LKFS measurement of 5.1

I have a 5.1 PBS special which must conform to a loudness measurement using Bs.1770-3 LKFS metering.

Is it known whether WL9 is able to properly measure a 5.1 file faithfully to this standard?

Thanks,

I think so, if the following is correct, saying that LUFS and LKFS are equivalent:

But I don’t know for sure, and I’m sure somebody else knows better that I. (anybody else?)

If they are equivalent, the Wavelab Meta Normalizer should be able to create or finalize exactly to your target LUFS level in the Metanormalizer Montage Output or Master Section Output section, well within the client specified tolerances.

If you render a 5.1 multichannel file of your result, you can put it in the batch processor, with the monopass audio analyser plugin, file format - “no audio output”, and get a report of the integrated LUFS. It should be the same as the value you set in the Meta Normalizer. Of course you can analyze the stereo files with Global Analysis.

There’s always the loudness metering, but the analysis and auto-targeting tools are there too.

Everything I’ve seen says LKFS and LUFS are equivalent, so it should be fine.

Just wondered, what is their requirement? -23, -24 LKFS ? +/- 1 ? And a maximum True Peak?

btw, I’ve delivered 5.1 files that had to conform to R128 this way, that have passed without issue. Just wasn’t familiar with the term LKFS, sorry.

LKFS creation, analysis, metering should be fine in Wavelab, shouldn’t it PG?

Thanks for the info Bob99. The spec was -24LUKS +/- 2dB. TP -2dB max

This project is actually being created in ProTools 11. I bounced out an interleaved 5.1 bwf. When that file is then opened in WL9 it appears as only 2ch stereo, and then WL crashes! both Mac and Windows version crash. So I never even got to try and measure the audio cuz WL keeps coughing up a fur ball.

ProTools opens the 5.1 file just fine. Am I missing something?

I feel for you WYCA.

If you are under pressure, this meter may work for you in ProTools until you get WL sorted:

http://www.orban-europe.com/products/data/lmeter/supp_loudmeter_1.html

Manual: ftp://ftp.orban.com/Orban_Loudness_Meter/Documentation/FreeOrbanLoudnessMeter_2.9.0_Readme.pdf

I can’t comment on the WL issue.

Good luck!

Are you using file/open/audio file? Because that doesn’t work. (although it doesn’t crash if I try that).

If you haven’t done so already, make a Surround montage (there’s a surround template), set the montage sampling rate, and right click in a track and select Insert Surround Audio File at the bottom of the list. That will put the surround file on two stereo tracks (Lf Rf and Ls Rs) and two mono tracks (C and LFE). If that’s what you mean when you say it’s opening as Stereo, maybe it’s getting stuck and not completing all of the channel openings. Have you tried another file?

Alxo, did you try just Analysis in the Batch? You don’t need to “open” the file to do that. Make a new Batch Process, and drag the 5.1 file into the batch File Area. Add the monopass Analyzer plugin, set Output: No Audio Output, and run the batch to get the Integrated LUFS (LKFS) and TP values.

Both methods work fine with 5.1 interleaved files here, although I don’t have a Pro Tools 5.1 file handy to try. But I’ll try that later.

I’ve tried with Pro Tools 5.1 interleaved files, and it works fine here.

Thanks Bob99, those suggestions functioned for me.

I tried two methods.

  1. Imported files into montage and used the metanormalizer for measurements.

  2. Directly imported file into batch processor and used the audio analyzer plugin.

Both returned the same value for TP. However they disagreed on LUFS measurement, returning values which differed by 0.8dB.

So which (if any) of these tools is accurate, I wonder?

The batch processor analyzer readings run on the final delivery files are the closest they’re going to see on their end when they analyze your files.

You say you used the metanormalizer for measurements. I use the metanormalizer to create the final delivery files exactly at the LUFS target. If you render to a 5.1 multichannel file with the metanormalizer set with final LUFS exactly as you want in the delivery, and then analyze that rendered 5.1 multichannel file in the batch, what do you get? The difference (if at all) should be less than about 0.1db I think. There shouldn’t be any leveling going on after the metanormalizer LUFS setting you set, because that would make a difference. Like if there was a plugin or fader setting after the MN LUFS setting or even another MN setting after the LUFS that was affecting the LUFS. i.e. I don’t think there should be a 0.8db difference in what you set and rendered with the metanormalizer, and what the render then says in the batch analyze, in my experience.

