[Robert Orban]

In article -
west.giganews.com, says...


In article Gd-
,
says...

I think that this computer may have some problems with DC
offsets
in the
digital domain. Some of the Orban metering code might not
ignore
them.

The meter doesn't block DC. That gives most accurate peak
indications because AC coupling introduces tilt into the
waveform
unless the cutoff frequency is lower than around 0.15 Hz.

Moreover, if the meter blocked DC and there was substantial DC
on
the audio, the meter would not detect clipping due to the
asymmetry.

....that being said, both loudness meters block DC because of the
highpass part of their equal-loudness contour weighting filters. It
is the VU and PPM meters that do not block DC.

[Arny Krueger]

"Robert Orban" wrote in message

In article -
west.giganews.com, says...


In article Gd-
,
says...

I think that this computer may have some problems with
DC offsets in the digital domain. Some of the Orban
metering code might not
ignore
them.

The meter doesn't block DC. That gives most accurate peak
indications because AC coupling introduces tilt into the
waveform
unless the cutoff frequency is lower than around 0.15 Hz.

Moreover, if the meter blocked DC and there was
substantial DC
on
the audio, the meter would not detect clipping due to
the asymmetry.

...that being said, both loudness meters block DC because
of the highpass part of their equal-loudness contour
weighting filters. It is the VU and PPM meters that do
not block DC.

I would argue that the reasons for not blocking DC apply equally to all 4
meters. If it is OK to block DC for two of them, then it is OK for all of
them.

[Robert Orban]

In article
,
says...


"Robert Orban" wrote in message
-
west.giganews.com
In article -
west.giganews.com, says...


In article Gd-
,
says...

I think that this computer may have some problems with
DC offsets in the digital domain. Some of the Orban
metering code might not
ignore
them.

The meter doesn't block DC. That gives most accurate peak
indications because AC coupling introduces tilt into the
waveform
unless the cutoff frequency is lower than around 0.15 Hz.

Moreover, if the meter blocked DC and there was
substantial DC
on
the audio, the meter would not detect clipping due to
the asymmetry.

...that being said, both loudness meters block DC because
of the highpass part of their equal-loudness contour
weighting filters. It is the VU and PPM meters that do
not block DC.

I would argue that the reasons for not blocking DC apply equally to
all 4
meters. If it is OK to block DC for two of them, then it is OK for all of
them.

Ther are good reasons for blocking DC to the loudness meters
because the ear does not responds to DC, so DC las no loudness
level.

However, the PPM, absolute peak, and VU meters are level meters.
The insidious thing about blocking DC to peak level meters is that
you no longer see the correct values of the digital samples so you
can't detect clipping accurately.

[Don Pearce]

On Sat, 12 Apr 2008 11:41:58 -0800, Robert Orban
wrote:

In article ,
says...


On Fri, 11 Apr 2008 11:33:40 -0800, Robert Orban
wrote:

In article ,
says...


On Fri, 11 Apr 2008 06:52:15 -0400, "Arny Krueger"

wrote:

"Robert Orban" wrote in
message
-
west.giganews.com

The software accepts two-channel stereo inputs. The VU
and PPM meters are split to indicate the left and right
channels. The PPM meter also displays the instantaneous
peak values of the L and R digital samples.

When I started the program, the Vu and PPM meters started
reading
mid-scale
and full-scale respectively, with no music playing. This seems
counter-intuitive.


Works as expected for me. my only issue is that the PPM isn't
scaled
as a normal PPM.

I spent some time soul-searching about the issue of scales and
finally
chose to put all of the meters on identical scales to make it easier
for
users to compare the meters. In future version, we may make
scales a
user-definable parameter.

Bob Orban

Can you explain a little more about the variable gain/sensitivity
controls available for a couple of the scales? I'm not really too sure
how one should be using those.

Because the scales are all -30 to 0 dBfs, setup is really somewhat
arbitrary. If you set the VU Gain to +10 dB, the VU meter will indicate
the same as the peak meter on a sinewave. (We sill probably change
this is the next revision so that that this will happen with the VU Gain
set to 0.)

As for the CBS loudness meter, there was never an "official"
calibration published. As stated in the readme, if I want to compare it
with the BS.1770 meter, I set the CBS gain to +7 dB if I want to
compare the peak excursions of the meter wit BS.1770 and to +10 if I
want to compare the experimental long-term CBS loudness to BS.1770.


