Old 04-14-2016, 08:40 AM   #41
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by heda View Post
Hi TBProAudio
One question regarding SL, in both the JS version and dpMeter.
I've noticed that SL values don't refresh as often as ML or IL. Could it be improved and be smooth as ML or IL?
Thank you for your great plugins.
Well this is correct:
new ML/IL value is calculated every 300ms, SL every 2000ms, according EBU R 128 spec.
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-14-2016, 08:48 AM   #42
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by RJHollins View Post
? does the AB_LM have the 1.9 update ?

Hmm, maybe i do not understand you?

AB_LM does not use the same internal lib, just because of efficiency reasons (no TP/IL/LRA measurement, different update speed for SL/ML).
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-14-2016, 09:01 AM   #43
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by Castanuelas View Post
I am happy dpMeter supports all the sample rates I need.

I looked at the manual and the web site for dpMeter, and I didn't see any information about which version of BS.1770 it uses. I see that BS.1770-4 is the current version: http://www.itu.int/rec/R-REC-BS.1770/en

I recall BS.1770-3 has gating on the integrated loudness, which is important for movies and classical music. I don't know what changes BS.1770-4 introduced.
dpMeter supports latest specification EBU R 128 V3 (2014) completely. All test cases are passed.
This includes gated IL/LRA measurement (acc. ITU-R BS.1770-4).
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-14-2016, 09:21 AM   #44
heda
Human being with feelings
 
heda's Avatar
 
Join Date: Jun 2012
Location: Spain
Posts: 7,239
Default

Quote:
Originally Posted by TBProAudio View Post
Well this is correct:
new ML/IL value is calculated every 300ms, SL every 2000ms, according EBU R 128 spec.
I know you know it is 400ms and 3000ms, but maybe there was a way to calculate a partial buffer if 3000ms have not yet passed. I don't know. Don't worry about it. I asked because I've seen SL being smoothly represented in other EBUR128 plugins from other developers. Maybe they just interpolate values.


Another question, thanks:
here and also in dpMeter. Why the -60dB minimum value? I think many users need lower values, including me. -100 would be nice.
heda is offline   Reply With Quote
Old 04-14-2016, 09:50 AM   #45
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by heda View Post
I know you know it is 400ms and 3000ms, but maybe there was a way to calculate a partial buffer if 3000ms have not yet passed. I don't know. Don't worry about it. I asked because I've seen SL being smoothly represented in other EBUR128 plugins from other developers. Maybe they just interpolate values.
Calculation of IL/SL/ML/LRA is implemented according EBU spec including respective timing.
New ML/IL values are calculated every 100ms and SL every 1000ms (i was wrong with the 2000ms). If there is no new value, old is used, no interpolation.

Quote:
Originally Posted by heda View Post
Another question, thanks:
here and also in dpMeter. Why the -60dB minimum value? I think many users need lower values, including me. -100 would be nice.
This is a good question: decision by design. I could set it to -144.0 if requested.
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-14-2016, 10:01 AM   #46
heda
Human being with feelings
 
heda's Avatar
 
Join Date: Jun 2012
Location: Spain
Posts: 7,239
Default

ok now I understand better about how it has to be calculated. thanks!

yes -144.0 minimum would be great! at least you could only change it in the automation outputs of dpMeter without changing the design.
heda is offline   Reply With Quote
Old 04-14-2016, 11:52 AM   #47
notwa
Human being with feelings
 
Join Date: Jul 2013
Posts: 44
Default

Quote:
Originally Posted by TBProAudio View Post
I have not yet found any translation to 88/96k, so I would say: not available.

But I don't think that coefficients for 88/96k make really sense, because current TV play-out rate is 44k or 48k. But this may change in future.
I use these settings for my BS.1770-3 implementation in python:
a 1501 Hz second-order highshelf with 4dB of gain and a bandwidth of 1.0,
a 38.135457 Hz second-order highpass with a Q of 0.5003268,
and 1.004995 (linear, not dB) overall gain.

here's a quick and dirty plot of error against the coefficients provided by the standard. it's pretty insignificant.

