How to find LAeq of a music mix?

I'm assuming you are talking about using a spectrum analyser for mastering ???? To be honest and this is only my opinion, I wouldn't bother using one, the way I see it is, how can it determine what you are hearing from inside the box, some engineers I know like to use them for live work in what ever kind of venue (stadium etc)

Dolby have released a VST plug-in that can measure dialnorm, but it's still expensive of course.
http://www.dolby.com/prof...olby-media-meter2.html
$795.00 list.

I have a theory on how this might be done cheaply, but I haven't had a Dolby product to test it against yet..

The dialnorm measurement is just the RMS average power of the dialogue.
It's expensive to measure in a finished mix, but if you're creating the material yourself you can measure the dialogue stem on it's own.

SoundForge (and presumably other software) has a 'normalise to RMS' function which can also measure RMS level with adjustable time constants.
It should be possible to measure the dialogue / vocal stem with that, or normalise the dialogue to whatever is required then leave that level alone and adjust the other stems around it.

Does anyone have access to DMM2 or a LM100 to help test that ?

Oh yeah, I forgot about A-weighting.

Does SPAN do that ?
Didn't look like it from a quick glance in the manual.

I'll need to find an EQ plug that can do A-weighting,
or a meter plug that can measure it.

Dave Allison


Oh yeah, I forgot about A-weighting.

Does SPAN do that ?
Didn't look like it from a quick glance in the manual.

I'll need to find an EQ plug that can do A-weighting,
or a meter plug that can measure it.

I think NUGEN's VISUALIZER does all that.   I have it, but I'm not at my DAW so I can't double check.  But Nugen's site probably has the info.
 
 

Thank you very much for the responses, very helpful. I will try out the suggested eq/SPAN method and see how I go.

Tarsier in particular, thank you once again for being so helpful as I try to find my way in the 5.1 world. Not sure if you read my other thread about problems I'm having with Surround Bridge? http://forum.cakewalk.com/tm.aspx?m=2079873

I was wondering if you had encountered this, and if you have a work around? I'm on the verge of buying Nuendo which is such a waste as I'm not going to use half of the features and I REALLY want to stay with Sonar. I just can't be limited to the Sonitus plugs - that reverb sounds awful! Although having said that, at the end of yesterday I was noticing some strange things happening with just Sonitus plugins inserted, so I'll have to check later on today.

Thanks again, I'll post back here when I've tried the EQ/SPAN method.

Here's an EQ that features A-weighted filters: Refined Audiometrics Laboratory PLParEQ. And it's only $1,000.00!

OK, here's one for a lot less (~$150): Electri-Q.

OK so I've tried to create the best A-weighted curve I could in Sonitus - is this OK? 
 


I then have Voxengo's SPAN afterwards on the vocal stem. 2 questions: 

-how should I interpret each channel's RMS reading to arrive at my value for dialnorm? Surround Bridge inserts 4 linked instances of SPAN, only allowing us to view the reading for each set of channels (L/R, C, LS/RS, LFE), so do I take an average?


-Is the RMS reading I get from this going to give me the 'correct' dialnorm number? 

Ran the test with a stereo mix.  Given the above Sonitus eq curve followed by SPAN in DBFS+3 mode (very important) you get a per-channel RMS reading that is about half a dB off from the Dolby LM100. I'd say that's close enough.

Dolby conforms to the AES-17 standard of RMS measurements which is why you need to set the DBFS+3 mode in SPAN.

I dug out the Dolby guidelines for setting dialnorm and on re-reading, they basically said every program is different. For music mixes close enough is good enough. They're pretty specific about getting dialnorm precise for movie dialog tracks, but for music they seemed to be less picky. Just ensure that your dialnorm setting doesn't result in unwanted compression or overs.

Thank you, that's cleared that up - 0.5dB off is very good. 

Like you, I was re-reading the Dolby docs and found them a tad ambiguous when it comes to music. Perhaps, like you suggest, we can conclude that dialnorm is not so critical when it comes to music. Although we definitely do have to get it close, as I've had some terrible results in the past.

So I take it that the A-weighted curve I made in Sonitus will suffice?

I'm very jealous of your LM100 (thank you for doing the comparison by the way - extremely helpful) - although it seems we now have a simple way of measuring Leq(A), it sure would be nice to have a decoder on board to hear the results of encoding. But for now, I'm happy without that luxury if it means hanging on to $4000!

Tarsier, you seem like you've been around a bit in the surround world, care to listen to a mix of mine and critique it?!! I'm very happy with how my mixes are sounding actually, but I'd love to get an opinion of someone more experienced, especially as all these 5.1 projects I'm doing are for commercial release and I'm a total newbie in 5.1. I've already emailed Al Schmitt for advice - I got a very nice email back from his secretary saying "Thank you for your kind words, but I don't think Mr. Schmitt does consultancy." :-) Was worth a try though!!