Thoughts about Dynamic Range and Music Loudness

[@LoudnessWar]

Hi Unfa! Thanks for the thorough explanation and examples! Hard to believe it's been over 15 years since I made the old Loudness War video that went viral, and almost 10 years since the Audio Fundamentals course... I'm a fan of your Ardour and other Linux audio tutorials, and am working on transitioning my own audio/video stuff to open-source software. Thank you for your enlightening and entertaining videos!

[@unfa00]

It's an honor! Your video was one of the first things that introduced me to the concept, and it has never left me since. If you'll need some help with going Linux - you may want to join my community chat - when I'm not available, there's always a lot of great people there, ready to help :) chat.unfa.xyz (this is on self-hosted Rocket.Chat, but there's also a Discord server - that's proprietary though).

[@thedeltamode]

For everyone a little experiment : Take all your recent songs and measure the LUFS. lower every song to - 14 LUFS for the loudest parts. Then make a dJ MIX/SET. Put them in order, tell a musical journey. You will find out, the less squashed stuff pulls you in, the squashed stuff especially in comparison turns you off. This is how other people hear your music on spotify. They hear it after great productions and producers who have understood the loudness war is over. So be smart. Make your career long lasting by thinking about this stuff now. and dare to have people turn your music up instead of down. turning music up = dopamine hits = memories are being created. turning music down because it's too loud = you lose.

[@lovely-shrubbery8578]

Ive considered this a lot. I think the cause is that many people leave their volume at a constant level. In that context a less dynamic track will have higher perceived loudness. A related variable is that music played at a louder physical db level will be perceived as having louder highs and lows than at a lower db level. With that being said, dynamics are still better.

[@atacoman]

I knew unfa was a Dan Worall fan! Been waiting for you to discuss this topic for quite a while and I'm happy to say you've nailed it.

[@unfa00]

Thanks! I find Dan's videos excellent. I think he's capable of making even the most complex topics easy to grasp. And I find his style very entertaining as well :) I've been sitting on these thoughts for a long time. I was still hesitant to make this, unsure if it's worth people's attention. It was very spontaneous as well - hence a bit of chaos there - though I hope it's going to push my viewers to consider some interesting angles on the technical aspects of music production. I'm glad you've enjoyed it :)

[@atacoman]

@@unfa00 Hey unfa, I just wanted to come back and mention that Dan dropped another (quite amusing) sequel video to his "DJs are Idiots" video on loudness: kzbin.info/www/bejne/qZCkf3iHqpp_qtE . In it he mentions some counterarguments pointing out how "loud mastering" is used somewhat nowadays as more of an aesthetic choice than that of a intentional technical decision. I won't spoil the video for you, but the hilarious (and actually quite interesting sounding) example mix he uses to "defeat the loudness war" actually lends itself to a rather cool and unique sound, despite it being created to present a technical point on loudness. What I'm trying to get here is that, while I do think it is silly for, say, the nth remaster of a classic rock or pop album to be mastered at a unreasonably loud level for its genre, I think the pushing of the LUFS's can work quite tastefully for certain genres and aesthetics. I agree that baking in artificial loudness for the sake of artificial loudness at the mastering stage is a poor decision, but if the mixing decisions leading up to mastering inherently result in a loud mix that works well with the genre at hand, there should be no reason to chastise the mix for being "too loud." At the end of the day, I think it is best to approach the limiter the same way you would approach any other plugin in your DAW or even as an instrument: that is to view it more as an artistic tool that augments your mix, as opposed to a crutch that desperately tries to support something that isn't supposed to be there in the first place.

[@unfa00]

@@atacoman Thanks! I've watched this video earlier today and we had a little discussion about it on my community chat. I think that chasing LUFS levels is totally pointless, because to get a distorted, squashed sound you don't need to use the master limiter, and you van do that at the sound design stage all you want. I voce distortion and it's very useful and spicy. I just don't do that on my master bus. I agree with Dan that achieving the clear, punchy mix only to worsen it with mastering is silly, and I'd much rather have the "un-mastered" mix with all the dynamics in and more punch. Loudness normalisation makes chasing LUFS numbers iirlevant - it'll be turned down anyways, so you might just as well use that headroom for something. If you want your bass to sound like Dan's "I won the Loudness War" track then sure! It's as valid as any creative decision, but it is utterly pointless to do that if the goal is to be louder than others.

