Optimizing my Dolby Atmos Music System
“786,432 Taps Polishing the Inside of the Snow-Globe.”
December 16th, 2022, was commissioning day. Hip, Hip, Hurray! Everything is functional. I have essential listening position control of playback using a single mobile device. The system sounds so good.
Early on, I had the system playing quietly while working around the office and basement. When I first walked into the listening area, I felt the sound start to envelop me like walking into a Snow Globe of music.
Very few of us have control over the rooms we use. I had to move some family members’ cheese to put this system where it is. I have to slide a speaker out of place to open the closet door! So, I only get to work on the inside of my Snow Globe. I cannot and do not want to try to make room changes. I want to optimize from the inside out.
Phase I Punch list
- Better remote control - Luna Display
- Clean up some of the cable routings
- Use 19-inch desktop rack rails to get more desk space.
- Eliminate more of the clutter in the room.
- Add Hue lamps for listening position mood lighting.
Phase II Project goals
- Keep the project scope within the Snow Globe.
- Setup for room and speaker measurements: Windows laptop, software, mic connections.
- Work with Accurate Sound to measure and correct the listening experience in DSP.
- Set up Accurate Sounds’ Hang Loose Convolver to work in the system.
- Get a different/extra chair.
- Figure out how to invite others to listen, keeping things friendly and safe!
The most crucial goal of this phase is sound quality improvements using DSP.
Many audiophiles need to be more informed about what DSP is and what it is capable of. This is a case where technology seems like magic to many. I know enough to be dangerous so let me cross that line into “The Danger Zone” and see if I can summarize.
Let me start with a confession. Hello, my name is Bob, I am an Audiophile, and my room is not ideally designed for audio. It is the wrong shape, unbalanced, and too much stuff is in it. My speakers are not perfect sound radiators, and my ears are over 70 years old. My brain needs to be tuned to understand music.
I am not alone in the above statements. We are entering a couple of fields of study unknown to most of us. They are acoustics and psychoacoustics. I could start down the path of research I have been dabbling in for a few years, but we do not have to do that. The tools and the knowledge to correct these problems are at our fingertips.
Here is what I intend to accomplish with DSP.
- Offload bass frequencies away from the speakers (eleven of them) and feed that to the subwoofer
- Tweak the sound from each speaker to correct anomalies
- Correct for room “issues.”
- Align the first time of arrival sound for 12 speakers to the listening position
- Tweak the sound for my preferences and ear
- Have the ability to switch between different filters in real-time
The how of this is a very deep well that you do not have to fall into!
From the aspect of the user, this is pretty simple. Buy some software and a measurement Microphone. Set the software up according to a well-designed process. Hook up your USB DAC and Mic to the computer and make some measurements. Hire THE REAL “Top Gun” to build your custom filters. Purchase and install the Hang Loose Convolver for your playback system to use HLC and the filters that @mitchco makes for you, and sit back and listen.
Here are a couple of side-bars and background you can study at your leisure:
Here is a link to Mitch’s book if you want to dig deeper - Accurate Sound Reproduction Using DSP
Let’s get started
We are using Audiolense XO from Juice Hifi to measure the speakers in the room from the listening position. Audiolense is used to build our DSP filters for frequency correction, digital crossovers, and their True Time Domain correction to align the first arrival time for all frequencies and speakers.
Audiolense runs on Windows; you will need a computer and a measurement microphone. I used the UMIK-1, but there is a potential disadvantage to that mic as it has a separate clock from the DAC. This can cause timing issues for building the True Time Domain corrections.
The setup needs two USB ports, one for the DAC and one for the Mic. Make sure to install the drivers for your DAC and have the calibration file for your Mic to run the measurements and for Mitch to build the filters. In multi-channel, systems care is needed on the channel order; it is easy to get confused.
The measurement mic is placed vertically, and the 90-degree mic calibration file is used. The mic must be placed in the listening position where your head is during listening. Remember that one of the goals is getting the first time of arrival to your ears just right!
Double-check the speaker locations and assignments, then run at least three sets of sweeps. This will play the spectrum to each speaker. When done, load up all the measurements and check the timing on each speaker to ensure no significant delays are shown that would indicate potential clock issues between the DAC and the Microphone.
Mitch will then evaluate the measurements and build filters for the convolver. This is a rather complex process that I would need help executing. I call it DSP magic.
The Blackhole Audio Driver for macOS allows applications to pass audio between each other with no latency. The driver is open-source and free. To use HLC for Apple Music and some others, we need to install Blackhole.
*** Remember to turn off system sounds or direct them to another device.
With HLC running and the Mac sound output pointed to the Blackhole 16 driver. I can use Apple Music, etc., to play music on the system.
NOTE: The HLC Host application must run for the driver to pass audio.
With my first set of filters in place and working, I can concentrate on some hardware setup/cleanup activities. Yes, I am listening, but my son and his wife are here from Germany, so I cannot be an audio hermit!
I have removed the Thunderbay 8 from the system and replaced it with an OWC miniStack STX with a 4 TB SDD and an 18TB HD installed. I am keeping a display on the system as a control center for demonstrations for at least the short term.
The large hard drive gives me the space I need to play with things like PGGB and keep a working copy of my music library on the system along with iTunes/Music library and 60,000 photos.
