The goal of this thread is to provide members with a forum to give and receive tips when creating a master for streaming playback.

The internet has resurrected the business it almost extinguished 20 years ago. Last year was the third consecutive year of global growth for the music industry with revenues totaling 20.2 billion U.S. dollars.There is no doubt that digital music streaming has been the clear driver of this growth with 9.2 billion U.S. dollars in revenue.This represents 45.54% of the global market. It is projected that digital music streaming revenue would increase to 17.2 billion U.S. dollars by the year 2021 which will represent 62.55% of the global market.* Are your mixes optimized for the streaming revolution?

There are a few things to consider if you’re looking to mix and deliver a master for streaming services. I have included a few and members are encouraged to contribute with different ones or to expand on the ones below.

1) Almost every major music streaming service is using loudness normalization. In the loudness normalization environment the perceived loudness of all the songs on the service is automatically adjusted to sound approximately the same. The main platforms playlist now only differ by 3 LU between the loudest and the softest. See the included graph.

2) It is important that you understand what the 'LUFS' metering system is. The different streaming services have different loudness standards and algorithms to take measurements and apply the normalization but for the most part they use the basic unit system of loudness measurement called LUFS or LKFS. LUFS stands for loudness units relative to full scale. It was developed to enable normalization of audio levels, and matches how our ears actually perceive sound rather than just the electrical signal.The higher the number after the minus, the less loud it will be perceived, so -12LUFS is louder than -18LUFS. Being able to understand how our music masters are metering with this scale is useful to see what will happen when they are streamed on different services (i.e. will the algorithm gain them up or down to meet the target or not?). A different master for each medium and service is ideal but not practical. Somewhere in the middle of -16LUFS and -11LUFS might be the best target loudness for your music to ensure optimum playback across all platforms.

3) Dynamic is the new Loud. The loudness war is over. The use of heavy compression and limiting to make your track sound louder than the other will no longer work and therefore when mixing and mastering for streaming you must maximize the dynamic range.

4) Optimize master for various codecs. Various data compression codecs are used by different streaming sites.Tools are available that are purposely designed to let you hear how your mixes will sound under different streaming codec settings.Software based processors such as Apple's RoundTripAAC,Nugen Audio’s MasterCheck Pro,Sonnox’s Fraunhofer Pro-Codec and iZotope’s Ozone 8 Advanced allow the potential side-effects of data compression to be evaluated in real time.


---------------------------------------
* Source: https://www.statista.com/statistics/239276/growth-of-the-global-music-revenue-by-type/

 

Last edited by a moderator: Apr 13, 2018

You should write for one of the magazines that people under 25 buy. While it might take a while to sink in, one line you put will change the whole face of music if they listen. While it may not seem better because so many like red-lining, when they try it they will be surprised at the new life in their music.
You might have to explain the method and the benefits even though I know what they are and many others probably do too. This is the only pitfall of an under 25's music younger generation mastering magazine. Great stuff.
What a great subtitle line:
"Dynamic is the new Loud."

Perfect.

 

A few nitpicks:

Sorta. For streaming (and some for downloads). S̶p̶e̶s̶h̶u̶l̶C̶l̶o̶u̶d̶ SoundCloud does not, AFAIK.

It's nice to know, though metering plugs will normalize your material with a mouse click. Select top circled, click bottom circled.

Most streaming services (including S&*!Cloud) transcode ALL uploaded content. Meaning the 192kbps MP3 you upload will be transcoded before being streamed @ 192kbps. Just like the 32-bit-deep .wav.

People who type about music claim dynamic is the new loud. Those who listen to & buy music feel that loud is the new loud, just as it has always been. I'm not up on current music, but guessing that teh dynamic range of current top 10 bangaz is no more HDR than it was 5, 10, or 20 years ago. Just a guess, but that's my story and I'm sticking to it. Until I see proof (as in measurments, data) to the contrary.

Don't they all stream MP3? The actual code could be different, I guess. Could you tell me the codec YT uses, and how it differs from Vimeo?

 

@Lambchop even if you decided to call it "nitpicking" I appreciate your reply and welcome your participation.
You are correct, Soundcloud is not currently using loudness normalization but it is coming. The graph I attached explains what I just told you but I have edited the thread to reflect it.
When I mentioned that different sites used different data compression codecs is because it is a fact. MP3 is sometimes used but the most widely used codecs are AAC which iTune uses, and Vorbis which Spotify uses. On the other hand, Tidal uses Lossles. As for You Tube and Vimeo it is AAC-LC .

 

Ah, I see they updated the graph I posted a while ago.
https://audiosex.pro/threads/loudness-on-different-online-platforms.25671/

Yeah, nice thought in theory but could you please name just one distribution platform that lets you upload several different masters for several different platforms as the same release?

 

Is not that what I said? A different master for each medium and service is ideal but not practical. With that said, allow me to explain that we have been providing clients with 2 different masters. One for streaming services and another for physical media. I believe that is the best and most practical compromise.

 

Bah, (embarrassed smiley) my mistake! Really should know better. I blame ̶s̶o̶c̶i̶e̶t̶y̶ https://www.telechargerunevideo.com/en/

The rest stands tho -- lossy compression; not much you can do about what happens to you upload. It gets transcoded regardless of upload format, so uploading uncompressed seems like the the smartest thing (to me, could be wrong). As far as loudness normalization thing goes, the worst a streaming platform will do to your (overly loud) mix is lower its overall volume for you. I think.

