Defeat the Loudness Deception ! How to take loudness out of the equation when mastering

[@martijn_nl]

Taking loudness out of the equation using perception A|B has helped me a lot to make better decisions choosing the right fx. Loudness is the most deceiving aspect of music I can think of. Perception A|B should be a built-in feature of every DAW.

[@michal.ochedowski]

15 minutes well spent. Thanks for sharing the knowledge.

[@ProductionAdvice]

Pleasure !

[@MarlonMehr]

Fantastic video, thanks much!!

[@lamentistriangle281]

Great job! It's a nice demonstration and the entire process, your workflow gives such a vibe that modern times are a bliss when a work is done well.

[@Audiojunkk]

Your mastering is so good! Another great video

[@ProductionAdvice]

Thanks !

[@AvithOrtega]

10:55 I would also suggest to select Volume level "Quiet", just as you have set up in your preferences, so the LUFS threshold is even lower (as far as I understand that's what it does). But what I would like to know is which LUFS level "Quiet" volume level is, instead of "Normal" which I guess are -14 LUFS, do you know what value is it?

[@ProductionAdvice]

Yes, the Quiet setting is -23 LUFS and the Loud setting is -11 LUFS. I think it makes most sense to test @ -14 though, since this is the default setting and what >83% of users will hear

[@MR_Cellarpop]

Thanks Ian.

[@ferrograph]

Excellent video. I was please to see -14db as reference. A lot of new audio folk dont know the background behind some of the levels used in times past. The old days of BBC PPM meters would naturally leave some headroom because the target middle marker on PPM's (4) which engineers would aim for was around this level. It allowed sufficient headroom for louder passages without clipping. If tracks are normalised to Full scale it leaves no room for anything, like writing in CAPS. Its good to see that Spotify understand this even though they have a normalise setting in their app that normalises everything to full scale by default.

[@quintessenceSL]

Neat plug-in. I see a marketing opportunity of selling an "audiophile reference" version of the song with lowered output. At triple the price.

[@ProductionAdvice]

I think you could be right ! Hopefully not 3 times the price, though 😛

[@vitaminfian]

Thank you for this

[@laubsch75]

Very cool! Nice plugin.

[@ThisMichaelBrown]

Question: But if you aimed for -14 LUFS, wouldn't you get the same perceived loudness and get the extra headroom for transients etc...??? perhaps I misunderstand some of that? Thanks for the great content!

[@EmmzBeats1zx]

yh im having alot of trouble when uploading to youtube. My beats sound good only when turned up on youtube to the max but i want it to be loud half way loud like most producers

[@lordberly]

is this gonna be an episode on mastering podcast?

[@Sesaon3]

I think this message has been in every episode ;-)

[@ProductionAdvice]

Interesting - we don't usually do things with audio examples for copyright reasons, but maybe Lij would let us use the song... do you think people would like it ?

[@lordberly]

@@ProductionAdviceFor sure. We need some audio example too ✌️

[@billsherrington5996]

Thanks Ian. One question, can you explain why your final master was created at around -10LUFS so requiring it to be turned down 3+ db when normalised? Is there any advantage in doing this rather than just mastering it to -14?

[@ProductionAdvice]

Sure - the short answer is "sound not numbers", but I can give more helpful details that that: First, my choices are based on what has consistently sounded good to me over the years. (And also sounds great on some of the best-sounding albums of all time, which is no coincidence !) In practise that works out as mastering the loudest sections no louder than -10 LUFS (short-term). For this song, with similar levels throughout, that gives an overall value of -10.7 LUFS. If the song had more internal dynamic contrast you'd get a lower overall level, for example if there was a bigger contrast in arrangement between the verse & choruses, or if it had an acoustic introduction or similar. In that case the overall value might be closer to -14, or even lower. For songs intended to be quiet, the overall value could be even lower still. I’ve written more about this approach in this blog post: www.productionadvice.co.uk/how-loud And there’s a demo of how it works in this video: kzbin.info/www/bejne/mqKpY5x4fpZpeqM Second, I’m all about *balanced* dynamics. Loud enough, but not too loud. There’s a benefit for some material in being louder - and no particular reason to make it less loud. In fact since many streaming platforms don’t turn up quiet songs, making a loud song quieter could backfire. And finally even for clients like Lij who appreciate the effect of normalisation and the benefits of dynamics, they still like it to be good and loud ! Thirdly, -14 LUFS is probably a temporary number anyway. The AES TD1008 guidelines for streaming services recommend -16 LUFS for music, and eventually I wouldn’t be surprised if an even lower number is used, as it is in broadcast. So “aiming” for a specific value doesn’t really make sense. And finally, aiming for targets doesn’t work musically. You wouldn’t want an acoustic ballad to be as loud as this song, for example. I do what feels right to me for the music, then double-check it works when normalised as well. Hope that helps !

