How Do I Normalise The Audio In My Music, Video Collection?

Ok – as an audio professional who works every day on music and sound for tv, film, radio etc – I’d have to say this article is a little concerning.

There ares miss-truths, important information that has been glossed over (or is missing) and potentially advice that would see people wasting thousands of dollars.

The original question asked about normalizing the levels.
Normalizing has nothing to do with compression. At all. All that does is set the peak audio level to a pre-defined point.
As is stands, 99.9% of CD’s you buy, and mp3′s you download LEGALLY (ie, itunes etc) are normalized to (or just under) what is known as digital 0 (0dBFS)
Even classical music – which has a much larger dynamic range.
It is just a fact of life (and the limits of current digital audio delivery systems) that you may need to adjust levels for things like classical audio if you have just listened to a rock CD or MP3. This is because of dynamic range differences, and limits of the systems involved.

You should not have to do this for rock or pop – if the music was mastered correctly. If you are having issues with any of this style of music, I would be seriously looking at how the music was delivered / captured. There is a problem in the system somewhere if it is not close to the “normal” level. Most of these problems occur with illegal music downloads, with 3rd parties capturing the audio and not really understanding what they are doing.

And the quick way around this is to (as the article suggested) normalize the audio file within a program like audacity. But, given you have 1000′s of files, I’d be looking at a batch convertor. On the mac, I use a program called “sound converter” which will do whole directories of audio at a time. Quick, easy, and almost impossible to stuff up if you understand the settings.

DO NOT compress the audio (as in use a compressor – there are 2 types of audio compression… one involves dynamic range, which is what this article is mainly dealing with. The other has to do with product delivery formats, and should also be avoided if the original is not PCM data).
It takes many years for engineers to learn how to use an audio compressor (and actually understand what it is doing to the audio.) Attack, release times, compression ratios and thresholds are very important to the overall enjoyment of music – and difficult to get right. 99.9% of the time, these have been set VERY carefully (in a mastering studio) – and any EXTRA compression can ruin the audio beyond belief.

Now, video is harder. Why? Because standards differ so much – even in different parts of the world. However, again, normalizing should see you good. But it is not as easy to achieve as straight audio files inside a small audio utility. You may have to demux the video (if it is mpeg2, you will have to re-save the video at the very least.
Most TV / Video audio that is NOT destined for movie theaters (they have their own, very very special audio requirements) have a “brickwall limiter” implemented somewhere along the chain…(normally at the end) to a predefined value. It is then just a matter of working out what that value is (again, audacity should have the tools to do it) and then bringing that peak level UP to digital 0 (if you want it to match the level of your music!)

However, some playback systems for video have auto normalization in them. This is what I would be looking at as an option first.

It is a complete waste of money to get an analog external compressor. They are expensive (thousands) and if you don’t know what you are doing, the audio will be worse than you started. They require constant changes (ie, you’d need to set it for every time you hear a level change!) and you will experience the music as it WASN’T intended by the people who mixed it / created it.

In addition, the comment about latency and hardware compression is a little miss leading. All analog boxes (like the buzz audio box in your picture) have effectively ZERO latency. The only boxes that COULD incur latency would be digital boxes, and even then, these are likely to be unnoticeable. Latency occurs in the digital world, NOT the analog world.

For audio personally I just use MP3Gain. I’m sure the quality of it isn’t the best, but I’ve never noticed a problem and it’s definitely the fastest.

What about using replaygain on an app like foobar to normalise? One benefit of this is, it doesn’t change the MP3 file itself, it just changes the way foobar plays the file. If you want to read more I’d just look on the foobar website or the foobar – Hydrogenaudio forums.