OK. here is what I get.

Generate a tone at -18dB in Audition, and it shows -14dB on both VU
and PPM, with VU gain set to 10dB. Is that what you would expect?

BS 1770 also shows roughly -14dB, and I can make CBS match that if I
wind its gain up to 18dB.

I am going to try some loudness testing later. I will choose several
different music types and change their amplitudes until they are
subjectively as loud as each other. Then I will see how the CBS and
BS.1770 respond. I presume the tell-tale left behind by the CBS is
roughly equivalent to the main bar of the BS1770?

d

--

d

[Robert Orban]

Our developer is actively working on bug reports and contacting
users who report them. This is hard for USENET posts because
many people use phony email addresses to foil spambots. In
addition to posting them here, please send bug reports to the email
address on the Settings page of the meter. (It's also in the readme.)

Bob Orban

[Robert Orban]

In article ,
says...


On Sat, 12 Apr 2008 11:41:58 -0800, Robert Orban
wrote:

In article ,
says...


On Fri, 11 Apr 2008 11:33:40 -0800, Robert Orban
wrote:

In article ,
says...


On Fri, 11 Apr 2008 06:52:15 -0400, "Arny Krueger"

wrote:

"Robert Orban" wrote in
message
-
west.giganews.com

The software accepts two-channel stereo inputs. The VU
and PPM meters are split to indicate the left and right
channels. The PPM meter also displays the instantaneous
peak values of the L and R digital samples.

When I started the program, the Vu and PPM meters started
reading
mid-scale
and full-scale respectively, with no music playing. This seems
counter-intuitive.


Works as expected for me. my only issue is that the PPM isn't
scaled
as a normal PPM.

I spent some time soul-searching about the issue of scales and
finally
chose to put all of the meters on identical scales to make it easier
for
users to compare the meters. In future version, we may make
scales a
user-definable parameter.

Bob Orban

Can you explain a little more about the variable gain/sensitivity
controls available for a couple of the scales? I'm not really too
sure
how one should be using those.

Because the scales are all -30 to 0 dBfs, setup is really somewhat
arbitrary. If you set the VU Gain to +10 dB, the VU meter will
indicate
the same as the peak meter on a sinewave. (We sill probably
change
this is the next revision so that that this will happen with the VU
Gain
set to 0.)

As for the CBS loudness meter, there was never an "official"
calibration published. As stated in the readme, if I want to compare
it
with the BS.1770 meter, I set the CBS gain to +7 dB if I want to
compare the peak excursions of the meter wit BS.1770 and to +10 if
I
want to compare the experimental long-term CBS loudness to
BS.1770.


OK. here is what I get.

Generate a tone at -18dB in Audition, and it shows -14dB on both
VU
and PPM, with VU gain set to 10dB. Is that what you would expect?

I would expect a sinewave peaking at 0 dBfs to indicate 0 on the
loudness meter's PPM and absolute peak meter. I would also expect
the VU meter to read 0 dB when its gain is set for +10 dB. Because
this is admittedly confusing, the next release will be modified so that
when the VU gain is set to 0 dB, the VU meter agrees with the
PPM and absolute peak when fed with a sinewave.


BS 1770 also shows roughly -14dB, and I can make CBS match that if
I
wind its gain up to 18dB.

I am going to try some loudness testing later. I will choose several
different music types and change their amplitudes until they are
subjectively as loud as each other. Then I will see how the CBS and
BS.1770 respond. I presume the tell-tale left behind by the CBS is
roughly equivalent to the main bar of the BS1770?

There are two hold bars in the CBS. One jut holds the peak indication
of the meter, while the other (the cyan bar) is an experimental long-
term loudness measurement. Please see the readme for details.

The experimental long-term loudness indication still can get fooled by
certain types of program material, so I expect I will refine the algorithm
further in future releases.

[Don Pearce]

On Sat, 12 Apr 2008 23:38:59 -0800, Robert Orban
wrote:

OK. here is what I get.

Generate a tone at -18dB in Audition, and it shows -14dB on both
VU
and PPM, with VU gain set to 10dB. Is that what you would expect?

