There is a fair amount of content available about mixing Dolby Atmos® for music. And you can find a lot of useful information, as well as a Dolby Atmos Pro Tools session, that you can download and explore at Music That Moves.
However, for this blog article, I wanted to concentrate on some different aspects and share my tips and tricks to keep handy when mixing Dolby Atmos for music. These are points that I follow in my own mixes, and I hope they help you as well. I also have a mix breakdown that will take you through how I approach mixing a song.
When it comes to delivering Dolby Atmos music, there are two main digital audio compression schemes that are used by streaming services: EC-3 and AC-4 IMS. As a mix engineer or music producer, you might not be familiar with these, but it is good to know why they are important and how they will affect your mixes.
Without going into technical detail, the EC-3 file format is primarily intended for playback using speakers. Amazon Echo and Blu-ray players use this, for example. This is also what is used in Apple Music for Dolby Atmos playback. The EC-3 format does not consider the binaural information that is baked in when we mix because it is intended to play back through speakers, whereas binaural requires headphones. The exception here is when the EC-3 is played on the iPhone. At that point, Apple uses custom spatialization to be able to deliver the Atmos content over the approved headphones or the iPhone speaker.
AC-4 is used by services such as TIDAL, Hungama, and Anghami that play back on your Android phone. This format does make use of the binaural metadata that we embed when mixing.
Let's look at some tips and tricks I follow when I mix, keeping the binaural format in mind.
- Getting the Atmos mix to match an existing stereo mix took me some time to start off with. It is common to use stereo wideners or harmonic processing in the mastering of a conventional mix, but these techniques don't work if you are mixing in Atmos. If you are mixing a song that is already released as a stereo version, you might find the binaural to have spatial accuracy but be lacking in punch. This is why you need to be clear on which objects need to have the binaural positioning enabled or disabled, and if enabled, what the distance should be. If used correctly, the binaural version will sound truly immersive and give you a far better listening experience than the stereo version.
- When you begin your mix—and before you create a master file on the Dolby Atmos Renderer—it's a good idea to have all the objects enabled in the session. I usually make sure I have at least 64 depending on the project. The reason for this is that once you create a master, you won't be able to punch in the same file if you create additional objects. So, having them ready when you need them is a good idea.
- Use the Dolby Atmos Binaural Settings plugin in Pro Tools. It's indispensable. I always have a timecode track as part of my Track Preset where the first insert is the Binaural Settings plugin and the second is the Dolby Timecode plugin.
- Use the near, mid, and far settings for binaural placement creatively to make space. Though this will change the frequency spectrum, if used well, it will give clarity.
- Elements on the center path will not contribute to any binaural experience. So, if you pan a sound down the middle of the room, it won't provide any binaural response.
- Your mind will quickly mask a static pan. If you have even a tiny bit of movement, the mix will come to life. I do this by flipping my pans to faders and randomly moving them slightly. I sometimes set the auto glide time (in Preferences > Mixing tab) to match the tempo. This way, even though the faders (Pan) are pushed randomly, they come back musically. Your mind will pick up the randomness and yet find closure in the tempo-based return.
- You don't have the opportunity of M/S (Mid/Sides) mastering that you have for stereo, but you can decide the width in a much more powerful way with placement and binaural distance.
- Be modest with the overhead placements. The slight phase and tonal change won't be noticed if it's static or there are too many elements at that point—even on the same plane. At the same time, remember that there is no absolute left and right in binaural. Keep this in mind to prevent your binaural mix from clipping.
- Consider frequency when you are positioning. Sending low frequency towards the heights rarely works. A good idea is to split the frequencies and pan them separately. I use the Avid Pro Multiband Splitter a lot for this positioning method. This allows me to have different frequency ranges like the Low, Low Mids, High Mids, and Highs, on different auxes and pan each of the auxes separately in the room. The Pro Multiband supports what is called Auxiliary Output Stems. This allows the different frequency bands to be sent as separate outputs to Aux Inputs.
