1

Topic: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Hello,

a new version of DIGICheck is available. It includes a new function called ITU 1770/EBU R128. The implemented functionality is straight from the book and comes with no surprises for professional users.

For the normal user this meter is more or less useless. It is tailored for TV/broadcast. Whatever 'normal' signal from your daily work you feed it with, you will just see a fullscale level most of the time ;-)

Maybe you will find the included measurement of the 'Loudness Range' useful. It is supposed to show how much 'dynamic range' is left in your mastered recording - a means to fight against the so called Loudness War, but unfortunately not always working correctly (which might indeed be nearly impossible). But if used and read in a sensible way it will give you a good idea and especially a standardized comparison to others.

Pro users please provide feedback in this thread!

https://archiv.rme-audio.de/download/dc_54_win.zip

https://archiv.rme-audio.de/download/dc_065_mac.zip

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Thanks,

works well with->FFUC-W7 home 64bit.

Regards

FF UCX - ADI 8 QS(DOTec Madi/Usb i64) - Madi HDSPe(Tco 4)

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Thanks MC!

Works great on WinXP SP3 32bit.

Could we also get temporarily "Time Total" minus "Time Gated" display? For example if you click on "Time Gated" field and hold your mouse click down... you temporarily get the total time the gate was closed (or a settings option or something like that).

BTW: If I press F2 the title says "Surround Audio Scope Setup".

Cheers!

    b.

https://granurise.com

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

No oversampling (OVS)?
"Weighting Filter" also missing.
Would be nice to have the ITU weighting for all meters too (not just the Surround Audio Scope). A bit more consistency would be nice.

5

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Sorry, but you just disqualified yourself ;-)

As I wrote: The implemented functionality is straight from the book...

OVS is part of the spec, it is always on and can not be switched off (you did not see the +3 dBFS scale?).
ITU Weighting is part of the spec and therefore can not be disabled for the right three meters. It could be for the left, but that is something for later or never, depends on the feedback that we get.
ITU Weighting for other meters is on our list, but no date.

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

MC wrote:

Sorry, but you just disqualified yourself ;-)

As I wrote: The implemented functionality is straight from the book...

OVS is part of the spec, it is always on and can not be switched off (you did not see the +3 dBFS scale?).

Ok. My fault. Just was looking for the button.


MC wrote:

ITU Weighting is part of the spec and therefore can not be disabled for the right three meters. It could be for the left

That's what I meant.


MC wrote:

ITU Weighting for other meters is on our list, but no date.

Ok.

7 (edited by georgob 2012-03-14 15:08:39)

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

I would like Digicheck to display when the gate is active.

Additionally I made some tests comparing measuring in Sequoia and with Digicheck:


Sequoia
INT -4,8
True peak -10,2
LRA 22,0
INT time 2:56


Digicheck:
INT -5,5
True peak -10,2
LRA 18,7
INT time 2:56

Which one should I believe?

Georg

8

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

DIGICheck.

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Cool smile

10 (edited by georgob 2012-03-14 16:47:49)

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

I did some more tests with the EBU testsignal Nr8 of Document 3341. The test tones were all the same in Sequoia and Digicheck, but not the real life signals (the INT should be 0 +-0,1).
Especially the handling of the gate seems to be different.

Sequoia
INT 0,3
LRA 14,8

Digicheck
INT 1,6
LRA 4,5

Georg

11

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Are you talking about seq-3341-2011-8_seq-3342-6-24bit-v02.wav? That one gives me INT -0.1 (and LR 14.1).

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Exactly this one. I will check it tomorrow again.

Georg

13

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Make sure you have the latest v3 test set downloaded (the above file is changed to -10 LUFS relative gate, although still named v02).

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Yes I have.
http://tech.ebu.ch/webdav/site/tech/sha … setv03.zip

I use Digicheck 5.40Beta2.

Georg

15

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Me too. What hardware? DC in ASIO Playback mode? Which player is used? Mixer settings all unity gain?

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

I apologize. Obviously I made some mistake yesterday, I could not reproduce my results today.

I did more tests today. I use Sequoia as playback, and a Madiface.
Playback of the mentioned file gave these results: INT-0,1 and LRA 14,2
Playback from another computer and input metering in Digicheck: INT-0,0 and LRA 14,1
Still there is a small discrepancy to Sequoia: INT 0,3 and LRA 14,1

Sorry again.
Georg

17

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

In that case you have to call Magix. The files allow +-0,1 dB, DIGICheck matches that...

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

I think the main difference between Sequoia's and Digichecks metering is in the handling of the gate. Sequoia's gate gets active when the current level is 10dB below the current INT (relative measuring). This threshold can change during time of the measurement.

Digicheck's gate gets active when the absolute level is below -70dB (according to the display "time gated").

