1

Topic: DIGICheck ITU

Hello,

after getting lots of requests about the latest ITU standard of weighting/measuring levels, we made a first test version of DIGICheck that has ITU weighting available. If you are interested please download:

https://archiv.rme-audio.de/download/digicheck_itu.zip

and copy the extracted digicheck.exe into your existing DIGICheck program folder, overwriting the current DC 5.23 file.

In this version ITU weighting is only available in the Surround Audio Scope. Using F 2 you will find the new option 'Weighting', which can be set to RLB (the old option) and ITU, according to ITU-R BS.1770-1.

When selecting ITU, the level bars show RMS with ITU weighting. Additionally the sum of all active channels will be displayed on a separate level bar (labeled ITU). This one works similar to RMS slow, so it is easy to read the 'loudness' value.

Remember that you can use this meter with only 2 channels as well, simply deactivate all others (F3). The Surround display can not be deactivated, but made smaller by dragging the window size horizontally.

If we get some useful feedback on this, we will add the new weighting to 2-bar, 4-bar, Multi and Global level meter as well.

Regards
Matthias Carstens
RME

Re: DIGICheck ITU

GREAT!

I'll test at once. Been waiting for this.

P

Re: DIGICheck ITU

Hi,
I'm a TV mixer in Los Angeles. I've been using digicheck as my main metering system for years. Love it. Thank you for this update. ITU is becoming a new standard for TV loudness in America. RME is always ahead of the technology curve. Quick question. I've been using the RMS slow un-waited for Leqa loudness requirements and it's worked great. Is this ITU just a eq curve? I've mixed a show with the new meter and like it so far.
Thanks,
Brandon

4

Re: DIGICheck ITU

RMS slow plus RLB is nearly the same (at least in the end result) as ITU. Technically it's quite different, especially the weighting frequency response. The biggest difference is when you calculate the sum of multiple channels (surround), which adds special factors for summing the channels into one single ITU loudness value.

Regards
Matthias Carstens
RME

Re: DIGICheck ITU

Excellent news.
Just found this thread via the Nuendo forums, and have DL this.
Will happily report back - top one, guys.
Many thanks indeed

Re: DIGICheck ITU

Works great, thank you RME!

Regards,

    b.

https://granurise.com

Re: DIGICheck ITU

Quick (silly?) question.
When using the ITU setting, should the +3dB RMS switch also be used, as well as K-System - or are these meaningless in these circumstances please?

8

Re: DIGICheck ITU

Both simply shift the scale to your liking, so they have no meaning regarding the measurement of the value itself. ITU calls for +3dB active though.

Regards
Matthias Carstens
RME

Re: DIGICheck ITU

any ITU weighting planned for Mac users?

10

Re: DIGICheck ITU

Sure. The final version will have the same weighting options on both platforms.

Regards
Matthias Carstens
RME

Re: DIGICheck ITU

MC wrote:

Sure. The final version will have the same weighting options on both platforms.

great !!!!

Re: DIGICheck ITU

Hi,
I don't mean to contradict Matthias but the ITU BS.1770 standard specifies that a 1kHz sine wave reads 3.01dB lower than the peak level. In other words the +3dB option in DigiCheck must be turned OFF. With the appropriate channel weightings you should get the following:

1kHz tone @ -20 dBFS 1 channel (L or R channel): -23.0 LKFS
1kHz tone @ -20 dBFS 2 channels (L + R): -20.0 LKFS
1kHz tone @ -20 dBFS 5 channels (LRCLsRs): -15.4 LKFS

Also note, the Digicheck application shows what could be considered as short-term levels, much like a VU or a PPM meter, though with the correct frequency and channel weightings. BS.1770 describes a cumulative measure. As an example, the loudness of a 30 second commercial would be the average level for the entire 30 second duration. Currently, there are no recognized ballistics for a short-term loudness meter.

It may be useful to some, but the ATSC has published a document for measuring loudness for television. The document, A/85, is available here:  http://www.atsc.org/standards/practices.php 

Cheers,
M

13

Re: DIGICheck ITU

Michel wrote:

Hi,
I don't mean to contradict Matthias but the ITU BS.1770 standard specifies that a 1kHz sine wave reads 3.01dB lower than the peak level. In other words the +3dB option in DigiCheck must be turned OFF.

Thanks, you are right. It also means that we should disable (grey out) the +3 dB option when ITU weighting is selected.

Michel wrote:

Also note, the Digicheck application shows what could be considered as short-term levels, much like a VU or a PPM meter, though with the correct frequency and channel weightings.

Correct. IMHO that is what 99% of the users want to see and use.

Michel wrote:

Currently, there are no recognized ballistics for a short-term loudness meter.

Which is really bad and seemed to have caused headaches for all companies offering this new metering.

Michel wrote:

BS.1770 describes a cumulative measure. As an example, the loudness of a 30 second commercial would be the average level for the entire 30 second duration.