Analysis I’ve been sent back, probably done with Dolby tools, has been within about 0.1 or 0.2 db for me as I recall.

I never use the metanormalizer to actually process files. I do that in the analog domain.

For the measurement, I set the metanormalizer to Reference: loudness of final mixsown. Check Shoew Log & Test Only. Hit apply and read the Log. It is there where I read a level which is 0.8dB higher than the batch audio analysis plug.

Your report about returned analysis being within ~0.1dB of the plug is reassuring. Appreciate your sharing.

So this leaves an unaddressed question. Why was I not able to open a 5.1 interleaved file in the audio file window? Is this not a supported operation?

OK, you’ve got me there. I never tried the Metanormalizer Test Only. In MN Master Section Output, you put in a LUFS target, and Loudness of Final Mixdown, and in the report you get Original Peak and LUFS, and what you would get if you rendered with your MN LUFS setting target. So comparing the “Originals” result in that report to the Batch Analyze results, I do see a 0.5db difference in the LUFS and about 2db difference in the Peak, at least on the Pro Tools file I’m trying.

I don’t know, I guess that’s a question for PG, about the difference. But I would go with the Batch analyze.

And yes, 5.1 interleaved file open in the audio file window is not supported yet afaik.

It looks like a repeatable problem of bussing when the 5.1 file is imported to the montage. My Ls Rs track always ends up bussed to Rf. However, if I buss it correctly to Ls Rs, the metanormalizer Test Only - Original LUFS reading is still not the same as the batch. It’s about 0.5db off the other way.

PG, can the import bussing be fixed at some point? It busses the multichannel files correctly when imported in Wavelab 8.5.

Also, do you know why the MetaNormalizer Original Test readings would be different from the Batch Analysis? So far I get about the same readings in both MN and Batch with Multichannel files made in Wavelab, but different readings (~0.5db) with Multichannel files made in Pro Tools. I guess that makes absolutely no sense at all, but if you have any idea why that might possibly happen… ?

If you think there is a difference between 8.5 and 9, please give me a short file so that I can reproduce. I don’t remember any change in this area.

Concerning batch / montage differences. I would need look, but I guess a different per-channel-type weighting factor is used.

Philippe

Thank you. That makes sense. And the WL/PT difference I saw is probably not because of the different software programs, but more likely because of the channel content of the files. I don’t have same program material for both. But making copies indicates no WL PT file difference.

I’ve made screenshots using two different surround files. The top red box in the pic is the meta normalizer analysis. The bottom red box is the batch analysis. With File 1, there’s a 0.4 dB difference in the loudness reading. With File 2, there’s a 1.0 dB difference.

You see the same sort of thing if you try the WAV test files on this page:
https://www2.iis.fraunhofer.de/AAC/multichannel.html

Doesn’t it seem like the weighting should be the same in both cases in Wavelab, whatever it is.

Also, the Peak readings in the meta normalizer analysis don’t say if they’re True Peak or Digital Peak.


Indeed, I should add if a peak is reported as true peak or digital peak. For the time being, if the box shows ‘ignore peaks’, understand ‘digital peaks’.

For the weighting issue, I will need to have a closer look, but I am too busy in other stuff to do it now.

Philippe

Thanks PG, and thanks for putting these features in there. I think it’s pretty great you can get a complete analysis of a 5.1 file in no time.

WYCA, I’ve found two free surround loudness meters that are probably a good crosscheck. TBProaudio DPMeter and Melda MLoudness. In Reaper you can render thru them pretty quickly and the overall results are displayed. I couldn’t get results on render to work with them in Wavelab. Don’t know why because it seems like it should work. But you can play the whole program thru them in Wavelab to get the results. I don’t know how they would work in Pro Tools, but you might want to try them out as a crosscheck. The Melda takes a lot of configuring settings each time to get surround metering on it in Wavelab, but that could be just me.

Just testing one surround file they were both closer to the Wavelab Metanormalizer analysis than the Wavelab Batch analysis, but I would have to do some more with more files before saying which weighting is probably more common.

Haven’t tried the Dolby, Waves, or Nugen meters yet.