TPS086: Loudness Normalization with Georg from Auphonic (Part 2)

Tweet about this on TwitterShare on FacebookShare on Google+Email this to someone
Play

Loudness normalization of your audio is a “simple” process for measuring and processing for the perceived level of your audio. Perceived levels are how the human ear interprets levels. In recent years, new standards for measuring perceived level have arisen out of a lot of research and engineering.  Georg Holzmann and I discuss the history of measuring loudness and these new standards more in depth in part one of this interview. For an even more complete understanding of these newer standards please see the links at the end of this post for useful resources.EBU R128 logo

In part two of our talk, Georg and I take a look at loudness normalization as provided by his service available at Auphonic.com. We also talk about how a podcaster can start to begin measuring and performing loudness normalization in their own audio editing system.

Also discussed are the ways in which you can better prepare your audio in the post production stage. Strategies for making your audio level across a single track as well as across multiple tracks are discussed and we also touch on effects such as peak normalization, EQ and compression.

For podcasters, measuring audio levels is usually handled using the tools most available to all of us…peak meters. The issue here is that peak meters tell us nothing about the actual level of your audio. They only measure exactly what they were made to measure, the peaks of your audio waveform. If you produce a podcast to peak consistently around -4db, on a peak meter, the only thing you know is that a lot of your peaks are reaching -4db.

With loudness normalization we now have an objective standard that removes the guessing process from producing audio to a particular level. Loudness standards for the most part are a spec made for broadcasting but research has been done into the best levels for Internet/mobile device delivery and the level most agreed upon is -16LUFS. Producing your audio to a perceived level of -16LUFS will result in a file that is loud enough to be heard on lesser speakers (i.e. earbuds) and over background noise such as traffic while removing the need or desire to push your audio levels too high, most likely stripping it of its dynamic range and essentially squeezing the life out of it as often happens in over-compressed, super loud audio.

The advantages of performing loudness normalization are many:

  • You will know exactly how loud your audio is. No more guessing
  • The level of your audio across your entire library will be consistent from episode to episode.
  • You’ll meet agreed upon standards for online distribution.
  • Your show will be on the same level with other shows that also use the standard.

Not producing audio that meets standards for loudness could result in annoying your audience at best and at worst could actually be painful to listen to. If one podcaster produces audio that is -23LUFS and the next is at -12LUFS the differences when going from one podcast to another are going to be very jarring.

If all podcasts were produced to an agreed spec for loudness then the listening experience would be much more pleasurable for everyone. This also helps push our medium (podcasting) past “amateur hour” by establishing and producing consistent audio that meets professional standards.

*Links mentioned on this episode (in order of appearance).

Do you need a reliable host for your audio or video podcast? Consider moving your show to Libsyn.com (my chosen host) and get your first month free when you use promo code: PODCASTHELPER at checkout. If you’d like to use Blubrry to host your podcast (my other recommended host) you can get a free month by using the same promo code: PODCASTHELPER

To learn more about “loudness” standards, the history and the solution, here’s a fantastic talk.

Sign up for the TPS newsletter and never lose contact with the show.

Subscribe to The Podcasters' Studio Podcast
itunesRSSMiroStitcher
* all links should be considered affiliate links. Read my ethics statement for more info.
About Ray Ortega

Full-time podcast producer and host of The Podcasters' Studio and Podcasters' Roundtable, I enjoy sharing my ten years of experience making podcasts to help others improve or start their own show.

Comments

  1. Woo Hoo! Was waiting for Part 2 – thanks, Ray!

  2. Hi Ray, Regarding your conversation about True Peak and inter-sample peaks, the True Peak measurements in EBU R-128 specifies 4x oversampling, which gives a maximum possible under-read error of just less than 0.7dB, hence the -1dBTP setting just to be on the safe side.

    However, higher oversampling ratios have associated smaller inter-sample under-reads. E.g. 8x ratio gives a maximum error of just under 0.2dB, and 16x is accurate to 0.05dB. Higher oversampling ratios require greater processing power though.

  3. Hi Ray! Thanks for bringing this interview. I'm not sure if I understood everything, it was very technical. However, I led me to buy Auphonic. Do I just have to run Auphonic to my audio or is there anything else that I have to understand from everything that was talked about?

    Thanks!

    • Thanks and have fun with the Auphonic app. You will drop your completed audio edit into the Auphonic app and it will do its magic;) You can select various processes that Auphonic performs. Play with options to find out what works best for your audio and check out their blog for some good resources on using it.

  4. Alexandre Sales says:

    Great info Ray, i'm glad i've found you and your work through Youtube. Greetings from Brazil.

Speak Your Mind

*