I can not read that kind of approach or function out of the existing document. Also I am sceptical on the usefullness of such an application. Over such a length of time the 'Loudness value' has no real-world meaning. For example my commercial could be totally wrong and absurd (wrongly mixed, too low volumes at some parts) within the radio program, but still show the correct 'value'.

Please don't get me wrong, this is not against you, I just would like to have some discussion and more background talks about this new metering, so don't hold back to comment on the above.

Regards
Matthias Carstens
RME

Re: DIGICheck ITU

Hello,

I just found this interesting thread.

Thanks RME for following the current reorganization of level and metering normalization.

HDTV did introduce new practice and standarts. In France here are the latest norms:

http://www.afsi.eu/news/nouvelle-recomm … la-recomma
summarize 2 important papers explaining or defining new norms for french TV PAD:
http://www.ficam.fr/upload/documents/RT … _V2-21.pdf
http://www.cst.fr/IMG/pdf/Sensibilisati … 019_-1.pdf
in english there is only an older, no more valuable  one: http://www.cst.fr/IMG/pdf/RT_017_-_TV_- … glaise.pdf

it seems both long term and short term LKFS measurements are needed, for long programs (such as a movie for instance), long term measurement should be done on dialog only. see http://www.cst.fr/IMG/pdf/RT_016_-_TV_- … vel_V2.pdf
Dolby LM100 "Dialog intelligence" can recognize dialog from non dialogued parts, and do the according measure...

10 seconds seems to be the choice of "Short term" LKFS measurements. the meter we should be looking at...

So it would seem to me that Digicheck should include a 10 second "short term" LKFS measurement. and maybe a choice of other times as well
(see for instance http://sourceforge.net/projects/baptools/)

For short programs long term measurement via digicheck might also be interresting, but would be only to long program if "dialog intelligent", which might be more complicated...


It seems EBU will soon issue very similar recommendations. see the group P/loud



I still myself is trying to understand all that, so I don't want yet but to give this crude summary, in order to avoid any misconprehension I could have...

Anyway, it would be great if Digicheck could be compliant to thoses new norms...


Best regards

Laurent.

15 (edited by lgabiot 2010-04-05 10:32:19)

Re: DIGICheck ITU

Just to summarize the papers I mentionned in my previous post:

According to the new norms, we need the following meters to work:

A din45406 (10ms). Stereo program material must not exceed 0 PPM.
An (ideally) oversampled FS meter (5.1 material must not exceed -3FS, for downsampling compatibily reasons).

alignement: 0 dB VU = -18 dB FS = -9 dB PPM DIN45406

A loundness meter:

that comply to ITU ? R BS.1771 and ITU ? R BS.1770.1

units of measurement are in dB LKFS (Loudness K weighted relatively to a Full Scale) so on a Full Scale we have a loudness measure K weighed (that is R2LB). the scale must be in 1 dB increments.

The loudness must be measured in 2 ways: Short Term and Long Term
Long Term measurment is to be performed on the entire program for program during up to 15 seconds (for instance commercials). On longer programms the measure must be made on the dialog only for as much as possible of the program... That is called the Dialog Level, and will be used for Dolby E Metadata.

The Target is at -25 dB LKFS (+/- 2 dB).


Short Term means that Loudness should be measured in a 10 second window.

norm allow a variation +/- 7 dB LKFS for Dialog on short Term, and +/- 12 dB LKFS for complete mix around the -25 dB LKFS target

So we need 2 modes: Long term ans Short term (10 sec).

Scale is Full scale, but visually centered around -25 LKFS, the value -13 LKFS being a max, -18 LKFS max for dialog, etc... (deduct from the +/- tolerances).


So not speaking of phase issues, here is the description of the meters we need now...

Re: DIGICheck ITU

ITU works great. Now can we have it in OSX?

Have a nice week end

17 (edited by Dandruff 2010-11-11 14:06:40)

Re: DIGICheck ITU

Matthias,


I always have "Always on top" enabled but sometimes on DIGICheck startup it doesn't stay on top although the option is enabled. Close and reopening DIGICheck doesn't help. I have to disable and enable "Always on top" to get it fixed ...

Re: DIGICheck ITU

Matthias,

when will we get the ITU weighting in the other meters?

smile

19

Re: DIGICheck ITU

Next release, I think. First we bring the Mac version to a similar level.

Regards
Matthias Carstens
RME

Re: DIGICheck ITU

Ok, cool!

Re: DIGICheck ITU

I discover the Digicheck 5.3.
Thank for Smpte channel order (L/R/C/LFE/Ls/Rs), correlation inter channels, LFE display and ITU weighting !!!

About this Loudness measure, i used it a lot, those last years, for dialog level estimation with Dolby products : LM100, DP570, Dolby Media Meter (recent plug Loudness meter + logs, graph).

Key point of the loudness measure, is ITU load ok, but,another key point is that this measure need differents averages time possibilities.

