Following my last post on tips on using loudness metering in audio post with Pro Tools that was partially triggered by questions from Adam Taylor, Adam has responded with some more questions which he has kindly agreed I can post here so that his questions and my answrs will help others who are in the same boat. Adam asks…
“I did find it odd that the WLM has the various ITU / EBU standards as presets, but then you can choose from a myriad of variants, all of which seem to affect the metering in some way. Surely if its a set standard you should not be mucking about with any options?”
Yes there are two main standards, EBU R128 for Europe and ATSC A/85 for US. There are a number of very small variations in how to display these but basically these two standards are very close to each other. However the Waves meter is not clear in the way it configures its settings and it doesn’t appear to support the ATSC A/85 standard.
“In the end i settled on using the EBU meter mode, with no weighting.”
As the EBU standard is based on BS1770 which uses a K weighting filter, I would suggest selecting EBU and the ITU1770 weighting option in the Waves WLM Meter. Other loudness meters like the Nugen Audio VisLM are much clearer in the way you set and they display which standard.
“The most confusing area for me is getting my head around the shift from a peak meter mindset to an averaged level. I understand the principals of peaking….if its over - you loose!”
Yes, effectively with the loudness workflow, you no longer tied to what level the programme peaks to, (in reality the spec says around -2dBfs True Peak should be the limit, hence the need for True Peak meters and limiters). You are now tied to the Loudness, so as I say in the videos, we now normalise to loudness where as we have been normalising to peak level, and yes it takes some getting used to. The same rule about overs on Loudness is true but the Long Term (I) value as Waves calls it is the only number that counts, but the short term (S) and momentary (M) meters help in the mix process. Treat the Momentary meter like the loudness equivalent of a PPM, with the Short and Long Term giving you an average figure, Short over 10 secs, Long over the complete programme.
“Whereas the averaging seems vague at best. If my longterm average ends up reading -22 or -21 will my mix be rejected, even though the peaks are around +10dB.”
The EBU spec is plus or minus 1dB with the target of 23dBLUFS. Most broadcasters are talking about a 1dB tolerance for pre-recorded content and 2db tolerance for live programmes in their delivery specs. If you get a pre-recorded programme wrong, and once you get used to it isn’t too difficult to get it right, and meters with histograms like the Nugen Audio VisLM or the TC Electronic LM5/6 do make it easier.
“As the average is worked out over the whole programme, how do you reduce it by a couple of dB without lowering everything?”
Just reduce the gain of the finished file and use an offline version to re-measure the modified file. In fact there are a number of offline AudioSuite type plug-ins now, for example - from Grimm Audio (LevelOne) and Nugen Audio (LM-Correct) to help in this workflow.
“The mix I just did is probably not the best material to use as it was stems from the US version of an advert, with the levels blasting out (although balanced relative to each other). I usually do have to reduce these a lot, but found with the loudness meter the easiest way was to reduce every track by -18db… this gave an average reading of -25db on the WLM. So i pushed the voice element of the song to around -13db and the average is now coming out at 23. To get the 2db rise on the average required a 5db increase of one track over the 20s commercial. hmmm.”
Yes the correlation between audio level and loudness isn’t very tight, especially when the content is heavily compressed. As an experiment, try an uncompressed voice track peaking at PPM 6 and measure the loudness. Then compress it to taste and measure the loudness again with it still peaking at PPM 6, it will be much louder (no surprise), which is the expected outcome of audio compression. Its why we do it of course, we want our ad to be louder than anyone else’s as long as it peaks 6. Now with the loudness specs, that won’t be possible any more, and that is at the core of the loudness movement of course, and why the US has introduced The Calm Act which makes the ATSC A/85 spec legally binding after 30th Dec this year.
“Overall it does sound pretty good (and i always felt the vocals were to low anyway) but what does surprise me is how much quieter the overall mix seems (and looks to be 2-3db lower on the ppm) to ones done the “old” way. I have always been aware of over using compression and try to avoid it if possible, and try retain some dynamic range.”
Yes a compressed mix peaking at PPM6 will fail the loudness spec, usually it will come out around -20 or 21, whereas the spec is -23. But it actually allows us to put in dynamics like explosions at way above PPM 6 and so we can create impact that way. Before loudness everything had to be limited to PPM6.
“My only worry now is that as long as the broadcasters in the UK stick with Peak, will my mixes be quieter than the ones its shown against? And as you have mentioned people turn the volume down, then my mix runs the risk of being extra quiet!”
Because of what I have just said the transition will be a challenge because a conventional PPM6 prog with compression will fail the loudness spec whereas if we put dynamics into it for impact, it will fail the PPM6 spec.
I hope this dialog will be helpful to others. I should point out that we are refering to the Uk BBC type PPM which is offically a quasi Peak Programme Meter. There is a European equivalent with a different scale. PPM6 that we refer to is -10dBfs. But there isn’t really a US equivalent. But the UK and Europe work to a simlar standard of no peaks above -9 or -10dBfs read with a Quasi PPM before Loudness.