Loudness Wars, Part 1 - Thomas Lund

Thomas Lund points out the pitfalls of data reduction, digital audio processing and sample peak metering in contemporary music production, distribution and playback - often exemplified by so-called hyperpop productions, a controversial trend characterized by obnoxious loudness and extremely narrow dynamics. Conversely, Lund also demonstrates the advantages of adopting loudness metering standards like EBU R128 to counteract such symptoms and deliver lively dynamics and consistent loudness levels regardless of genre, production method and distribution formats.
Please note
To get the most from the audio examples in this talk, we recommend that you select the highest possible playback quality.
About Thomas Lund
Thomas Lund has studied medicine, specializing in perception, has written more than 30 scientific papers on topics such as audio loudness, spatialization and distortion, and has worked with organizations like AES, ITU and EBU on international audio production and broadcast standardization. Lund has worked with high definition audio at TC Electronic for 15 years, and holds position as HD development manager.
About 'Rome Calling - Audio Seminar of The Year'
TC Electronic has gathered some of the industry's most influential speakers in order to cover, from a multitude of angles, the radical changes now taking place in production and distribution of broadcast, film and music as a consequence of the new ITU and EBU standards and recommendations. The seminars focus strictly on technology rather than specific products.
www.tcelectronic.com/rome

[@urbanfabric]

I think anyone who's into production be it music or for broadcast work should watch this video.

[@ChatBlancAudio]

Thank you so much for posting these videos! It's so helpful to have such valuable lectures available, so thank you for your generosity.

[@michal.ochedowski]

Brilliant stuff, with some good jokes on the side! Thanks for this vid, a pleasure to watch.

[@screendrem]

Highly useful information. Thank you!

[@kuglepen64]

Awesome presentation. Tak!

[@DrEMplushrest]

Thanks for this, very informative!

[@TheVintageCustom]

Actually, Foobar2000 ReplayGain now (since v1.0.6) employs the libeubr128 for it's calculations, and thus is using EBU-R128 gating to generate ReplayGain tags, for adjusting the playback volume/gain to make the song meet a target of -18 LUFS (not the -23 LUFS broadcast target).

[@woobilicious.]

16bit is enough for all forms of media if your Speaker system can do 96dB above ambient noise you would be deaf in about 5mins.

[@MrKlixon]

Yes! Yes, yes! Please!

[@AzaIndustries]

I may not understand all of this but thanks a lot for uploading this interesting video!

[@genuineuni]

Interesting video. Being a fan of stereophonic reproduction, I had to answer my own question why I enjoyed it. I prefer it due to less masking of sounds. Why I can understand and relate to those who enjoy surround sound with four (or more) discrete channels. Sound masking comes in handy, such as when George Martin mixed The Beatles, such to "conceal" what he and others had done to Ringo's drumming. :)

[@Gdlen1]

thank god for people like this, I discovered mid/side years ago and heard those space monkeys on many mp3's even in normal mode,and decided I would never pay for them, also at 128 and even slightly higher rates, frequency response drops to 16kHz or lower

[@jakobole]

The mastering-engineer is hired to serve a client. If the client is not happy, client finds another engineer.

[@TheVintageCustom]

ReplayGain has tried to do what iTunes is trying to do now. I think it's a semi-weighted curve with a level of -89 dB SPL ? set as a target, no gating or *intelligent* measurements though. I have employed it, running Foobar with 32bit float, trying to avoid dithering distortions etc. It is not always accurate, and the R128 standard measurement looks very promising. Another good presentation. Thanks for the upload video's these are great!

[@doltBmB]

to clarify, it was es posthumus, and the streaming demo tracks from their website were better sounding.

[@tlooknbill]

I heard the iTunes encoded distortions of the S/M playbacks. It explains the strange timbre of certain passages in songs I was noticing in my Amazon mp3's and is why I only buy pressed at the factory CD's now. Wouldn't want those distortions amplified in the source file, but is it OK for factory stamped CD's? I like the big and loud distorted bass line of Flaming Lips' "It Overtakes Me"-(recorded loud) which doesn't clip my back dash 6x8's. It's the scratchy highs I have to EQ most often.

[@screendrem]

The loudness war has reduced mastering engineers to such a level that they just submit. You either fall in line or are replaced. What else should they do? As a music producer/performing artist I wonder what the hell I should do about this. Currently I just share the info that I can.

[@doltBmB]

Well, even the worst encodes usually preserves the "main" sound to some degree. The devil is in the details so most people don't notice it. They just hear that the vocal or guitar riff or whatever sound is dominant is almost the same and conclude it's perfect.

[@doltBmB]

I'm thinking it's because there's less of a push to grab consumer attention with the sound, and more through flashy effects and "radical" cinematography. Which incidentally is exactly where you see the decay occur in film. On the technical side of things I think this frees up mastering producers for video to 'sperg out and just do it as good as they can. Although there has been a push for sharpening and high contrast in terms of the picture, the sound must be a bit of a safe haven from producers

[@Meteotrance]

In fact clipping in digital sample domain, it's like loosing some data on a compressed video file, making some horrible scrambled mess of pixel and in audio artefact, in Analog domain you can distord, reasonably it's also left the harshness, i love using some VHS hifi tape or old cassette to make the sound softer, i think we also need to go back with analog mixing desk, i have a lot of exemple of good DDD and ADD , CD the first edtion of Supertramp or Kate Bush are very good exemple.