[@billsherrington5996]

@@ProductionAdvice Thanks so much for this comprehensive answer! I shall bear this in mind in my future Mastering projects! I had to match an Abbey rd master recently and was quite surprised that the overall level was -8.6 LUFS!

[@ProductionAdvice]

Yeah, that's sadly pretty common these days. Those last couple of dB can make a big difference 😕

[@kye903]

Here's an unrelated question that you probably don't have the answer to: Is the digital to analogue conversion of LG TVs any good..? Reason I ask is that I'm convinced I enjoy listening to my HD660s from my TV more than my Scarlett 18i8. They definitely sound brighter to me just coming out of the TV... what do you think? There doesn't seem to be any EQ options anywhere on the TV, so it can't be that. Does the average TV have a good DAC?

[@ProductionAdvice]

Honest answer - I don’t know. But my guess is that it’s more likely to be the analogue electronics than the DAC. Could be an impedance issue, maybe ?

[@kye903]

@@ProductionAdvice Can you define these 'analogue electronics' that you speak of, and what effect they might have..? I'm not an impedance expert, but I seem to be able to get the headphones plenty loud with no shortage of headroom to spare. I don't know what this might tell you, but I can drive my HD660s and my Sundara equally well from the Scarlett, but while the HD660s also drive easily off the TV, I have to turn the volume up to 100 for the Sundara and it's still not quite loud enough...

[@ProductionAdvice]

@@kye903 I just mean that DACs are mostly off-the-shelf chips at the moment. I would expect both to be perfectly reasonable but not super-high quality. Beyond the DAC though, you need the analogue headphone amplifier stage, and there are far more variables here - shielding, power supply, gain structure, impedance, components etc. I'm not an expert either but if you're hearing a real difference this is where I would expect it to lie. HD660s apparently are 150 ohms, whereas Sundara are 30 ohms. They will have different sensitivity as well, and these factors explain the different volume levels you get from the TV. But it's more complicated than that, because the impedance of dynamic cans is frequency-dependent, to a degree. So potentially you could get a different frequency response (and different amounts of distortion) from two different devices powering the same pair of cans. So my guess is, one of the amps doesn't match the 660s as well as it might, and this causes an audible difference. You prefer the result from the TV (less bass) with 660s and that's fine. Be aware though it may not be "right" - for example HD650s are warm-sounding, with a broad boost around 200 Hz. And if it's less technically accurate you may be getting extra distortion as well as a different tonal balance. See if you can find impedance specs for the converters - generally you want the cans to be at least 2 or 3 times the impedance of the amp.

[@genejas]

Is there a comparison somewhere with perception vs mcompare vs nugen mastercheck? I currently use mcompare, but I'd like to know if i was missing out on some features

[@ProductionAdvice]

I haven't used either in any depth, so people will have to correct me if I'm wrong about something, but as far as I know both MCompare & MasterCheck are intended for stereo file comparisons only. We put a ton of effort into Perception AB to make it really useful in mixing as well, and *very* fast to use, as this video shows: kzbin.info/www/bejne/gHy3pquBaLmbd80 (Jump to 2:50 if you're in a hurry for the demo) For stereo comparison only MasterCheck has streaming codec preview, and only MCompare has the ability to load multiple reference reference files. You can easily make the same kind of comparison using Perception AB though, by stacking up multiple reference tracks as I show in this video. Hope that helps !