I have set the pads on the XLR interconnects to -15db. I tested with 10DB pads, but the volume range was too narrow. With the DSP software running and the 15db pad, I have a good control range on volume and some speaker protection.
The Herbies Audio Labs gliders are perfect for moving the right speaker for closet access. I use two Hue Go lights for low-level ambient lighting in the room.
OK, the punch list is finished, and the first measurements are made. Let’s look to see what work the DSP has to do for me.
It looks good, but with 12 traces, there is a lot to unpack. The DYNAUDIO speakers all follow a very consistent flat pattern of reproduction. Listening to the system, even without correction, has been a joy. These speakers are great.
The SUB6 stands out here; good response below 20hz.
The impulse response timing curves show one of the things that we can correct with Audolense XO.
Mitch produced two filter sets for me the Dolby Atmos Audio target and the EBU 3276 target. I also have a third experimental filter that is pass-through. The two curve sets are shown below. Mitch says that the difference is “subtle.” My thought is “interesting.” They feel different, and some music seems “better” on one or the other. I have yet to decide.
EBU 3276 target
Dolby Atmos Music Target
M1 Mac mini Performance
With a 7.1.4 system setup, a filter set consists of 12 filters, one for each speaker and 11 more for bass offloading to the sub-woofer. I have three filter sets in HLC during my testing. One for the Dolby curve, one for the EBU curve, and a test curve with ****Pass Through****, so a total of 23*3 filters when playing music. It would be best if you had lots of buffer space. This allows for instant switching between filters for instant comparison. YES!
With the music playback software, macOS Activity Monitor, HLC, Blackhole driver, and screen capture running, I use about 12% of the CPU. Turn off screen capture, and that drops to around 8%. Activity Monitor accounts for a couple of percent of that. The M1 processor is not even thinking about working hard.
**** As of this writing, the Pass Through Correction filter is experimental and needs to be tested carefully. Low volume, etc.
The sum of all things.
In the screenshot below, you can see most of the elements running on the Mac mini. First is the Luna Display software, with its display dongle plugged into the Mac. Next is the system sound panel showing the 16-channel blackhole driver as the default audio device. The Hang Loose Convolver Host software is running. Below is Apple Music playing and Spotify standing by for testing. There will be more to see as we keep going.
With family here for the Christmas holiday, we only had time for some shorter listening sessions. So my new year's promise to myself is to listen a lot more. And here we are.
I always have too many things running at the same time. But we are cleaned up from holidays, so I can set up the system in the Living Room for another review I am working on. While listening to the gear, I felt something was off. I could not put my finger on the problem, so I moved on.
While working in my office, the “something off” upstairs was bugging me. So I played the same tracks on the ATMOS system, which sounded like I expected. What the heck? OH, slap me on the forehead! I have been listening to a room-corrected, time-aligned, bass-managed system for days. The system upstairs does not have that luxury, as the playback system does not support convolution filters. Whew, my brain is not broken. What is crazy to me is how fast I adapted to the new system.
As about 99.9% of my music is 2-channel, let me start there.
Robert Plant and Allison Krause
Sitting listening to the interplay of Robert and Allison as their vocals intertwine. There is a flow when they intermix as they interplay in dominance on a syllabic level. Listening on headphones does not even reveal this level of detail (YET). This has only been discernible to me with the DSP correction. The clarity of their vocals is stunning.
With the Convolver running, the complex interplay of BT’s composition leaps out of the system and spreads across the front of the room. I would love it if BT embraced ATMOS and re-mixed this one; please update the This Binary Universe 5.1 mix too! I can dream!
I cannot stop playing this album. It is genuinely an Opus.
Moving on to the immersive side of things, I turn the lights down, sit back and allow myself to be transported to where the music takes me. The following Apple Music Playlist is a short journey into my “Snow Globe.”
Suppose you have any of the supported Apple headphones. In that case, you can turn on Atmos playback on your phone, etc., Using the internal speakers, and you can also get a pretty good feel with any of the newer Macs or iPads listening with no extra hardware!
The other afternoon my friend Tony came down to listen to and record an upcoming episode of our YouTube channel, “The Three Techs.” Tony noted that it is hard to share this experience. I hope the taste you get when listening on your Mac or AirPods will lead to curiosity!
Artists, Engineers, and Mixers are starting to learn and experiment with this technology. They are expanding the horizons of music and how it is presented. I want to mention one more album that is a pretty grand ATMOS presentation;
Brian Eno has created an exciting experience here. Watch the video before you listen. Then give it a real shot! Here is Brian Eno discussing FOREVERANDEVERNOMORE on the Dolby YouTube Channel.
“I am a visitor in the middle of something.”, “Feelings are the beginning of thinking.”
FOREVERANDEVERNOMORE (Apple Music)
There is so much more I want to say and do. I am working with Mitch and Bernt on a tiny fanless PC. I plan to test a pair of the new Topping DM-7 eight-channel DACs. I need to get Audrivana working cleanly with HLC (@damien78 , we need your help here). I need to create a couple of playlists for demos and sharing. I need to process a couple of new albums I purchased, so my PGGB-256 machine will need more processing time.
I must finish the data backup and sync processes to keep my data safe and organized. Then there is discovering and listening to music. And none of this even scratches the surface of my wild ideas.