Which system is correct? According to Florian Camerer (in the latest VDT magazine S.12) there should be a gate with a relative threshold of -10LU.

Georg

19

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Not quite sure what you are hunting for. We are not here to explain the standard or our implementation of it. The gate time that we (!) display is the absolute gate time, because a display of the relative gate time makes no real sense. The fact remains that DIGICheck shows the correct value, Sequoia does not. So please stop questioning the validity of our implementation and turn to Magix for further inquiries.

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Just wanted to say that foobar2000 1.1.1 with -5dB ReplayGain preamp setting gives same EBU R128 results as DIGICheck! That's what's used in foobar2000: https://github.com/jiixyj/libebur128

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

I would like to forward some tests to you made by Kai Schluenz in the Sequoia Beta forum. The shown results are INT.
Georg

Here are Kais tests:

I used the following test signals:
- ebu-loudness-test-setv03
- "speech sample" from ATSC Audio Level Calibration (a_85-2011a.pdf), at -2.4dB relative:
http://www.atsc.org/...eech_sample.wav
- Eros Ramazotti "Foucco nel foucco", min. 1:44-2.54 at -15,3dB rel.
- Carmen Exerpts "2000 Digital Remastered" Prelude, min. 1.04-2:03 at -9.8dB rel.
- Flieg mit mir (unmastered german title, strong bass) min 0:00-1:07 at -6,9dB rel.
- "LFPinkNoise_-20dB" from ATSC Audio Level Calibration (a_85-2011a.pdf), at +5,1dB rel.:
http://www.atsc.org/...Noise_-20dB.wav
- Sine waves 20Hz-20kHz generated in SEQ at -23dBfs.


The tested devices:
RME DIGIcheck 5.40b2
Audication free AC-R128 Loudness Metering Plug-In EBU R128 1.0
Sequoia 12.0.1.110 R128

Results
RME Au. SEQ Signal
-0.1 +0.0 +0.0 EBU #1 Sine wave -23dB
-0.1 +0.0 +0.0 EBU #4
-0.1 +0.0 +0.0 EBU #5
-0.1 +0.0 +0.1 EBU #7 short audio sequence
-0.1 +0.8 +0.3 EBU #8 long dyn. audio sequence
-0.3 +0.2 -0.1 Speech_sample
-0.1 +0.0 +0.0 Eros Ramazottti
-0.8 +0.0 -0.5 Carmen
-0.1 +0.0 +0.0 Flieg mit mir
-0.1 +0.0 -0.7 LFPinkNoise
-14.1 -13.0 -19.9 20Hz
-6.4 -6.3 -8.5 40Hz
-2.0 -1.8 -1.2 100Hz
-1.1 -1.0 -0.7 200Hz
-0.8 -0.7 -0.6 440Hz
-0.1 +0.0 +0.0 1kHz
+2.2 +2.3 +2.4 2kHz
+3.1 +3.3 +3.3 4kHz
+3.2 +3.3 +3.3 8kHz-20kHz


So, as already mentioned, Sequoiatude uses a different high pass filter.
As far as I can see from the EBU Tech 3343-2011v2 Practical guidelines page 9, the Magix version is right.
There are older versions, e.g. in the i.t.u.-requirements from 2006 proposing a flatter filter.

One more thing a noticed: the RME gate reaction is slow, after the end of a file (e.g. sine wave) the LU level often drops by 0.1dB

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Help/Search doesn't open the "Search" tab in the help file. It opens the "Inhalt" tab instead.

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Dear Mattias:

With a little work we could get rid of the K-system in the Digicheck!

I haven't downloaded the beta because I think from your reports so far that this implementation is bare bones "by the book".

One thing which exemplifies the new ITU approach:

Hide the peak reading. Peak readings only encourage people to peak normalize. So, a good ITU/EBU meter should only give a peak warning, and only if the true peak level reaches the -1 dBTP limit. Otherwise, don't show the peak reading. That's my first advice.


The other thing which would let us get rid of the k-system is to make a K-system for the modern world:

Keep the R2LB weighting filter.  Hide the peak unless it exceeds the threshold. And give us a variable 0 LU point (which if course "violates" R128) but gives the user flexibility until the music world comes around to EBU-land.


Hope this helps, please send me a note at bobkatz[atsign]digido.com if you implement any of this, I'll be the first to put it into my work.


Best wishes,



Bob Katz

24

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

The final release version of DIGICheck 5.4 and 0.65 is now available for Win and Mac. Compared to the Win betas these are the latest changes:

- Updated included help

- EBU R128: ITU Weighting (the formerly called R2LB filter) can be switched off for the left (non EBU R128) level meters

- Setup window title fixed

- Note: you can select +9 or +18 scale for the right meters. But you can also deactivate both options. Then the scale of the left meters is taken over. While this violates EBU R128 it gives a nice comparison option.

