adopt the conventions of the
graphics world, we'd talk about CDs as "32-bit" audio. But we don't :)
--
Paul Winkler
http://www.slinkp.com
Yes, but this 48bit representation of color is
just 16bit x3 colors
[channels]. In other words the 48bit representation is 3 unrelated
16bit [channel] representations concatenated, one each for the 3 primary
colors/CCD sensors. The equivilent for audio would be just 16bit as
there is only one channel.
Hmmm, wouldn't there actually be TWO channels for audio - stereo?
23 Feb
23 Feb
2:37 a.m.
Paul Winkler wrote:
On Thu, Feb 22, 2007 at 12:24:47PM -1000, david
wrote:
Hmmm - I've only seen that in games and DVD movies.
Rick Wright wrote:
No :-) Sure, two channels is still the most common delivery format
(eg. CD). But mono files will still be important for a long time
(eg. as individual regions in a multi-channel editor session), and for
delivery, the six-channel "5.1" format is becoming pretty common. Yes, but this 48bit representation of color is
just 16bit x3 colors
[channels]. In other words the 48bit representation is 3 unrelated
16bit [channel] representations concatenated, one each for the 3 primary
colors/CCD sensors. The equivilent for audio would be just 16bit as
there is only one channel.
Hmmm, wouldn't there actually be TWO channels for
audio - stereo? So in the audio world, by convention we talk about bit
depth per
channel, not total. If the graphics people used our conventions,
they'd talk about a "16-bit, 3 channel" image, rather than saying "48
bit" and leaving the number of channels implicit. (Which is
weird. What about when you switch color spaces, eg. from RGB to CMYK?
Is it still 16 bits per channel? Do you call the CMYK file a "64-bit"
file or what?)
CMYK uses 8-bits per color, so it uses 32-bits of color information. But
its actual color space works out to less than the RGB color space that
JPG uses.
I don't know anything about the mathematics of converting from CMYK to
RGB. RGB and CYM (not including K for Black) are complimentary colors,
or something like that, and the conversion process has to turn part of
each pixel of RGB into an 8-bit shade of black. I just let the programs
do it ...
Conversely, if the audio world were to adopt the
conventions of the
graphics world, we'd talk about CDs as "32-bit" audio. But we don't :)
Just wait until some recording industry marketing droid somewhere gets
the bright idea of pitching "premium 32-bit CDs!" ;-)
--
David
gnome(a)hawaii.rr.com
authenticity, honesty, community
22 Feb
22 Feb
6:51 p.m.
Hmmm, wouldn't there actually be TWO channels for audio - stereo?
There are as many channels as you hear. They are just not represented
in the same way as images, there are in actuality, as you surmised, two
16 bit 44.1 kHz channels, so we would have a 32 bit setup there if you
did the same thing graphics do, we just choose to give the overall
resoltuion and the amount of channels as two seperate numbers, as each
channel will usually have the same bit depth.
Seablade
26 Feb
26 Feb
2:50 p.m.
Rick Wright wrote:
This is true, but as Folderol wrote, 32 bits should be *plenty* of
dynamic range for audio. In fact, it has been argued that ~22bits is
sufficient as beyond this you get into the h/w noise floor, hearing
limits, etc.
24 bit means 135dB signal/noise ratio.
I haven't seen yet analog gear capable of 135dB. Hints:
- 90dB is the difference between a jackhammer and silence
- the human hearing range is 120dB - above that permanent injury occurs
- the space shuttle taking off generates "noise" somewhere in the 130..140db
Note: yeah, I know, rigorously speaking I'm mixing here several
different scales that should not be mixed. Don't kick the messenger into
the pit, even if "this is Sparta". ;-)
(bonus points for figuring out the reference)
--
Florin Andrei
http://florin.myip.org/
3:56 p.m.
Vincent Jaubert wrote:
On 2/26/07, Florin Andrei
<florin(a)andrei.myip.org> wrote:
Right, keep an overhead while processing, while the requirements for
delivery are lower.
But I would argue that I'd like a tad more than 16bit for delivery.