I just glanced at the pdf, and the coefficients for BS.1770-4 seem to be unchanged from version 3.

edit: sorry, I just realized you were talking about the true-peak oversampling coefficients. I hope these filter parameters are useful regardless!

Last edited by notwa; 04-14-2016 at 12:03 PM.
notwa is offline   Reply With Quote
Old 04-14-2016, 01:11 PM   #48
Castanuelas
Human being with feelings
 
Join Date: Sep 2012
Posts: 74
Default

Quote:
Originally Posted by TBProAudio View Post
dpMeter supports latest specification EBU R 128 V3 (2014) completely. All test cases are passed.
This includes gated IL/LRA measurement (acc. ITU-R BS.1770-4).
I am really impressed. Great job! You may want to clarify the documentation about versions of specification compliance.

Quote:
Originally Posted by TBProAudio View Post
TP measurement of dpMeter supports all 44/48K "times factor" sample-rates.The signal is always 4 times up-sampled, even if this may not be necessary for 88K and higher according BS1770.
IL, ML and SL measurement works for all sample-rates.
Since I only use plugins with 64-bit audio-path bit-depth (double type variables in C/C++), I am very happy that dpMeter supports this. However, dpMeter is a little heavy in CPU at 192k. With the EBU R128 mode, I measured 3.5X CPU at 192k compared to 44k, and I have more trouble achieving real-time processing at 192k. I request disabling upsampling for TP measurement at 176k and 192k.
Castanuelas is offline   Reply With Quote
Old 04-14-2016, 01:48 PM   #49
musicbynumbers
Human being with feelings
 
musicbynumbers's Avatar
 
Join Date: Jun 2009
Location: South, UK
Posts: 14,214
Default

Isn't the -60dB thing because of the gate that stops silence being included in the calculation?

Maybe not, but just a thought
__________________
subproject FRs click here
note: don't search for my pseudonym on the web. The "musicbynumbers" you find is not me or the name I use for my own music.
musicbynumbers is offline   Reply With Quote
Old 04-14-2016, 10:07 PM   #50
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by heda View Post
ok now I understand better about how it has to be calculated. thanks!

yes -144.0 minimum would be great! at least you could only change it in the automation outputs of dpMeter without changing the design.
OK; next version of dpMeter will have minimum of -144.0.
Except IL with minimum of -70.0 (EBU R128 absolute gate).
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-14-2016, 10:17 PM   #51
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by notwa View Post
I use these settings for my BS.1770-3 implementation in python:
a 1501 Hz second-order highshelf with 4dB of gain and a bandwidth of 1.0,
a 38.135457 Hz second-order highpass with a Q of 0.5003268,
and 1.004995 (linear, not dB) overall gain.

here's a quick and dirty plot of error against the coefficients provided by the standard. it's pretty insignificant.

I just glanced at the pdf, and the coefficients for BS.1770-4 seem to be unchanged from version 3.

edit: sorry, I just realized you were talking about the true-peak oversampling coefficients. I hope these filter parameters are useful regardless!
Sorry for confusing. I mixed some stuff:
Yes, we were talking about TP measurement. I talked about up-sampling filter coefficients, There is a definition in the spec BS.1770 for 48k. But there are no coefficients defined for 96k, because 48k coefficients are used.
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-14-2016, 10:22 PM   #52
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by musicbynumbers View Post
Isn't the -60dB thing because of the gate that stops silence being included in the calculation?