[@unfa00]

@@atacoman BTW: I find parts of that insane track really catchy though, and we've talked about this on my chat - it's probably gonna spawn a genre of 2-bit music :D

[@420bobby69]

tfw you realize you're watching a video about dynamic range on an 11 year old macbook pro with one broken speaker lol. Also, i just realized your dedication to FOSS means you do your mastering without using a proprietary crutch like izotope ozone. respect!!!

[@GarthClarkson]

In terms of commercial production and the traditional studios they have always tailored their masters to the most popular playback device of the day. I remember when I had a studio and the Yamaha NS-10M studio monitors first came out. Their selling feature was not that they were the highest quality or had the best performance but that they had the sound that most closely represented the average home HiFi system. They soon became the reference. People don't seem to be that in to home HiFi any more. The closest they get is home theatre which usually has small satellites and a sub. Even now I think their popularity is waning. I think a lot of the problem is that the target playback device these days appears to be the smartphone. If that becomes your target (or the 2x6cm speakers in an LED slimline TV screen) it becomes "don't worry too much about the bass, don't let the drums go crazy", etc, etc. Gone are the days of actual HiFi. Regardless of how amazing our recording and production tech is, it comes down to what sounds acceptable on a 12mm speaker...

[@unfa00]

Well, I would suspect the typical phone earbuds to be the widest target, with semi-decent headphones being a close second.

[@Flutrix994]

Thanks for wonderful Explanation, Loudness on youtube is very confusing.

[@unfa00]

I hope it's not as confusing any more :) If you have questions - let me know!

[@justinflowers2111]

Great points. Good examples and well articulated. I am happier with my mixes and feel they started coming together easier after I started using K-20 or K-14 metering. Particularly when it comes time to add the bass, because that "leftover" dynamic range can still be used in case I need to punch up the track in some way.

[@desreb740]

Great perspective. Thnx

[@vudejavudeja]

I love the fact that you use the very material in which you're bringing accross your ideas to also highlight the concepts you're talking about. e.g. You talk about noise floor while pointing out that the sound of leaves rustling in the background is the noise floor reducing the dynamic range of the signal, which is your voice.

[@unfa00]

Thanks! I also tried to draw parallels to the video (less light means less distance between the signal and the noise floor, which means we need to boost everything to get any image and the noise floor becomes more and more apparent).

[@4polelowpass]

War is, in any context, the ultimate form of failure! You have some great examples in this video and you explain this topic well. No need to apologize for the bad video quality/focus and the audio is great. The information presented here is of a much higher importance.

[@unfa00]

Thank you!

[@sergiovegaz]

Very nicely explained amigo, thanks as always and ... Saludos desde México!

[@dimitrivanlieshout1615]

Respect Unfa ... you explain it very well

[@unfa00]

Thanks! I wasn't sure if this video is worth making, but it seems it was!

[@dimitrivanlieshout1615]

@@unfa00 you have a gift , to explain everything very clear , thats what we need .... greetings from the Netherlands ;)

[@georgemartinez1334]

Amazing video Unfa, as usual :D

[@andx4024]

i like you vids UV#20 Basics of dynamic range compression (w/ Calf and UV#21 Advanced dynamic range compression w/ Calf

[@digitalspecter]

Physical limitations can sometimes be used to push theoretical limits. Way back when PCs used to have a buzzer speaker that only played 1-bit sound. It was either 0 (off) or 1 (on). So, in theory you could play different pitched square waves with it. However, if you drove it *very* fast the poor speaker couldn't keep up.. you could switch the direction before the membrane had reached its intended state. This meant that you actually could generate much more interesting sounds.. We would play something like 11KHz 4-bit sample by sending it as 88KHz 1-bit sound.. and it worked! I still remember the state of awe I was in the first time I heard that polyphonic music coming out of the speaker that used to only make those screeching beeps.