20bit would give us around 115dB if I'm not mistaken. 22bit would be
somewhere in the 120's. It would be nice to cover the whole dynamic
range of the human ear, which is 120dB.
Come to think of it, it's probably easier to just make it 24bit (3x8)
and forget all the nitpicking. Let processing be done on 32bit and that
would be it.
--
Florin Andrei
http://florin.myip.org/
24 bit means 135dB signal/noise ratio.
I haven't seen yet analog gear capable of 135dB.
24 bits is useful for processing. For the end user, 16 bits is enough. 
27 Feb
27 Feb
5:43 a.m.
2007/2/26, Florin Andrei <florin(a)andrei.myip.org>rg>:
Vincent Jaubert wrote:
You have amps and speakers doing 120dB? Show them to me.
Most time when I try to reach that dynamic range (with my home stereo
or with a professional) PA I get several problems:
- Neighbors banging on the door because of the loud passages in the music.
- Listeners complaining about not hearing anything in a silent
passage after 2 minutes of 100dB.
- To small mainlines to get the current for powering 120dB(A).
- To small budget to be anywhere near 120dB dynamic range.
On 2/26/07, Florin Andrei
<florin(a)andrei.myip.org> wrote:
Right, keep an overhead while
processing, while the requirements for
delivery are lower.
But I would argue that I'd like a tad more than 16bit for delivery.
20bit would give us around 115dB if I'm not mistaken. 22bit would be
somewhere in the 120's. It would be nice to cover the whole dynamic
range of the human ear, which is 120dB. 24 bit means 135dB signal/noise ratio.
I haven't seen yet analog gear capable of 135dB.
24 bits is useful for
processing. For the end user, 16 bits is enough. Come to think of it, it's probably easier to just
make it 24bit (3x8)
and forget all the nitpicking. Let processing be done on 32bit and that
would be it.
While 24bit or higher is nice for processing, the only reason I see
for more bits on the consumer-side is to have a higher SN-ratio. But
most consumer electronic is worse than noise of 16bit dac...
Arnold
--
visit http://www.arnoldarts.de/
---
Wenn man mit Raubkopien Bands wie Brosis oder Britney Spears wirklich
verhindern könnte, würde ich mir noch heute einen Stapel Brenner und
einen Sack Rohlinge kaufen.
7:57 a.m.
tisdag 27 februari 2007 10:43 skrev Arnold Krille:
2007/2/26, Florin Andrei
<florin(a)andrei.myip.org>rg>:
There is a Swedish inventor that has constructed a sub that can replace 40 sub
membrane. It is build with turbine instead of membrane. I might be such a
beast.
Swedish ref.
http://www.nyteknik.se/uploaded/image/bilder/bildarkiv/RotoSubhogtalare.jpg
http://www.nyteknik.se/art/41212
Small ref. from CeBit (in English)
http://www.rotosub.com/f_CeBit.html
--
- Bengan ------------------------------------------------------------
- KTHNOC - No1 Internet Offspring | http://www.noc.kth.se/~bengan ---
- 08-7906586 (ej vidarekopplat), 070-6443325 ------------------------
Vincent Jaubert wrote:
You have amps and speakers doing 120dB? Show them to me. On 2/26/07, Florin Andrei
<florin(a)andrei.myip.org> wrote:
Right, keep an overhead while processing, while the requirements for
delivery are lower.
But I would argue that I'd like a tad more than 16bit for delivery.
20bit would give us around 115dB if I'm not mistaken. 22bit would be
somewhere in the 120's. It would be nice to cover the whole dynamic
range of the human ear, which is 120dB. 24 bit means 135dB signal/noise ratio.
I haven't seen yet analog gear capable of 135dB.
24 bits is useful for processing. For the end user, 16 bits is enough. 
3:07 p.m.
On Tue, Feb 27, 2007 at 10:43:39AM +0100, Arnold Krille wrote:
You have amps and speakers doing 120dB? Show them to
me.
Are you talking sound pressure level, or dynamic range?
