5 Oct
2013
5 Oct
'13
4:56 p.m.
Hi,
In the wake of overhauling Ardour's bar-graph meters, I've taken the
time to implement corresponding needle-style meters and more.
meters.lv2 features 10 needle-style meters. Mono and stereo variants of
DIN, Nordic, EBU, BBC and VU types (which are based on jmeters by Fons
Adriaensen).
Additionally it includes a Stereo Phase Correlation Meter (needle
Display), an EBU-R128 Meter (also based on Fons' implementation) with
Histogram and History display, a Digital True-Peak Meter (4x
Oversampling, using zita-resampler), a Stereo Phase Scope (Goniometer)
and a 31 Band Spectrum-Analyzer.
Source-code, info and screenshots can be found at
https://github.com/x42/meters.lv2
They have been tested in various environments and fine-tuned to match
their corresponding real hardware type by Chris Goddard and Axel Müller.
Currently the plugins come in Gtk2 (LV2 GtkUI) and openGL (LV2
#external-ui) variants. Both versions are installed in parallel.
The main reason for this is portability and compatibility. Yet the GTK
variant UI does run in the host's GUI-thread and may slow down UI
response (in particular the goniometer). The openGL versions of the
plugins do not have this issue, but require very recent versions of
libcairo, libpixman and libpango which are thread-safe. Please see the
README included with the source.
Caveat: Note that the plugins use the LV2 idle-interface (lv2 >= 1.4.2)
The plugin-host must to be compiled with 1.4.2 or a later version of the
LV2 SDK. It also must support the http://lv2plug.in/ns/ext/resize-port/
extension. At the time of writing Ardour 3.4 and jalv.gtk from SVN >=
5129 do fulfill these criteria.
as always, any feedback is welcome.
robin
7 Oct
7 Oct
10 a.m.
On 10/05/2013 06:56 PM, Robin Gareus wrote:
Additionally it includes a Stereo Phase Correlation
Meter (needle
Display), an EBU-R128 Meter (also based on Fons' implementation) with
Histogram and History display, a Digital True-Peak Meter (4x
Oversampling, using zita-resampler), a Stereo Phase Scope (Goniometer)
and a 31 Band Spectrum-Analyzer.
(insert usual friendly ridicule of animated VU meters here)
but hey, it does look nice! i've always had this fondness for jellyfish
meters. and the radar-style r128 meter is a very welcome addition to the
toolbox, having this history information at a glance is very helpful
indeed. thanks for sharing this!
built without problems and seems to run very smoothly.
i believe radar meters have a bright future in different applications.
for instance, to provide limiter history. fons' zita-dpl1 has short-term
history already, which is extremely helpful. but to imagine something
that listens to transport information and provides history of a whole
song, plus "locate to peaks" from the history so that i can manually
tame parts where the limiter cut in too sharply, that would be awesome.
i'm currently trying to make sense of the libcairo version number in
opensuse tumbleweed, but i guess our lady of the segmentation fault will
let me know if there is a problem...
thanks a lot,
jörn
--
Jörn Nettingsmeier
Lortzingstr. 11, 45128 Essen, Tel. +49 177 7937487
Meister für Veranstaltungstechnik (Bühne/Studio)
Tonmeister VDT
http://stackingdwarves.net
6:34 p.m.
On 10/07/2013 11:59 AM, Jörn Nettingsmeier wrote:
On 10/05/2013 06:56 PM, Robin Gareus wrote:
lol. yeah, VU meters just exist to look nice :)
Even most of the hardware ones are utterly useless (see "[LAD] vu-meter
DSP" this July). At least this one does conform to spec. Thanks Fons.
Additionally it includes a Stereo Phase
Correlation Meter (needle
Display), an EBU-R128 Meter (also based on Fons' implementation) with
Histogram and History display, a Digital True-Peak Meter (4x
Oversampling, using zita-resampler), a Stereo Phase Scope (Goniometer)
and a 31 Band Spectrum-Analyzer.
(insert usual friendly ridicule of animated VU meters here) but hey, it does look nice! i've always had this
fondness for jellyfish
meters.
The goniometer(TM) aka Stereo Phase Scope is not a jellyfish display. JF
would be a surround 5.1 variation.
I was actually positively surprised how much information is displayed by
a goniometer. It takes a while to learn to read it, but it's amazing.