[@unfa00]

Oh man. I remember playing an old DOS game called "Electro Body" (it was published under a different name by then called Epic Megagames) and it did play actual PCM audio through that buzzer. I still don't know how it did that, but I guess it must be the method you've described? ALSA on Linux has a pc-speaker audio output driver and I believe it could do this too, but when I tried it on a semi-modern PC, it would sound like something is frying inside my PC case, so I didn't investigate too much :D

[@digitalspecter]

@@unfa00 Yeah, it probably used a method similar to that. One problem with it was that it used quite a bit of what little resources machines back then had because it had to "hand feed" every bit to the speaker. Haven't tried it in years (well, decades), but I'll take your test as an indication that it's better left to the past :D

[@unfa00]

@@digitalspecter Heh. I was thinking if I could use the PC speaker pins to drive an LED, but I thoght I'll better not try, since I'm not knowlegable about electronics enough to be sure I'm not going to blow up my motherboard :P

[@smashSpikeMC]

me: mentions that the generic radio music sounds flat and boring. everyone in the car: then turn up the volume! me: {face palms}

[@sealica6159]

unfa unexpected made a asmr

[@marbangens]

It will get smoothed out by sinc function. For normal volume 16 bit should be enough. Even with high volume should sound fine. Dynamic plugs often do calculations in 64-bit. If you gonna play something with extreme volume you can export to wav 64 bit. Its not really limited to 16 mostly on cd. But usually we dont care, 16 bit is enough. But I would personally always export to 24 bit wav.

[@marbangens]

​@jordanrodrigues1279 The interpolation could be considered when you think of bit depth if you want. When we program there is only the scale 1.0 to -1.0 where 0.0 is no current drag on the element of the speaker. I build my own speakers with plywood, I had to use programs to calculate the air volume and base on cubic form of the speaker to get extra base, base reflect. How it works is that -1.0 will push the drum inwards and 1.0 will push the drum out with a electromagnetic force. Now consider a float moving like the drum, so y = ax, where a is amplification if you like, now consider Domain and Range (its 0 to 1 all real number cus we have to stop at 1*1 or 1*(-1)) of a = x/y do you see? You can consider change and amplitude equal or directly related. Now by definition we have been given that 0 dB means a=1, meaning to "a" also to the decibel or like a string point if you like. Im gonna stop here assuming you know how decibel and log scale works. You also need to know how floats work to understand this fully. I don't disagree with you because you dont need to think about it so much. Use you're ears it will be fine :) " 24-bit makes it about as quiet as thermal noise - the molecules of air themselves." I like this :)

[@BrunoVernay]

In the same spirit , radio station used to play the albums faster and faster to sound more lively than the others. There is a Supertramp song about that, long time ago

[@unfa00]

Oh wow. I had no idea about that!

[@lambda7652]

Hey Unfa can you pleases show us a view good open source real time Analyzers? I Used Baud line and its Awesome but sadly its not open source :(

[@unfa00]

That's a good idea for a video! Thanks :) I've used to use Baudline years ago as well, for the lack of a more Libre option. But now there's plenty of open-source choices. Wolf-Spectrum is a very nice spectrograph: github.com/wolf-plugins/wolf-spectrum There's also Spectacle, a very nice spectrogram: github.com/jpcima/spectacle x42 Spectr[a] is the most precise spectrogram out there - it seems a bit strange but it's super accurate: github.com/x42/spectra.lv2 Calf Analyzer has a nice spectrogram and vectorscope in one: calf-studio-gear.org/doc/Analyzer.html LSP Spektrum Analyser is also nice: lsp-plug.in/?page=manuals&section=spectrum_analyzer_x2

[@lambda7652]

@@unfa00 Wow, Thanks a lot. I also know about kokkinizitas stuff like Jaaa, Japa and his Ebumeter. Even experimented with GNU radio because there are platy of nice FFT waterfall displays available. but nor relay suited for audio.

[@davidr2421]

Oh no, Unfa has moved out to a cabin in the woods like Luke Smith

[@georgemartinez1334]

Ohhh by the way, do you know any Open Source headphone calibration software? Thanks in advance :3

[@unfa00]

Hmm. I have never needed or searched for any...

[@georgemartinez1334]

@@unfa00 ohhh don't worry! Thank You for the answer anyway :D

[@Zephyrus0]

Tried StarGate DAW yet?

[@InnerTranquility]

The loudness war is beyond stupid. The best sounding albums have dynamics.

[@gx1tar1er]

the worst i've heard is Death Magnetic and Californication

[@baumschubser2249]

Everything I listen to is -3 LUFS what the hell

[@unfa00]

Sweet Squarewaves!

[@InsideOfMyOwnMind]

Loudness for the sake of the artist's intent is fine and if that sucks well then the music sucks by definition but if a well recorded piece gets absolutely aborted by a radio station just to get more catches on a seek button then it's a disservice to everyone frome the artist to the listener