It's really not hard to reach 120 dB SPL at close range. I own a bass
cabinet and an amplifier that I could, if I chose, drive to 130 dB SPL
at 1 meter, and be well within the system's limits.
The cabinet is rated at 103 dB at 1 watt / 1 meter, and has impedance
of 4 ohms, and maximum continuous power handling of 1000 watts. My
amplifier can deliver 800 watts at 4 ohms. You get roughly 3 dB per
doubling of wattage, so let's see: 1 watt << 9 = 512 watts is well
within the system's limits, and that puts the SPL at 103 + 9 * 3 = 130
dB at 1 meter. You could back away to 3 meters and still be at roughly
120.
Dynamic range, though... highly doubtful :-) I do not know what the
noise floor of my system is, but at those settings I'd guess it to be
at least 20 dB SPL with no input connected, and this is a very good
"quiet" system. With a bass guitar connected (humbucking pickups),
forget it... the noise floor skyrockets, I would guess it to be well
over 40 dB. These are just guesses. The manufacturer rates the
amplifier's noise at "-85 dB or better". A popular brand of
professional power amps rates their noise level at -107 dB, but that's
just a power amp... add a complete system and you'll get a lot more
noise (mine includes a preamp with equalizer).
Most time when I try to reach that dynamic range (with
my home
stereo or with a professional) PA I get several problems:
Your comments below again seem to mix up dynamic range with sound
pressure level.
- Neighbors banging on the door because of the loud
passages in the
music.
agreed :) but that's purely SPL.
- Listeners complaining about not hearing anything in
a silent
passage after 2 minutes of 100dB.
agreed :)
- To small mainlines to get the current for powering
120dB(A).
Again, that's just SPL. And it depends on the amplifier. Mine is a
class D system that draws about 1000 watts at maximum rated output, or
roughly 8 amps at 120 volts AC. A standard 20-amp household circuit
can easily handle that, unless it is already pretty heavily loaded.
- To small budget to be anywhere near 120dB dynamic
range.
agreed :) The noise floor is going to be high.
120 dB dynamic range is an incredible amount, it corresponds roughly
to the difference between whispering and a rifle fired at 1 meter.
Even the humble CD can capture the difference between whispering and a
chainsaw.
And please remember folks, even 85 dB SPL can be enough to damage your
ears with long-term exposure.
--
Paul Winkler
http://www.slinkp.com
5:41 p.m.
2007/2/27, Paul Winkler <pw_lists(a)slinkp.com>om>:
On Tue, Feb 27, 2007 at 10:43:39AM +0100, Arnold
Krille wrote:
Well, both :-)
Actually if it is dynamic range and an amp can do a dynamic range of
120dB or more but I turn it down so the loudest level is 60dB SPL at
my ear, the smallest level will be 60dB-120dB=... well, it will be
silent enough to be not to be heard. (*)
You have amps and speakers doing 120dB? Show them
to me.
Are you talking sound pressure level, or dynamic range? It's really not hard to reach 120 dB SPL at close
range. I own a bass
cabinet and an amplifier that I could, if I chose, drive to 130 dB SPL
at 1 meter, and be well within the system's limits.
We are talking about music-production for the mainstream and PA, not
musicians amps. :-)
I haven't seen anyone firing his home equipment to have 120dB and
sitting 1m or less away from it...
<snip>
120 dB dynamic range is an incredible amount, it
corresponds roughly
to the difference between whispering and a rifle fired at 1 meter.
Even the humble CD can capture the difference between whispering and a
chainsaw.
120dB is the difference between just loud enough to be heard by your
ear (1dB) and the loudness when the bones in your ear bow to protect
it. At higher levels (>120dB) your ear aches but it will recover if
the noise isn't permanent. At least thats the definition I learned...
And please remember folks, even 85 dB SPL can be
enough to damage your
ears with long-term exposure.
Agreed! :)
(*) At least if my physics aren't completely wrong...
--
visit http://www.arnoldarts.de/
---
Wenn man mit Raubkopien Bands wie Brosis oder Britney Spears wirklich
verhindern könnte, würde ich mir noch heute einen Stapel Brenner und
einen Sack Rohlinge kaufen.