Download link in first post. Thanks for testing!

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

https://archiv.rme-audio.de/download/dc_54_win.zip

Official download site isn't updated yet.

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Hi Mattias,
this is a great add on to Digicheck. 
Im using it for a TV film mix and when I'm done I'll compare results with some other meters like Pinguin and Grimm.

I have a question however.
Does Digicheck 5.4 work with the old Digi 96 card (driver 2.11, Windows XP, SP3)?
The last Digicheck that worked fine with this card was 4.5. The Digicheck 5.4 wrongly identifies this card as HDSP MADI.
FF400, Digiface and 9632 they all work fine with 5.4 on W7.
Is there anything I can do to make it work?

thanks,
Peter Gruden

27

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Sorry, no.

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Ok, thanks.

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

I just made some real world comparision between 4 different meters: Digicheck, Grimm, Nugen and ToneBoosters,
on a 12 minute long section from a film mix.
The results of all four meters were identical for parameter I (-1.0 LU in this case).

There were some slight differences for LRA:
Digicheck 13.8
Grimm 14.8
ToneBoosters 14.6
Nugen 14.6

I also measured the whole film which has some silent areas as it is unfinished yet and the integrated loudness was still identical, while the LRA parameters were a bit further apart.

I find Digicheck easy to follow and the resizable window is nice, but TP display would make it even more usefull.

30

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Dc 5.41 and 0.65 are now available on our website. 5.41 is identical to 5.4 except for the version number. This was necessary for the installer to overwrite the beta version.

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

nedur wrote:

There were some slight differences for LRA:
Digicheck 13.8
Grimm 14.8
ToneBoosters 14.6
Nugen 14.6

I also measured the whole film which has some silent areas as it is unfinished yet and the integrated loudness was still identical, while the LRA parameters were a bit further apart.

It seems that the standardisation of the LRA, which is given by the EBU is not precise enough. There seems to be too much tolerance in how to implement the LRA measuring.
Apparently the exact amount of overlap between the measured blocks is implementation-dependent. This can give very different results for the LRA (I had differences up to 4 dB when measuring the LRA with Sequoia compared to Digicheck).

Georg

32

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

I'm wondering whether there are any plans to implement alternate schemes, such as ATSC A/85, for those of us here in North America?

Frank Lockwood
https://LockwoodARS.com
Fireface 800, Firmware 2.77
Drivers: Win10, 3.125; Mac, 3.36

33

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Reading the docs it seems they are identical. You have a source that lists the differences? That would be helpful.

Regards
Matthias Carstens
RME

34

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

nedur wrote:

I just made some real world comparision between 4 different meters: Digicheck, Grimm, Nugen and ToneBoosters,
on a 12 minute long section from a film mix.
The results of all four meters were identical for parameter I (-1.0 LU in this case).

Which is good!

nedur wrote:

There were some slight differences for LRA:
Digicheck 13.8
Grimm 14.8
ToneBoosters 14.6
Nugen 14.6

I also measured the whole film which has some silent areas as it is unfinished yet and the integrated loudness was still identical, while the LRA parameters were a bit further apart.

We had a look at the LRA implementation again, and it seems that the standard test files that we used will not reveal small implementation errors, only when doing what you did it will show up. A new version of DC with a corrected LRA measurement is available here:

https://archiv.rme-audio.de/download/dc_543_win_exe.zip

Would you be so kind to test it again against the others?

Regards
Matthias Carstens
RME

35

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

MC wrote:

Reading the docs it seems they are identical. You have a source that lists the differences? That would be helpful.

While based on ITU-R BS.1770, ATSC A/85 uses a Target Loudness level of -24 LKFS, + or - 2 LU as its target, instead of -23. Note that the Target Loudness is shows on the meter as -24 LKFS, not 0. Absolute peaks should not exceed -2 LU True Peak.

Apart from the different Target level, the main difference between the Recommended Practice in Europe and that for North America has to do with the intent and goal of the regulation. The EBU wishes to impose Loudness regulation to all forms of media, whereas the C.A.L.M. Act in the USA (and hence, the rest of North America) deals specifically with broadcast, in order to deal with level disparity between program material and commercials.

This main manifestation of this disparity results in the adoption of gating - whether or not low level signal will have an effect on a Loudness Meter's reading. The EBU has adopted the provisions of ITU BS.1770-2, which recommends that gating should be employed all the time. The C.A.L.M. Act refers to BS.1770-1, where the focus is on the "Anchor Element" - the main dialog, the lead vocal, essentially the most important and prominent element in a sound mix. Under the terms of the Act, Loudness measurement should only take place on the sections of the program where the Anchor Element is present. The rest of the time, gating can be used, but its use and its threshold are left to the discretion of the individual broadcaster. Most people I have spoken with, involved in the film and television broadcast industry, are not using gating at all, preferring to measure loudness over the entire duration of a given program's material. To say that this state of affairs is ambiguous is no understatement. It would seem prudent to produce a Loudness Meter for the North American market which has the level and time constants defined in ATSC A/85, with the option to engage or not engage gating, with a user definable threshold, but since the legislation seems to be in flux, it's understandable why nobody has rushed such a design to market.