Maybe not, but just a thought
Well right, but absolute gate should be -70.0 (according EBU R128), so I will set IL automation minimum value to the same value.
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-14-2016, 10:44 PM   #53
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by Castanuelas View Post
Since I only use plugins with 64-bit audio-path bit-depth (double type variables in C/C++), I am very happy that dpMeter supports this. However, dpMeter is a little heavy in CPU at 192k. With the EBU R128 mode, I measured 3.5X CPU at 192k compared to 44k, and I have more trouble achieving real-time processing at 192k. I request disabling upsampling for TP measurement at 176k and 192k.
Well, right. I will consider a different, more sample rate sensitive implementation.
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-15-2016, 09:43 AM   #54
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Update to V2.0:
- TP measurement adjusted according BS 1770 beyond 44/48K

In fact:
4x Os for srate <= 48kHz
2x OS for srate <= 96kHz
no OS for srate > 96kHz
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-15-2016, 09:58 AM   #55
heda
Human being with feelings
 
heda's Avatar
 
Join Date: Jun 2012
Location: Spain
Posts: 7,239
Default

Quote:
Originally Posted by TBProAudio View Post
OK; next version of dpMeter will have minimum of -144.0.
Except IL with minimum of -70.0 (EBU R128 absolute gate).
thank you

Regarding the SL update rate... I've read this in the new January 2016 document "EBU Mode"
in page 6: https://tech.ebu.ch/docs/tech/tech3341.pdf

Quote:
The short-term loudness uses a sliding rectangular time window of length 3s. The measurement is not gated. The update rate for ‘live meters’ shall be at least 10 Hz.
So, that's 10 times than current refresh rate for dpMeter.

it would be great to make dpMeter "EBU Mode" compatible.
heda is offline   Reply With Quote
Old 04-15-2016, 10:15 AM   #56
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

should not a big issue
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-15-2016, 10:47 AM   #57
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Update to V2.1: SL/ML update every 100ms according EBU R128 spec

BTW: dpMeter will get all those fixes very soon.
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-17-2016, 11:23 PM   #58
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Update to 2.2, small fix

BTW: dpMeter is now updated as well
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-18-2016, 12:07 AM   #59
planetnine
Human being with feelings
 
planetnine's Avatar
 
Join Date: Oct 2007
Location: Lincoln, UK
Posts: 7,924
Default

I haven't downloaded this yet (I have my own LU meters), but I'd like to say I'm impressed with the effort and attention to detail to support a "free" plugin.

Kudos to you sir!



>
__________________
Nathan, Lincoln, UK. | Item Marker Tool. (happily retired) | Source Time Position Tool. | CD Track Marker Tool. | Timer Recording Tool. | dB marks on MCP faders FR.
planetnine is offline   Reply With Quote
Old 04-18-2016, 08:55 AM   #60
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by planetnine View Post
I haven't downloaded this yet (I have my own LU meters), but I'd like to say I'm impressed with the effort and attention to detail to support a "free" plugin.

Kudos to you sir!
>
Thank you Sir!
Well it uses the same code base as dpMeter, so it is fair keep both up to date
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 04-18-2016, 09:53 AM   #61
Tale
Human being with feelings
 
Tale's Avatar
 
Join Date: Jul 2008
Location: The Netherlands
Posts: 3,645
Default

I don't have time to really look at this right now, but this seems very useful. Thanks for sharing!
Tale is offline   Reply With Quote
Old 04-19-2016, 05:31 AM   #62
Castanuelas
Human being with feelings
 
Join Date: Sep 2012
Posts: 74
Default

I tested dpMeter v1.4.2, and the CPU is about 40% of v1.4.1 at 192k. Thank you!
Castanuelas is offline   Reply With Quote
Old 04-19-2016, 08:49 AM   #63
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by Castanuelas View Post
I tested dpMeter v1.4.2, and the CPU is about 40% of v1.4.1 at 192k. Thank you!
I hope it is better now
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 08-19-2017, 06:28 AM   #64
DarkStar
Human being with feelings
 
DarkStar's Avatar
 
Join Date: May 2006
Location: Surrey, UK
Posts: 19,677
Default

Just been working on this (the orange numbers are the maxima):



But the Loudness Range is not the same as in dpmeter2.
Q1: which one is right?
Q2: can the other one be fixed?
__________________
DarkStar ... interesting, if true. . . . Inspired by ...
DarkStar is offline   Reply With Quote
Old 08-19-2017, 06:44 AM   #65
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by DarkStar View Post
Just been working on this (the orange numbers are the maxima):