6:39 p.m.
On Tue, Feb 27, 2007 at 10:40:46PM +0100, Arnold Krille wrote:
We are talking about music-production for the
mainstream and PA, not
musicians amps. :-)
But you said "or with a professional PA". Surely you know that
commonly installed public address systems are capable of far more than
my puny 800 watt bass amp with its two 12" speakers?
A typical 300 capacity club here in NY will have 5000 watts or so
driving a 3-way system with the midrange drivers less than a meter
from the closest audience members. 120 dB at 1 meter would be about
108 dB at 4 meters, still plenty dangerous...
I haven't seen anyone firing his home equipment to
have 120dB and
sitting 1m or less away from it...
You never said anything about distance :-)
and i see that every time I go out to a club.
<snip>
I got my numbers from http://en.wikipedia.org/wiki/Decibel#Safety but
obviously their numbers represent some sort of typical examples; I
been in a number of discotheques (here in the US we can't pronounce
that so we just say "clubs") that had sound levels far above 100 dB.
The amount of time the ear can be exposed to a given sound pressure
level before sustaining permanent damange clearly varies a lot from
person to person. But government standards are probably a good
guideline to start with, eg. OSHA table G-16:
http://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=STANDARDS&…
TABLE G-16 - PERMISSIBLE NOISE EXPOSURES (1)
______________________________________________________________
|
Duration per day, hours | Sound level dBA slow response
____________________________|_________________________________
|
8...........................| 90
6...........................| 92
4...........................| 95
3...........................| 97
2...........................| 100
1 1/2 ......................| 102
1...........................| 105
1/2 ........................| 110
1/4 or less................| 115
____________________________|________________________________
Footnote(1) When the daily noise exposure is composed of two or
more periods of noise exposure of different levels, their combined
effect should be considered, rather than the individual effect of
each. If the sum of the following fractions: C(1)/T(1) + C(2)/T(2)
C(n)/T(n) exceeds unity, then, the mixed exposure should be
considered to exceed the limit value. Cn indicates the total time of
exposure at a specified noise level, and Tn indicates the total time
of exposure permitted at that level. Exposure to impulsive or impact
noise should not exceed 140 dB peak sound pressure level.
--
Paul Winkler
http://www.slinkp.com
120 dB dynamic range is an incredible amount, it
corresponds roughly
to the difference between whispering and a rifle fired at 1 meter.
Even the humble CD can capture the difference between whispering and a
chainsaw.
120dB is the difference between just loud enough to be heard by your
ear (1dB) and the loudness when the bones in your ear bow to protect
it. At higher levels (>120dB) your ear aches but it will recover if
the noise isn't permanent. At least thats the definition I learned...
28 Feb
28 Feb
7:44 a.m.
Hi Paul and all (that rhymes :-),
2007/2/27, Paul Winkler <pw_lists(a)slinkp.com>om>:
On Tue, Feb 27, 2007 at 10:40:46PM +0100, Arnold
Krille wrote:
I stand corrected, my numbers where from my physics-book which states
that 120dB is the "Schmerzgrenze" and 0dB is the slightest whisper
human ear can hear. Problem is that these values vary on the frequency
so there is a correction curve standardised(?) in some A-something
paper so these corrected values for the human ear are called db(A)...
Otherwise these logarithmic power- and pressure-levels are bubbling my
mind and if I don't stop thinking about theme, I will get even more
screws loose in my head.
So, for the archives and the record and Paul, don't trust the words I
said, I don't really have a clue. At least not good enough to do an
argument.
Happy coding and musing(?),
Arnold
--
visit http://www.arnoldarts.de/
---
Wenn man mit Raubkopien Bands wie Brosis oder Britney Spears wirklich
verhindern könnte, würde ich mir noch heute einen Stapel Brenner und
einen Sack Rohlinge kaufen.
120dB is the difference between just loud enough
to be heard by your
ear (1dB) and the loudness when the bones in your ear bow to protect
it. At higher levels (>120dB) your ear aches but it will recover if
the noise isn't permanent. At least thats the definition I learned...