 

Lower its volume? It will destroy your song! A "very loud song" when normalized will sound flat and overly compressed. A song much quieter will suffer less normalization and hence provide the listener with a more musical and pleasurable listening experience. That is why optimizing for streaming is recommended. Look at the images below.
Both tracks have been peak-normalised, the track on the right has been treated to significant loudness processing. The second screen shows the same two tracks matched for loudness as defined by the new standards — and the audible differences are as startling as the visual ones!

 

Not sure what you mean by "normalization." Normalization alters the content in exactly the same way as a volume control knob. Here's an example of normalizing content to 0dB:

As far as your pics go, this (assuming red is the input file & green is not float):

shows clipping in the green.

This is not normalization, or, rather, inane normalization to +something dB (clipping). No streaming service will do that to your content (turn it up to make it clip). Normalization done by streaming platforms is strictly downward, as in "turn it down if too loud," never "turn it up into heavy clipping if loud enough." Not sure what the pics are meant to show :\
P.S. "loudness 2.jpg" also shows clipping, albeit played at a lower volume.

 

Last edited: Apr 13, 2018

http://productionadvice.co.uk/taylor-swift-loudness/

http://productionadvice.co.uk/stats-for-nerds/

 

If that shows anything, it sure ain't "dynamic range is the new loud." It seems to show just the opposite --

...but it doesn't. All it shows is some guy @ DynamicRangeDay.com (lol) doing some srs cherry-pickin'

 

The comparison was made by a personal friend,Hugh Robjohns, for an article that he wrote in 2014. By his own words:

These are screen grabs from Cockos Reaper comparing two stereo tracks ripped directly from commercial CDs. The one on the left (coloured here in red) is Chris de Burgh's 'Devil's Eye', taken from an early 1980s CD release, while the one on the right (in green) is Grand Prix's 'Samurai', taken from a much more recent, remastered 'greatest hits' album. I apologise for my (lack of) musical taste, but these tracks highlight the issues nicely.

As you can see, the Chris de Burgh track isn't heavily limited or compressed at all, and it exhibits quite wide-ranging musical dynamics building and falling throughout the track. Using a BS.1770-compatible loudness meter I measured the True Peak level at -1.4dBTP, and the Integrated Loudness measured -18.8LUFS for the entire track. In contrast, the Grand Prix track has a true peak level of +0.5dBTP (in other words, some inter-sample clipping is present!) and an Integrated Loudness of -7.8LUFS, which is 11dB louder than the de Burgh track. This loudness disparity is painfully obvious if you play the two tracks back-to-back. Interestingly, I also have the original CD album containing that Grand Prix track (I was a closet head-banger back in the day!) and that version measures 0dBTP and -9.6LUFS. Clearly, the 'greatest hits' re-release suffered some further destructive compression and peak limiting during its re-mastering!

The lower screen grab shows what happens to these same two tracks in a loudness-normalised environment. In this case I employed a target loudness of -23LUFS — the HDTV standard, and a little lower than the current iTunes Radio setting. Now, the incredibly loud Grand Prix track (in green) has been attenuated to bring its loudness in line with the target value, with the result that its peak level is now just -13.7dBTP, with virtually no variation other than for the intro and a break towards the end. The Chris de Burgh track needed less attenuation to match the loudness target and, importantly, its more dynamic nature means its transient peak level reaches -4.5dBTP, a full 9dB higher than the Grand Prix track, despite both having equal overall loudness.

So, the track that was remastered to sound amazingly loud and 'in your face' has actually come out sounding flat, overly compressed and aurally uninteresting compared with the far more dynamic, but originally 11LU quieter, track. Not surprisingly, it's the Chris de Burgh track which comes across as the more interesting and musical, and is a more pleasurable listening experience. I'd encourage anyone to try this sort of comparison for themselves: hearing is believing.
---------

By the way, thanks for your participation. I know that others may share your feelings and concerns.

 

Again, normalization is nothing more than turning the volume knob. Nothing. No compression, no expansion. Yes, things sound deader and flatter at low volumes, that's why people like things louder. Yes, things that were mastered with higher PLR will be turned down less.

I assumed red & green were before/after, my mistake. Your friend picked some ...unusual examples to make his point:

Both are normalized to YT's -13LUFS (because YT)

This is what Grand Prix's 'Samurai' looks like, streamed by YT.
Wait for it.

Your friend doesn't just cherry-pick tunes, he also plays it fast & loose when it comes to numbers. Yeah, -23 is a little lower than iTunes' -16. How much lower, you ask? Well, certainly more than the difference between YT and iTunes (iTunes normalizes to -16LUFS).
It's ...this much:

 

The streaming revolution: Are your mixes optimized for it?

Very intelligent thread topic fine sir.

This is my mindset about this-

Dynamic as possible and RMS DB at or around -12 ( with zero limiters)
unless some hefty EDM then -10 RMS db ( brickwall limiter shaving 1.5 db off tip of transients)

Also important to say achieve extra clarity so that when the artificial intelligence "auto limters" algorithm add their touch of harmonic distortion it will still be considered HD audio.

 

https://www.facebook.com/DynamicRangeDay