But the Loudness Range is not the same as in dpmeter2.
Q1: which one is right?
Q2: can the other one be fixed?
Hey, this looks great!
Let me look into my code, i´m sure JSFX needs a small update...
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 08-19-2017, 08:30 AM   #66
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by TBProAudio View Post
Hey, this looks great!
Let me look into my code, i´m sure JSFX needs a small update...
Thanks for the hint, there was indeed a bug in the LRA algo, fixed long time ago, but not published with "EBU R128 JSFX".
So please download V2.4 here: http://www.tb-software.com/TBProAudio/download.html
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 08-19-2017, 09:19 AM   #67
DarkStar
Human being with feelings
 
DarkStar's Avatar
 
Join Date: May 2006
Location: Surrey, UK
Posts: 19,677
Default

Thank you - getting closer, but still a Range difference (3.3 to 3.7):

__________________
DarkStar ... interesting, if true. . . . Inspired by ...
DarkStar is offline   Reply With Quote
Old 08-19-2017, 10:10 PM   #68
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by DarkStar View Post
Thank you - getting closer, but still a Range difference (3.3 to 3.7):
Hmm, cannot find the deviation here. Could you please pm me the test file, thanks.
BTW: maybe it helps if you enable "Sync" in dpMeter2 as well.
__________________
www.tbproaudio.de

Last edited by TBProAudio; 08-19-2017 at 10:17 PM.
TBProAudio is offline   Reply With Quote
Old 08-20-2017, 02:14 AM   #69
saastara
Human being with feelings
 
Join Date: Mar 2011
Posts: 64
Default

I'm not quite shure I've understood the automation writing.. The plugin writes volume automation so that the track level stays at the desired loudness level?
saastara is offline   Reply With Quote
Old 08-20-2017, 08:27 AM   #70
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by saastara View Post
I'm not quite shure I've understood the automation writing.. The plugin writes volume automation so that the track level stays at the desired loudness level?
It just writes the levels.

I you need constant levels use either Auto Gain Staging JSFX or AB_LM or Gainrider2
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 08-20-2017, 11:21 AM   #71
DarkStar
Human being with feelings
 
DarkStar's Avatar
 
Join Date: May 2006
Location: Surrey, UK
Posts: 19,677
Default

Quote:
Originally Posted by TBProAudio View Post
Hmm, cannot find the deviation here. Could you please pm me the test file, thanks.
BTW: maybe it helps if you enable "Sync" in dpMeter2 as well.
AH ha; the numbers are (almost) the same when Sync is enabled. What does that do?

Oddly, the values shown in Reaper envelopes were correct when writing them, but afterwards they were different for the dpMeter2 envelope. Have a look at:

>>> https://i.imgur.com/G5rYmU1.png

>>> https://i.imgur.com/dSXa0wC.png
__________________
DarkStar ... interesting, if true. . . . Inspired by ...
DarkStar is offline   Reply With Quote
Old 08-20-2017, 01:18 PM   #72
saastara
Human being with feelings
 
Join Date: Mar 2011
Posts: 64
Default

Quote:
Originally Posted by TBProAudio View Post
It just writes the levels.

I you need constant levels use either Auto Gain Staging JSFX or AB_LM or Gainrider2
Gain rider looks interresting..

As a little side track question: Why do so few gain riding and compressor plugins use rms instead of lufs / ebu r128 based calculations? In 'leveling' type compression scenarios, you'd think that lufs-based calculations would be an interresting option?
saastara is offline   Reply With Quote
Old 08-20-2017, 10:50 PM   #73
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by DarkStar View Post
AH ha; the numbers are (almost) the same when Sync is enabled. What does that do?
Sync just resets the measurement before start playing...


Quote:
Originally Posted by DarkStar View Post
Oddly, the values shown in Reaper envelopes were correct when writing them, but afterwards they were different for the dpMeter2 envelope. Have a look at:
dpMeter2 and EBU R128 JSFX use different min/max values for writing the automation values (see slider min/max value definition). The values themself are the same.