I got my
numbers from http://en.wikipedia.org/wiki/Decibel#Safety but
obviously their numbers represent some sort of typical examples; I
been in a number of discotheques (here in the US we can't pronounce
that so we just say "clubs") that had sound levels far above 100 dB.
The amount of time the ear can be exposed to a given sound pressure
level before sustaining permanent damange clearly varies a lot from
person to person. But government standards are probably a good
guideline to start with, eg. OSHA table G-16:
http://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=STANDARDS&…
6:24 p.m.
Arnold Krille wrote:
Hi Paul and all (that rhymes :-),
2007/2/27, Paul Winkler <pw_lists(a)slinkp.com>om>:
Fletcher-Munson did the first (I believe) curves:
(http://en.wikipedia.org/wiki/Fletcher-Munson_curve
Others have refined.
see more: http://en.wikipedia.org/wiki/Equal-loudness_contours
On Tue, Feb 27, 2007 at 10:40:46PM +0100, Arnold
Krille wrote:
I stand corrected, my numbers where from my physics-book which states
that 120dB is the "Schmerzgrenze" and 0dB is the slightest whisper
human ear can hear. Problem is that these values vary on the frequency
so there is a correction curve standardised(?) 120dB is the difference between just loud enough
to be heard by your
ear (1dB) and the loudness when the bones in your ear bow to protect
it. At higher levels (>120dB) your ear aches but it will recover if
the noise isn't permanent. At least thats the definition I learned...
I got my
numbers from http://en.wikipedia.org/wiki/Decibel#Safety but
obviously their numbers represent some sort of typical examples; I
been in a number of discotheques (here in the US we can't pronounce
that so we just say "clubs") that had sound levels far above 100 dB.
The amount of time the ear can be exposed to a given sound pressure
level before sustaining permanent damange clearly varies a lot from
person to person. But government standards are probably a good
guideline to start with, eg. OSHA table G-16:
http://www.osha.gov/pls/oshaweb/owadisp.show_document?p_table=STANDARDS&…
in some A-something
paper so these corrected values for the human ear are called db(A)...
Yeah... dB doesn't mean a whole lot unless there is a reference.
see also: http://en.wikipedia.org/wiki/A-weighting
The OSHA standards _are_ important, so it's good that you listed them.
--
brad fuller
http://www.Sonaural.com/
+1 (408) 799-6124
22 Feb
22 Feb
4:43 p.m.
On 2/22/07, Folderol <folderol(a)ukfsn.org> wrote:
On Thu, 22 Feb 2007 13:48:04 -0600
millward <millward(a)ms.umanitoba.ca> wrote:
The higher bit rates (AFAIK) are used purely for recording and editing
within an application (Ardour uses 32-bit floating point WAV
internally, for instance), to keep a high precision of audio
resolution (similar to the HDRI resolutions used for 3D rendering).
You do have to ultimately downsample/downgrade to a lower resolution
for CDs and MP3s, of course.
-- Brett
------------------------------------------------------------
"In the rhythm of music a secret is hidden;
If I were to divulge it, it would overturn the world."
-- Jelaleddin Rumi
I've got a sound card capable of 96 bits, (
M-Audio audiofile 24/96 )
but my sound editor, Audacity for Linux, only goes up to 32 bits.
Is there a sound editor for Linux that can do higher than 32 bits?
Maybe I'm missing something here, but what on earth does anyone need
such high resolution for?
ISTR The humble CD is recorded at 18bit. which is over 200,000:1 and
more than 100dB. The orchestral dynamic range is I believe quoted at
90dB. I accept some extra headroom is nice, and the calculations aren't
quite so simple, but even 32bit comes out at mind boggling 200dB. 
6:39 p.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Folderol schrieb:
On Thu, 22 Feb 2007 13:48:04 -0600
millward <millward(a)ms.umanitoba.ca> wrote:
Since the first time, I compared the output of ams at 16/44.1 to the
output of the very same patch at 24/96 I know, what hi-res is good for ;-)
I've got a sound card capable of 96 bits, (
M-Audio audiofile 24/96 )
but my sound editor, Audacity for Linux, only goes up to 32 bits.