I would expect a sinewave peaking at 0 dBfs to indicate 0 on the
loudness meter's PPM and absolute peak meter. I would also expect
the VU meter to read 0 dB when its gain is set for +10 dB. Because
this is admittedly confusing, the next release will be modified so that
when the VU gain is set to 0 dB, the VU meter agrees with the
PPM and absolute peak when fed with a sinewave.

No, I definitely see a 4dB discrepancy between peak sine wave level
and the indications on PPM and VU (+10). The meters read higher than
the actual signal level. The sound card is an M-Audio Audiophile 2496
if that is any help.

See this screen shot for the situation CBS loudness gain is +18, VU
meter gain is +10.

http://81.174.169.10/odds/orbanppm.gif


--

d

[Robert Orban]

In article ,
says...


On Sat, 12 Apr 2008 23:38:59 -0800, Robert Orban
wrote:

OK. here is what I get.

Generate a tone at -18dB in Audition, and it shows -14dB on
both
VU
and PPM, with VU gain set to 10dB. Is that what you would
expect?

I would expect a sinewave peaking at 0 dBfs to indicate 0 on the
loudness meter's PPM and absolute peak meter. I would also
expect
the VU meter to read 0 dB when its gain is set for +10 dB. Because
this is admittedly confusing, the next release will be modified so
that
when the VU gain is set to 0 dB, the VU meter agrees with the
PPM and absolute peak when fed with a sinewave.

No, I definitely see a 4dB discrepancy between peak sine wave
level
and the indications on PPM and VU (+10). The meters read higher
than
the actual signal level. The sound card is an M-Audio Audiophile
2496
if that is any help.

See this screen shot for the situation CBS loudness gain is +18, VU
meter gain is +10.

http://81.174.169.10/odds/orbanppm.gif

I will investigate this.

My first guess is that the 4 dB you are seeing is actually 3.922 dB [20
log10 (2/pi)], which is the difference between the peak and average
value of a sinewave. If I recall correctly, Audition allows you to scale
levels according to the peak or average value of a sinewave, so this
is what I would investigate.

[Michael Rempel[_2_]]

On Apr 11, 2:33 pm, Robert Orban wrote:
In article ,
says...

On Fri, 11 Apr 2008 06:52:15 -0400, "Arny Krueger"





wrote:

"Robert Orban" wrote in message
-
west.giganews.com

The software accepts two-channel stereo inputs. The VU
and PPM meters are split to indicate the left and right
channels. The PPM meter also displays the instantaneous
peak values of the L and R digital samples.

When I started the program, the Vu and PPM meters started reading
mid-scale
and full-scale respectively, with no music playing. This seems
counter-intuitive.

Works as expected for me. my only issue is that the PPM isn't scaled
as a normal PPM.

I spent some time soul-searching about the issue of scales and finally
chose to put all of the meters on identical scales to make it easier for
users to compare the meters. In future version, we may make scales a
user-definable parameter.

Bob Orban

Hey a couple more suggestions:

Add sensitivity to the list, and possibly a compensating EQ setting to
get it flat with a known measurement mic.

Add ASIO drivers to your support list too, I dont know too many
serious audio users who use anything else most of the time.

[Robert Orban]

The first update is now available.

Version 1.0.1:

--Reduces CPU load caused by refreshing the meter’s display. This
allows slower computers to operate at the meter’s maximum 100 Hz
refresh rate, minimizing flicker.

--Changes the color scheme to improve appearance and to make the
meters easier to read.

--Changes the graphic design of the single-bar meter display elements
like peak hold to make them easier to read and to prevent them from
being obscured.

--On startup, checks whether the computer’s CPU supports the SSE2
instruction set and exits gracefully if it does not.

--Increases the gain of the VU meter by 10 dB for a given setting of
the VU Meter Gain control. When the control is set to 0 dB and the
meter is fed by a sinewave, the VU meter will now display the same
level as the absolute peak meter.

--Clarifies the readme to better explain how the meter interacts with
your computer’s sound device(s).

--Moves the Audio Input selector to the Settings page.

The update is available he

http://www.orban.com/meter/

[Don Pearce]

Robert Orban wrote:
The first update is now available.

Version 1.0.1:

--Reduces CPU load caused by refreshing the meter’s display. This
allows slower computers to operate at the meter’s maximum 100 Hz
refresh rate, minimizing flicker.