We (Chris, Axel and I) went through various stages including
video-taping the LV2 plugin display in order to improve readability and
usefulness. The default preset is very close to that of a DK-Tech MSD600
and the radar-style r128 meter is a very welcome
addition to the
toolbox, having this history information at a glance is very helpful
indeed. thanks for sharing this!
built without problems and seems to run very smoothly.
great. thanks for the feedback.
i believe radar meters have a bright future in
different applications.
for instance, to provide limiter history. fons' zita-dpl1 has short-term
history already, which is extremely helpful. but to imagine something
that listens to transport information and provides history of a whole
song, plus "locate to peaks" from the history so that i can manually
tame parts where the limiter cut in too sharply, that would be awesome.
'locate to peaks' is a very nice idea. However I'm not aware of any
plugin standard that allows to control the host's transport. A LV2
extension could do that, though. zita-dlp1 could do this via jack-transport.
i'm currently trying to make sense of the libcairo
version number in
opensuse tumbleweed, but i guess our lady of the segmentation fault will
let me know if there is a problem...
Depends. Thread-safety and concurrency issues may rather produce strange
effects than segfault right away. It's mainly the EBUR128 meter (rapid
font rendering), Try adding two or more instances of the openGL variant.
It also depends on your graphics card. The plugins render directly to
graphic card memory. It works fine here with an old intel-945 and some
nvidia cards that I have tested: zero CPU. But Axel has various problems
with his Radeon on debian/stable. Also x11/xcb can make a huge
performance difference. -- Anyway, the concurrency issues
cairo/pixman/pango + x11/opengl have been fixed upstream and will soon
find their way int various distros if they have not yet done so.
Cheers!
robin
8 p.m.
On Monday 07 October 2013 15:53:52 Robin Gareus did opine:
On 10/07/2013 11:59 AM, Jِrn Nettingsmeier wrote:
Your attempt at applying a trade mark to that word is about 45 or 50 years
late, that was the term used to describe a rotary transformer used to
rotate the display in color tv (NTSC days) vectorscopes first built by
Tektronx.
On 10/05/2013 06:56 PM, Robin Gareus wrote:
lol. yeah, VU meters just exist to look nice :)
Even most of the hardware ones are utterly useless (see "[LAD] vu-meter
DSP" this July). At least this one does conform to spec. Thanks Fons.
Additionally it includes a Stereo Phase
Correlation Meter (needle
Display), an EBU-R128 Meter (also based on Fons' implementation) with
Histogram and History display, a Digital True-Peak Meter (4x
Oversampling, using zita-resampler), a Stereo Phase Scope
(Goniometer) and a 31 Band Spectrum-Analyzer.
(insert usual friendly ridicule of animated VU meters here) but hey, it does look nice! i've always had
this fondness for
jellyfish meters.
The goniometer(TM) aka Stereo Phase Scope is not a jellyfish display. JF
would be a surround 5.1 variation. I was actually positively surprised how much
information is displayed by
a goniometer. It takes a while to learn to read it, but it's amazing.
Yes it is. The learning part however, does separate the wannabees from the
doers though.
We (Chris, Axel and I) went through various stages
including
video-taping the LV2 plugin display in order to improve readability and
usefulness. The default preset is very close to that of a DK-Tech MSD600
Cheers, Gene
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
A prohibitionist is the sort of man one wouldn't care to drink with
-- even if he drank.
-- H. L. Mencken
A pen in the hand of this president is far more
dangerous than 200 million guns in the hands of
law-abiding citizens.
and the radar-style r128 meter is a very welcome
addition to the
toolbox, having this history information at a glance is very helpful
indeed. thanks for sharing this!
built without problems and seems to run very smoothly.
great. thanks for the feedback.
i believe radar meters have a bright future in
different applications.
for instance, to provide limiter history. fons' zita-dpl1 has
short-term history already, which is extremely helpful. but to
imagine something that listens to transport information and provides
history of a whole song, plus "locate to peaks" from the history so
that i can manually tame parts where the limiter cut in too sharply,
that would be awesome.
'locate to peaks' is a very nice idea. However I'm not aware of any
plugin standard that allows to control the host's transport. A LV2
extension could do that, though. zita-dlp1 could do this via
jack-transport.
i'm currently trying to make sense of the
libcairo version number in
opensuse tumbleweed, but i guess our lady of the segmentation fault
will let me know if there is a problem...