Is there a sound editor for Linux that can do higher than 32 bits?
Maybe I'm missing something here, but what on earth does anyone need
such high resolution for? -----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFF3huP1Aecwva1SWMRApR4AJ9WqKeRIwIAuNKg/cAJV2TErtEsFACfWjXm
VSbtA1e79X3beStCyoK7P2I=
=irhc
-----END PGP SIGNATURE-----
7:01 p.m.
======= On Friday 23 February 2007, Hartmut Noack wrote: =======
Since the first time, I compared the output of ams at 16/44.1 to the
output of the very same patch at 24/96 I know, what hi-res is good for ;-)
Which DAC have you used? The thing is, real resolution of almost all
DACs is not better rather 14-15 bits. Very rare and very expensive DACs
have real resolution about 20-21 bits.
23 Feb
23 Feb
4:43 a.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Andrew Gaydenko schrieb:
======= On Friday 23 February 2007, Hartmut Noack wrote: =======
Terratec EWX 24/96 / envy24
Since the first time, I compared the output of
ams at 16/44.1 to the
output of the very same patch at 24/96 I know, what hi-res is good for ;-)
Which DAC have you used? The thing is, real resolution of almost all
DACs is not better rather 14-15 bits. Very rare and very expensive DACs
have real resolution about 20-21 bits.
May it be :-) the switch from 16/44.1 to 24/96 is very well audible
anyway :-). I guess, for the realworld-usage of this card signal/noise
ratio is of greater importance then the question, if the DAC could
compute even greater resolution then the one I hear.
So having the very same DAC in a breakoutbox shielded from Fanmotors and
Powersupplies could increase the quality of the recordings...
best regards
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFF3qjc1Aecwva1SWMRAubIAJ903kgKzZu3gJAT1m10jyuRE+O3mwCdG/RF
1BC0o5UPFD0jt4c19UERBGc=
=XEmn
-----END PGP SIGNATURE-----
5:12 a.m.
======= On Friday 23 February 2007, Hartmut Noack wrote: =======
Well, I didn't say about 96KHz at all :-) What about difference between
16/44 and 24/44? As for me, I'm not sure, I'll detect difference at blind
testing. Probably, my ears aren't golden ;-)
At any case, IMHO the main problem is DACs realization rather format
limitation.
Andrew
The thing is, real resolution of almost all
DACs is not better rather 14-15 bits. Very rare and very expensive DACs
have real resolution about 20-21 bits.
May it be :-) the switch from 16/44.1 to 24/96 is very well audible
anyway :-). I guess, for the realworld-usage of this card signal/noise
ratio is of greater importance then the question, if the DAC could
compute even greater resolution then the one I hear.
22 Feb
22 Feb
5:48 p.m.
On Thu, 2007-02-22 at 13:48 -0600, millward wrote:
I've got a sound card capable of 96 bits, (
M-Audio audiofile 24/96 )
but my sound editor, Audacity for Linux, only goes up to 32 bits.
Is there a sound editor for Linux that can do higher than 32 bits?
You have misunderstood the specifications of that soundcard. 24/96
refers to a maximum depth of 24 bits (=32 bit float) and up to 96000Hz
sampling rate. This is /way/ enough for use with audible audio.
Sampo
6:04 p.m.
On 2/22/07, Sampo Savolainen <v2(a)iki.fi> wrote:
On Thu, 2007-02-22 at 13:48 -0600, millward wrote:
I have a further question on that... if I am using Jack (w/ Ardour),
does Jack automatically use the highest bit resolution possible for my
hardware (in my case, a Delta1010), and this is only overridden with
the 'force 16 bits' flag? (Overlooking any session exporting to a
file, of course).
-- Brett
------------------------------------------------------------
"In the rhythm of music a secret is hidden;
If I were to divulge it, it would overturn the world."