BTW: very nice usage of the EBU R128 library...
__________________
www.tbproaudio.de
TBProAudio is offline   Reply With Quote
Old 08-21-2017, 03:09 AM   #74
DarkStar
Human being with feelings
 
DarkStar's Avatar
 
Join Date: May 2006
Location: Surrey, UK
Posts: 19,677
Default

^^^
Yep, I understood about the different ranges used in the two plug-ins. But have another look at the envelopes for the dpmeter2 LRA parameter. When writing the automation the value was 20.0, but when reading back it was 17.1 at the same position.
__________________
DarkStar ... interesting, if true. . . . Inspired by ...
DarkStar is offline   Reply With Quote
Old 08-21-2017, 10:24 AM   #75
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by DarkStar View Post
^^^
Yep, I understood about the different ranges used in the two plug-ins. But have another look at the envelopes for the dpmeter2 LRA parameter. When writing the automation the value was 20.0, but when reading back it was 17.1 at the same position.
Well, it seems that there is something strange with Reaper automation: sometimes values under the cursor are not synced with TCP.
I will dig into it...
__________________
www.tbproaudio.de

Last edited by TBProAudio; 08-22-2017 at 03:57 AM.
TBProAudio is offline   Reply With Quote
Old 08-22-2017, 04:07 AM   #76
DarkStar
Human being with feelings
 
DarkStar's Avatar
 
Join Date: May 2006
Location: Surrey, UK
Posts: 19,677
Default GUI edition

Here we go, guys:



The attached zip file has TBProAudio's originals and my modified version. NB I have used the Verdana font, I don't know what will happen on a Mac Any GUI issues are down to me.

How to install a JS FX (in Post #4)
Attached Files
File Type: zip EBR128_LM_DS JSFX.zip (15.7 KB, 257 views)
__________________
DarkStar ... interesting, if true. . . . Inspired by ...

Last edited by DarkStar; 08-22-2017 at 08:24 AM.
DarkStar is offline   Reply With Quote
Old 08-22-2017, 06:53 AM   #77
TBProAudio
Human being with feelings
 
TBProAudio's Avatar
 
Join Date: May 2014
Location: Germany
Posts: 643
Default

Quote:
Originally Posted by DarkStar View Post
Here we go, guys:



The attached zip file has TBProAudio's originals and my modified version. NB I have used the Verdana font, I don't know what will happen on a Mac Any GUI issues are down to me.

How to install a JS FX (in Post #4)
This is very, very cool. Very good work, thank you for this!
__________________
www.tbproaudio.de

Last edited by TBProAudio; 08-26-2017 at 10:50 PM.
TBProAudio is offline   Reply With Quote
Old 08-22-2017, 08:15 AM   #78
DarkStar
Human being with feelings
 
DarkStar's Avatar
 
Join Date: May 2006
Location: Surrey, UK
Posts: 19,677
Default

Thank you

Slight buglet in the maximum Integrated value (and a couple of changes: cyan meters, button to bottom. Please download again (post #76).
__________________
DarkStar ... interesting, if true. . . . Inspired by ...

Last edited by DarkStar; 08-22-2017 at 08:25 AM.
DarkStar is offline   Reply With Quote
Old 08-22-2017, 08:35 AM   #79
bobobo
Human being with feelings
 
bobobo's Avatar
 
Join Date: Oct 2014
Posts: 1,356
Default

nice , grafic is ok
but the values are different between original and ds_version(click to view bigger image)


a demoproject
bobobo is offline   Reply With Quote
Old 08-22-2017, 08:41 AM   #80
DarkStar
Human being with feelings
 
DarkStar's Avatar
 
Join Date: May 2006
Location: Surrey, UK
Posts: 19,677
Default

The original has measurements expressed in LUFS, mine are in LU

LUFS +23 = LU
__________________
DarkStar ... interesting, if true. . . . Inspired by ...
DarkStar is offline   Reply With Quote
Reply

Thread Tools
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -7. The time now is 12:29 PM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.