Depends. Thread-safety and concurrency issues may rather produce strange
effects than segfault right away. It's mainly the EBUR128 meter (rapid
font rendering), Try adding two or more instances of the openGL variant.
It also depends on your graphics card. The plugins render directly to
graphic card memory. It works fine here with an old intel-945 and some
nvidia cards that I have tested: zero CPU. But Axel has various problems
with his Radeon on debian/stable. Also x11/xcb can make a huge
performance difference. -- Anyway, the concurrency issues
cairo/pixman/pango + x11/opengl have been fixed upstream and will soon
find their way int various distros if they have not yet done so.
Cheers!
robin
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user(a)lists.linuxaudio.org
http://lists.linuxaudio.org/listinfo/linux-audio-user
8:08 p.m.
On 10/07/2013 09:59 PM, Gene Heskett wrote:
well, reading a stereo scope is not too hard really. even if you only
pick out some obvious aspects, it's still useful.
making sense of a video color vectorscope however (never seen an ntsc
one, but i believe the pal ones have similar output) seems to imply a
pact with dark forces :-D i'm quite happy that the days of analog video
are almost over and i don't have to learn that anymore.
now with SDI you get to think about eye diagrams, which is not much
better :(
--
Jörn Nettingsmeier
Lortzingstr. 11, 45128 Essen, Tel. +49 177 7937487
Meister für Veranstaltungstechnik (Bühne/Studio)
Tonmeister VDT
http://stackingdwarves.net
The
goniometer(TM) aka Stereo Phase Scope is not a jellyfish display. JF
would be a surround 5.1 variation.
Your attempt at applying a trade mark to that word is about 45 or 50 years
late, that was the term used to describe a rotary transformer used to
rotate the display in color tv (NTSC days) vectorscopes first built by
Tektronx.
I was actually positively surprised how much
information is displayed by
a goniometer. It takes a while to learn to read it, but it's amazing.
Yes it is. The learning part however, does separate the wannabees from the
doers though. 
9:04 p.m.
* Jörn Nettingsmeier <nettings(a)stackingdwarves.net> [2013-10-07 22:08]:
On 10/07/2013 09:59 PM, Gene Heskett wrote:
well, reading a stereo scope is not too hard really. even if you
only pick out some obvious aspects, it's still useful.
making sense of a video color vectorscope however (never seen an
ntsc one, but i believe the pal ones have similar output) seems to
well, isn't
NTSC the abbreviation for "never the same color"?
(sorry, stupid joke)
P.
The
goniometer(TM) aka Stereo Phase Scope is not a jellyfish display. JF
would be a surround 5.1 variation.
Your attempt at applying a trade mark to that word is about 45 or 50 years
late, that was the term used to describe a rotary transformer used to
rotate the display in color tv (NTSC days) vectorscopes first built by
Tektronx.
I was actually positively surprised how much
information is displayed by
a goniometer. It takes a while to learn to read it, but it's amazing.
Yes it is. The learning part however, does separate the wannabees from the
doers though.
10:13 p.m.
On 10/07/2013 10:08 PM, Jörn Nettingsmeier wrote:
well, reading a stereo scope is not too hard really. even if you only
pick out some obvious aspects, it's still useful.
It's mainly about learning to recognize certain patters. and/or what
effects cause or counterbalance them.
What I found interesting that it visualizes pretty much every aspect of
the sound: Stereo-phase (obviously), frequency distribution (smooth vs
jittery lines), amplitude per channel,.. when used with a persistent
display also the history of that.
While implementing it, I realized how important various aspects of the
display are. It's not just connecting a scope and drawing Lissajous
figures. Compression (auto-gain) as well as screen-persistence and line
or dot thickness are important details that not only make it usable in
an audio-context but also very relevant for people who are used to
'read' real hardware goniometers. Many thanks go to Chris and Alex for
nitpicking about all this. We're still not quite done there, either.
2c,
robin
I was actually positively surprised how much
information is displayed by
a goniometer. It takes a while to learn to read it, but it's amazing.
Yes it is. The learning part however, does separate the wannabees
from the doers though. 
9 Oct
9 Oct
6:15 p.m.