- Use three stereo tracks for creative panning. Set one track to near, one to mid, and the third to far. This way you can take one element and move it from near to mid to far in binaural position by editing it across these tracks (the binaural distance cannot be automated). This can reduce clutter in the mix.
- In binaural mixes, you don't have absolute left and right positioning—it's always leftish or rightish. If you are not careful, you can overload the center part of your mix or the stereo track, thus reducing the width of the mix. Be creative on assigning (or not) the binaural processing. This way you can create a depth beyond the stereo mix.
- Once my mix is done, I export an ADM for mastering. I then make a new session, import this ADM file, and tweak it if needed, as I would approach mastering for a stereo mix. It is this master that I deliver with the loudness spec and tonal correction. This is helpful when I have an album or an EP to mix and need to maintain consistent levels across the songs. This is also why I maintain my bed and object count so that all songs have the same track count. At this point, I monitor in binaural.
- From the above master ADM, I do a "Save As…" of my Pro Tools session and set my monitor in the Dolby Renderer to stereo mode for stereo deliverables. This is where I use the new session to check the pans and levels in stereo. When you switch from binaural to stereo, it's always a good idea to check the lows, low mids, and the pan levels. When your objects are set to mid or far in the binaural render, there is a good chance that the low end has changed in the monitor to reflect that. So, when you switch to stereo, some EQ tweaks will be needed. If you have surround pans, then it's always a good idea to listen to them. The levels may drop considerably in your stereo mix. I export this ADM as an intermediary version called Stereo ADM and then import it into the Renderer to then output a stereo re-render. If you have an RMU or an RMW, you don't need to do this step, as you can record the stereo downmix at the same time.
- My default starting is Binaural off on the beds. It seems to add more punch and tonality compared to it being in the default mid mode. This is a personal preference though.
- When importing an ADM into a session, I follow these steps:
- Delete all the beds and outputs in the I/O Setup and then do the import.
- I check that the renderer has the same input configuration as the ADM. If not—and I received an external ADM to master—I first import that into the renderer to get the Input config.
- Import the ADM into Pro Tools. That way the mapping is automatically taken care of.
- For my last few mixes, I have my first bed set to VOX bed, followed by 10 Vox Obj, then Music Bed, and finally the rest as Music Obj. This way I can export a minus track from the mastered ADM itself.
- Adding elements to the surrounds is a good way to not just get an enveloping mix, but to also stay within the loudness spec. This helps to offset the loading of the fronts and increase spatial positioning.
- In the stereo (2.0) downmix, I personally have found the LtRt with 90-degree phase to be a much closer representation of the surround mix to my ears for music.
- I am very careful of using speaker snap in music. Cinema mixes are different, and I am a big fan of using it in that context. But for music, because of the way I understand clustering works, there is a chance of hearing a jitter in the pan position. So, unless it is really needed and that space is not clashing with other objects, I avoid speaker snap. It is the same with size. I am very careful with increasing size and don't usually go beyond 15–18. This is because the decorrelation that size introduces can create some phase issues.
- I tend to leave a lot of headroom in my objects. This way when they sum, the peaks aren't shooting too high. As long as the Atmos mix hits -18 LUFS, you will have headroom for this, although I personally have my reservations about this since TIDAL stereo is -14 LKFS and the binaural version of a -18 LKFS Atmos mix is -16 LKFS. So, there is a 2 dB loudness loss here when your Atmos mix plays against a stereo mix. For this reason, I usually end up with limiters only on my beds. If there are transient elements in an object, then I would use a limiter there, but it is rare I do so.
- I like to keep the LCR space powerful and keep the other elements moving. An easy way I found to do this is to flip the F/R pans of the objects to faders once the balance is right and slightly touch them randomly. That way the mix will constantly be breathing and moving, giving a sense of motion to the mix.
- Finally, I am a huge fan of the Dolby Atmos Music Panner. I convert the pan data from the Music Panner into Pro Tools pan data by recording it as Object Pan data on the same track. The Pro Tools offline ADM export is very helpful—especially if you start hitting CPU issues with plugins.