Loudness measure is now being included in spec about 5.1 programs for HD TV but also soon for Stereo pgms.(as very well describe up )

Solution choosed by broadcasters is to fix a Loudness Dialog Level, all programs in the future would have quite the same dailog level (but could have gap of dynamic around this) .
The result is that people at home would get better matching listening levels, when changing channels in their hundreds tv channels(DSL, Satelite, Cable..).

Peak levels was the only parameter checked first, in France -9 dBFs was the max peak accepted for Stereo pgms.
Now, dialog loudnesss is checked also and must be - 25 LKFs(-24 LKFs, -23 LKFs EBU P/Loud group), peak for 5.1 musn't exceed -3dBFs in France. (EBU P/Loud group -1dBFs Max)

So french mixers have to work at -25 LKFs, this value -25 (or else), must be an average of all (and only)the dialogs parts on all the program and summing all channels except LFE (at this time). This must be done at the end of the mix, Dialog detection is needed, and log is apreciate.
If dialog auto detection is not possible, we have to make it ourself, summing the averages of differents dialog parts, this with an "Infinite or Long Term Mode" associated with pause button to "pause" measure when no dialog is present, the system make an average of all the parts measured until u reset it with "reset button" (needed also) and a counter gives u the among of time measured .


The second average time most used  (not yet normalised as i know by ITU SMPTE, EBU... but choosed in french specs) is called Short Term Mode and have a 10 seconds average window (EBU P/Loud group talk about 10 or 8 seconds). This one is the most important at this time for TV, DVD,mixers. During the mix when dialog present, we have to be near -25 on loudness meter, and on loud or soft dialog scenes we can reach :-25 +/- 7dB.
Then with this Short Term mode we can checked the "loudness peaks" of the rest of the mix with -25 +/- 12 dB LKFs allowed.

KEY point here : The TV broadcasters will verify those gaps: with Short term 10sec !!
                       And dialog level with infinite Mode and Dialog Detection auto.
So if u want to be shure, ur production will be accepted by the broadcaster QC control, u need to have the same way of meetering.

Working in Short term mode (10sec) during the mix,checking ur -25 dialog level, matches most of the time with an entire final analyse with Dialog auto detection, the gap allowed is +/-2dB at this point (EBU P/Loud group talk about +/-1dB).

This Short Term Mode would be very usefull in DigiCheck with 10 sec window, and if u can :
- Differents gaps for this window (8, 10...etc infinte)
- A special window with only ITU measure (to have only this process runing and save screen place) shown as meter or as Value Window (3 or 4 cm) with Big    Numbers so we can put in a corner to get an eye on Loudness.


Last mode is the shortest one, not present in Dolby Meters, it varies depending on manufacturers,(EBU P/Loud group call this mode: Momentary and talk about 100 to 500 ms, ). This is not normalised yet as i know.


After this short introduction (Sorry) some questions :

- What is the actual momentanary time average of ur ITU loudness meter ?
- Do u think possible to make it work, with 10s average Short term Mode or more ?
- May be get logs and graphics ?

- I noticed a difference on ITU metering between front and rear channels:
     Test: Send same sample sound or pink noise, on only one at a time of 5.1 channels Surround audioscope ITU.
           You'll notice that measure ITU is about 1.3 or 1.4 dB higher when sound is in one off rear channels than in one of front channels.


Anyway this is already great, even if it stays like that i'm happy with for semi pro use at home ITU metering !
With Some modifications it could be used in Pro environement.

Very Good Job Guys !!

22

Re: DIGICheck ITU

Thanks for all the info and explanations. Very much appreciated! The single ITU window is indeed planned, and the 10 s mode should be possible with it too (don't expect this one quick, there are other things to do first). Some other of your sugestions will never make it into DC, simply because we don't want to spend lots of time with adding features that only a handful of users would ever use.

The rear channels must be more sensitive by about 1.41 dB as covered in ITU 1770.

Regards
Matthias Carstens
RME

Re: DIGICheck ITU

EBU technical review 2010 Q3:

http://tech.ebu.ch/webdav/site/tech/sha … amerer.pdf

But the most essential docs: EBU Tech Doc 3341 and 3342 are published, Tech Doc 3343/44 "Practical Guidelines..." are 'awaited  publication'.

br
Gerhard

Re: DIGICheck ITU

Great thread, many interesting things & comments too.
I am particularly looking forward to the further development of DC.

Re: DIGICheck ITU

Webinar by Florian Camerer, Chairman of the EBU PLOUD group about R.128 (60 minutes)

http://tech.ebu.ch/Jahia/site/tech/cach … f?id=13386

Re: DIGICheck ITU

Thanks for that - will sit down & view both parts later tonight.

Re: DIGICheck ITU

This thread is really really interesting and I'd love to use Digicheck and nothing else for metering. I'm in the sound for film and TV bussiness and there's so much metering to be aware of so if some of these featires would be possible I'd be more than happy.

Another thing! Do we have the same features in Digicheck on both Mac and PC these days?

Best regards / Mike