-- Jelaleddin Rumi
I've got a sound card capable of 96 bits, (
M-Audio audiofile 24/96 )
but my sound editor, Audacity for Linux, only goes up to 32 bits.
Is there a sound editor for Linux that can do higher than 32 bits?
You have misunderstood the specifications of that soundcard. 24/96
refers to a maximum depth of 24 bits (=32 bit float) and up to 96000Hz
sampling rate. This is /way/ enough for use with audible audio.
6:56 p.m.
On Thu, Feb 22, 2007 at 05:01:35PM -0500, Brett W. McCoy wrote:
I have a further question on that... if I am using
Jack (w/ Ardour),
does Jack automatically use the highest bit resolution possible for my
hardware (in my case, a Delta1010), and this is only overridden with
the 'force 16 bits' flag?
Yes, see alsa options in the man page.
--
Paul Winkler
http://www.slinkp.com
6:59 p.m.
On Thursday 22 February 2007, Brett W. McCoy wrote:
On 2/22/07, Sampo Savolainen <v2(a)iki.fi> wrote:
Yes (i use a delta 66 for jack):
loading driver ..
apparent rate = 48000
creating alsa driver ... hw:1|hw:1|1024|2|48000|0|0|nomon|swmeter|-|32bit
control device hw:1
configuring for 48000Hz, period = 1024 frames, buffer = 2 periods
ALSA: final selected sample format for capture: 32bit little-endian
ALSA: use 2 periods for capture
ALSA: final selected sample format for playback: 32bit little-endian
ALSA: use 2 periods for playback
It also tells you when it fails selecting a bit depth. I don't use 96khz yet
[too demanding disc space requirements]..
Flo
--
Palimm Palimm!
http://tapas.affenbande.org
On Thu, 2007-02-22 at 13:48 -0600, millward
wrote:
I have a further question on that... if I am using Jack (w/ Ardour),
does Jack automatically use the highest bit resolution possible for my
hardware (in my case, a Delta1010), and this is only overridden with
the 'force 16 bits' flag? (Overlooking any session exporting to a
file, of course). I've got a sound card capable of 96 bits, (
M-Audio audiofile 24/96 )
but my sound editor, Audacity for Linux, only goes up to 32 bits.
Is there a sound editor for Linux that can do higher than 32 bits?
You have misunderstood the specifications of that soundcard. 24/96
refers to a maximum depth of 24 bits (=32 bit float) and up to 96000Hz
sampling rate. This is /way/ enough for use with audible audio.
7:15 p.m.
On 2/22/07, Florian Schmidt <mista.tapas(a)gmx.net> wrote:
Yes (i use a delta 66 for jack):
loading driver ..
apparent rate = 48000
creating alsa driver ... hw:1|hw:1|1024|2|48000|0|0|nomon|swmeter|-|32bit
control device hw:1
configuring for 48000Hz, period = 1024 frames, buffer = 2 periods
ALSA: final selected sample format for capture: 32bit little-endian
ALSA: use 2 periods for capture
ALSA: final selected sample format for playback: 32bit little-endian
ALSA: use 2 periods for playback
It also tells you when it fails selecting a bit depth. I don't use 96khz yet
[too demanding disc space requirements]..
Yeah, I stick with 44.1 or 48.
-- Brett
------------------------------------------------------------
"In the rhythm of music a secret is hidden;
If I were to divulge it, it would overturn the world."
-- Jelaleddin Rumi
6468
days inactive
6474
days old
linux-audio-user@lists.linuxaudio.org
37 comments
22 participants
participants (22)
-
Andrew Gaydenko -
Arnold Krille -
Bengt Gördén -
Brad Fuller -
Brett W. McCoy -
carmen -
Chuckk Hubbard -
david -
Florian Schmidt -
Florin Andrei -
Folderol -
Frank Barknecht -
Hartmut Noack -
Joe Hartley -
millward -
Paul Davis -
Paul Winkler -
Rick Wright -
Sampo Savolainen -
Thomas Vecchione -
Tim Orford -
Vincent Jaubert