--Changes the color scheme to improve appearance and to make the
meters easier to read.

--Changes the graphic design of the single-bar meter display elements
like peak hold to make them easier to read and to prevent them from
being obscured.

--On startup, checks whether the computer’s CPU supports the SSE2
instruction set and exits gracefully if it does not.

--Increases the gain of the VU meter by 10 dB for a given setting of
the VU Meter Gain control. When the control is set to 0 dB and the
meter is fed by a sinewave, the VU meter will now display the same
level as the absolute peak meter.

--Clarifies the readme to better explain how the meter interacts with
your computer’s sound device(s).

--Moves the Audio Input selector to the Settings page.

The update is available he

http://www.orban.com/meter/


Excellent. That now behaves exactly as I would like a meter set to. It
is, as you say, useful to have all the meters scaled the same way, but
as the PPM is at the bottom of the stack, perhaps you could add a
selectable set of normal PPM scales *below* that meter.

d

[Don Pearce]

Robert Orban wrote:
The first update is now available.

Version 1.0.1:

--Reduces CPU load caused by refreshing the meter’s display. This
allows slower computers to operate at the meter’s maximum 100 Hz
refresh rate, minimizing flicker.

--Changes the color scheme to improve appearance and to make the
meters easier to read.

--Changes the graphic design of the single-bar meter display elements
like peak hold to make them easier to read and to prevent them from
being obscured.

--On startup, checks whether the computer’s CPU supports the SSE2
instruction set and exits gracefully if it does not.

--Increases the gain of the VU meter by 10 dB for a given setting of
the VU Meter Gain control. When the control is set to 0 dB and the
meter is fed by a sinewave, the VU meter will now display the same
level as the absolute peak meter.

--Clarifies the readme to better explain how the meter interacts with
your computer’s sound device(s).

--Moves the Audio Input selector to the Settings page.

The update is available he

http://www.orban.com/meter/


And something I forgot. Get rid of the Start/Stop button. Just stop the
meters when choosing the settings tab, and restart when returning.

d

[Robert Orban]

In article ao2dncaCm7psWIvVRVnygwA@plusnet, says...


And something I forgot. Get rid of the Start/Stop button. Just stop the
meters when choosing the settings tab, and restart when returning.

The Start/Stop button is useful for two things:

1. The meter must be stopped for certain settings to be changed, mainly
the integration time for the CBS and BS.1770 meters. Doing this on the
fly is tricky because the integrators weight all samples equally
within the time window, so there is no natural exponential decay for
bogus samples that get into the integrator. It's safest just to flush the
integrators and set all samples to zero when the integration time is
changed.

2. The meter can be to compute the long-term loudness of program elements
up to 10 seconds in length. Doing this correctly requires starting the
meter at the same time that the program element starts.

Bob Orban

[Don Pearce]

Robert Orban wrote:
In article ao2dncaCm7psWIvVRVnygwA@plusnet, says...

And something I forgot. Get rid of the Start/Stop button. Just stop the
meters when choosing the settings tab, and restart when returning.

The Start/Stop button is useful for two things:

1. The meter must be stopped for certain settings to be changed, mainly
the integration time for the CBS and BS.1770 meters. Doing this on the
fly is tricky because the integrators weight all samples equally
within the time window, so there is no natural exponential decay for
bogus samples that get into the integrator. It's safest just to flush the
integrators and set all samples to zero when the integration time is
changed.

2. The meter can be to compute the long-term loudness of program elements
up to 10 seconds in length. Doing this correctly requires starting the
meter at the same time that the program element starts.

Bob Orban


OK, but you what is wrong with just automatically stopping the meters
when you hit the settings tab? I can't see that what you say, while it
is clearly correct, has any bearing on this.

Now, how about the little boxes for individual instruments. When you
remove the cross to disable, the meter just sits wherever it was when
you hit the button. Much better to remove the bar and max telltales.

d

[Robert Orban]

In article 0eydncwpV7-CFIrVnZ2dnUVZ8qugnZ2d@plusnet,
says...


Now, how about the little boxes for individual instruments. When you
remove the cross to disable, the meter just sits wherever it was when
you hit the button. Much better to remove the bar and max telltales.

I agree. We made this change in v 1.02.