The official ATSC A/85 document, with details of measurement spectral pre-emphasis and time constants can be found here:
http://www.atsc.org/cms/standards/a_85-2011a.pdf

A very welcome document purporting to provide "Loudness Standards Disambiguation", written by Valentin De?caillet of Merging Technologies Inc., can be found here:
http://www.merging.com/uploads/assets/M … uation.pdf

Frank Lockwood
https://LockwoodARS.com
Fireface 800, Firmware 2.77
Drivers: Win10, 3.125; Mac, 3.36

36

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Frank, thanks a lot for this roundup!

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Mattias,
With 543, the results of all four meters are quite close, within +/-0.1 LU for both I and LRA.

I have measured the whole lenght of a TV film (52 minutes), which I think is a good representation of european style of production, with modern music, moderate effects, dialog inside and outside and a lot of ambiences, mostly in town.

I captured 4 times across the lenght of the film because in the first 10 or so minutes Nugen was off by 0.5 LU approximately, but recovered later.
I'll send you these captured images of all 4 meters to your personal mail.

Meters reacted to loudness changes within several seconds, Grimm being the slowest I think. At one time, two meters could show exactly the same value, on other occasion other two or three meters were in unison. Most of the time, Digicheck and ToneBoosters were the same, while the other two were at +/- 0.1 LU, which is quite good.

In my opinion, Digicheck has the easiest display to follow and nice large buttons. ToneBoosters has the most complete set of parameters, and Nugen has the nicest and most flexible graphical display. Grimm has the best normalizer.
Currently, I find that Digicheck with ToneBoosters is an excellent combination.
I wish Digicheck could run on the audio card processor, to accept external sources,  and add a decent TP display.
And accept start command from the host application play button.

Peter Gruden

38

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Thanks a lot for the helpful testing and comparison! Version 5.43 should be available this evening on our website.

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Is it possible to set it up so that measurement starts when the DAW starts, in other words - when DigiCheck starts getting a signal?
As it seems to stand right now, I need to start playback, and scurry across to DigiCheck and press START - whereas in PPMulator 3 I can lock it down to my host. Okay, that's a plugin & DigiCheck is not - understood, and this might therefore be a problem.
It would still be very nice to have the option though.

In all, very nice indeed and complements the other EBU-R128 plugins I use (TB-EBU, PPMulator and Steinberg SLM128)
My cup doth truly runneth over etc....

40

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Silence at the beginning is ignored for the measurement (as per the standard).

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Great move, thanks guys!
Now if that view could be in one window with the vector audio scope....

Thanks!
Oliver

42

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Put the windows wherever you want them, then save the workspace...

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

The loudness metering according to the K-system is very helpful here. When recognizing there is already enough loudness, I can concentrate on variations in gain staging. Nice, when vocals breathe again and instruments show a little more expressiveness.

So, thank you very much for this free addition.

44 (edited by jonasgruska 2013-12-16 22:25:34)

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Hello,

I am having an issue with this meter. All other modes work fine, but in this one, DigiCheck becomes very unresponsive and starts to consume over 50% of a CPU.
My computer is Macbook Pro Mid-2010, with latest system update (10.9) and with latest version of DigiCheck (0.671).
My soundcard is Fireface UC with latest firmware and driver.
Can you please look into that?

Regards,
Jonas

45

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Which hardware is used at what sample rate when this error shows up?

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

MC wrote:

Which hardware is used at what sample rate when this error shows up?

I tried various sample rates, and discovered it only occurs before the first change of the sample rate. So - I start the DigiCheck, sample rate is 44100 - it is behaving slowly and CPU usage goes to 80%. When I change to 48000 - it starts to behave normal and when i go back to 44100, it continues on working without problems. So, now I know the workaround, but it would be nice to not have it there at all.
My hardware is Fireface UC.

47

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

I can not reproduce, which is no big surprise. To see this effect it might be necessary to have the exact computer model, or at least the exact setup. Can you please send me a DIGICheck workspace file?

Regards
Matthias Carstens
RME

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

Sorry about the delay, I didn't get the notification (or missed it somehow) about the reply. Here is the workspace file: http://cl.ly/TYfA

49

Re: New DIGICheck Mac/Windows version with ITU 1770/EBU R128 support

I was finally able to reproduce strange CPU load problems with the ITU meter. We will check.

Regards
Matthias Carstens
RME