>>>> "R" == Robin Gareus:
R> What I found interesting that it visualizes pretty much every
R> aspect of the sound: Stereo-phase (obviously), frequency
R> distribution (smooth vs jittery lines), amplitude per
R> channel,..
These kinds of visualisations are very useful for mono-sources as well
(just combing a signal with a delayed version of itself - ie. lagged
embedding) simultaneously providing visual info from sample- up to
(micro)phrase-level. Looking at the characteristics of the waveform
along with the amplitude - even the 'noise-character' of the reverb-tail
is displayed.
I did a video once back in 1996 (on sgis, rendering frames for one full
week...) testing this approach:
http://users.notam02.no/~andersvi/works/OIOI/OIOI.avi (39Mb, music made
with CLM).
Guess its time to test these things in real-time now, adjusting delays,
colour-schemes, fade-times etc. on the go for the graphics.
-anders
26 Oct
26 Oct
8:34 p.m.
On Wed, Oct 09, 2013 at 08:15:36PM +0200, anders.vinjar(a)bek.no wrote:
Wow, that looks and sounds great.
Well done...
>>>> "R" == Robin Gareus:
R> What I found interesting that it visualizes pretty much every
R> aspect of the sound: Stereo-phase (obviously), frequency
R> distribution (smooth vs jittery lines), amplitude per
R> channel,..
These kinds of visualisations are very useful for mono-sources as well
(just combing a signal with a delayed version of itself - ie. lagged
embedding) simultaneously providing visual info from sample- up to
(micro)phrase-level. Looking at the characteristics of the waveform
along with the amplitude - even the 'noise-character' of the reverb-tail
is displayed.
I did a video once back in 1996 (on sgis, rendering frames for one full
week...) testing this approach:
http://users.notam02.no/~andersvi/works/OIOI/OIOI.avi (39Mb, music made
with CLM).
7 Oct
7 Oct
11:07 p.m.
On Monday 07 October 2013 18:46:19 Jörn Nettingsmeier did opine:
On 10/07/2013 09:59 PM, Gene Heskett wrote:
well, reading a stereo scope is not too hard really. even if you only
pick out some obvious aspects, it's still useful.
making sense of a video color vectorscope however (never seen an ntsc
one, but i believe the pal ones have similar output)
Yes. Pal versions I think show two burst vectors IIRC, one for each field.
At least thats what I remember after seeing one at the NAB in '79 or so.
Rather rare birds here in Never Twice Same Color land.
The
goniometer(TM) aka Stereo Phase Scope is not a jellyfish display.
JF would be a surround 5.1 variation.
Your attempt at applying a trade mark to that word is about 45 or 50
years late, that was the term used to describe a rotary transformer
used to rotate the display in color tv (NTSC days) vectorscopes first
built by Tektronx.
I was actually positively surprised how much
information is displayed
by a goniometer. It takes a while to learn to read it, but it's
amazing.
Yes it is. The learning part however, does separate the wannabees
from the doers though. seems to imply a
pact with dark forces :-D
I don't know as I'd go that far, but it does require a basic understanding
of the theory.
One of the things I learned early on was that despite the density of skin
tone, the color itself never changed. Lots of folks will never believe
that, but its true, the vector angle of a skin tone remains within 2 or 3
degrees from the whitest Caucasian to the blackest African. The really
black black may not show as much amplitude, but if the camera isn't lying,
not by enough to be able to look at the scope and say that person is black.
We have had, for many years at my station, a pair of cousins who are color
blind, and did all their production work by checking it on the vectorscope.
It was right spot on moreso than the help that claimed good color vision.
i'm quite happy that the days of analog video
are almost over and i don't have to learn that anymore.
now with SDI you get to think about eye diagrams, which is not much
better :(
That and the constellation. Without good performance there, you can't get
even a 100kw signal out of your shirt pocket.
Cheers, Gene
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Suburbia is where the developer bulldozes out the trees, then names
the streets after them.
-- Bill Vaughn
A pen in the hand of this president is far more
dangerous than 200 million guns in the hands of
law-abiding citizens.
4198
days inactive
4219
days old
linux-audio-user@lists.linuxaudio.org
9 comments
6 participants
participants (6)
-
anders.vinjar＠bek.no -
Gene Heskett -
Jörn Nettingsmeier -
Lieven Moors -
Peter P. -
Robin Gareus