The digital front end of my Atmos music system can be a little confusing for many audiophiles. I completely understand the confusion. This system is a bit different from the norm, but I believe it produces results that are a bit different from the norm as well. As one can see in my recent objective and subjective review of the system, it's capable of amazingly high quality. Here is a more detailed look at what I'm doing and why I'm doing it.
Atmos Music Done Right, Or At Least Right For Me
The digital side of my Atmos music system can be separated into three pieces, music servers, Audio over Ethernet, and audio hardware. Each of these items will likely be addressed in an article or two later, but a detailed overview is due first. Also, keep in mind that being on the cutting edge of anything often requires a little more work, but it's absolutely worth it in this case. Those uninterested in putting in the effort, can always user easier methods without any judgement from me. Whatever works for you, is all good.
I also have a note below about where this "should" go in the future to make it even more accessible, easier to implement, and if possible, even more enjoyable.
Here's a simplified diagram of the digital side of my Atmos system.
My Atmos Music Servers
I play all my Atmos content from two music servers, one running macOS 12.5, the other running Windows 11. I have an Aurender that also plays my 12 channel files, but given the company's current workload and priorities, I don't believe this feature will see the light of day anytime soon.
Why use a music server rather than an AppleTV or NVIDIA SHIELD? Using a music server enables me to use state of the art convolution for room correction for the best sound quality, enables me to use a variety of music playback and library management applications and their iOS/Android remote control apps, and enables me to NOT use HDMI in any part of the playback chain.
Room correction / DSP on a music server is more capable than anything available from a hardware processor. Hang Loose Convolver enables me to use 65,000 tap FIR convolution filters for time domain and frequency domain room correction. These filters are designed by Mitch Barnett, using the state of the art software Audiolense. The most advanced multichannel processors use roughly 4,000 taps. According to Mitch, when writing about my system, "We use Audiolense’s time domain (i.e. excess phase) correction capabilities at low frequencies to shape the timing response towards the ideal minimum phase response. This time domain correction is what makes the bass clear sounding. It is one thing to smooth the frequency response for even sounding bass, but quite another thing to have clear bass, which is what FIR filtering (with independent excess phase correction) has to offer over all other room correction filtering technologies. This is in addition to the FIR filters superior frequency resolution with 65,536 filter taps which is the equivalent to a graphic equalizer with 32,768 eq sliders. That is 1000 times the frequency resolution compared to a 31 band graphic equalizer."
Mac - MacBook Pro (15 inch, 2017) with an Intel Core i7 3.1 GHz quad core CPU, 16 GB LPDDR3 memory, and a 1 TB NVMe solid state drive. I use the Mac to play Atmos music a few different ways.
- Streaming Atmos music from Apple Music. Apple Music on this Mac outputs audio into Hang Loose Convolver, which sends audio out of the Mac. Apple Music streams lossy Atmos using the Dolby Digital Plus JOC codec, but for many album, this is the highest quality every released. I encourage everyone to listen to Elton John's Rocket Man streaming in Atmos from Apple Music. It's a fantastic experience
- Playing local lossless TrueHD Atmos files I've either purchased/downloaded or ripped from Blu-ray. I use both Audirvana Studio and JRiver Media Center 29 for these lossless 12 channel WAV files. Audirvana uses Hang Loose Convolver as a plug-in for room correction. JRiver has its own convolution engine, enabling me to use the same filters as I use in Hang Loose convolver.
- Playing local TrueHD Atmos files through the Dolby Reference Player. This is the only way to decode TrueHD Atmos files on a computer. I set the Dolby Reference Player to output to Hang Loose Convolver, just like Apple Music, for room correction before the music is sent to the rest of the system.
Astute members of the Audiophile Style community are likely wondering how I can use Audirvana and JRiver, if the Dolby Reference Player is the only way to decode TrueHD Atmos files. I setup a process whereby I use the Reference Player to convert all my TrueHD Atmos albums into 12 channel WAV files, playable by any application capable of playing 12 channel WAV files. I convert into 12 channels because my system is 7.1.4, or 12 channels. They can be converted into other configurations such as 5.1.2, 9.1.6, etc... the same way. The conversion takes time as it's a realtime process, but it run unattended, so I don't even have to be awake while it's running.
Windows - Custom built CAPS Twenty music server, written about extensively here.
- Playing local lossless TrueHD Atmos music files with JRiver and Audirvana. Much the same as I do with my Mac, and others do with stereo files, I'm playing 12 channel WAV files and using the iOS apps for remote control. Convolution works the same way, either through Hang Loose Convolver or built-in convolution engine in JRiver.
- External Upsampling with HQPlayer. My CAPS Twenty PC has quite a bit of power, so I can use it for upsampling my 12 channel files to high rate PCM or DSD in HQPlayer. HQP also has its own convolution engine that will handle my filters designed by Mitch Barnett. This is an avenue I've yet to fully pursue, but I look forward to jumping in with both feet. The benefits of upsampling have been heard for decades inside "every" DAC, in products such as dCS upsamplers, and in software such as HQPlayer.
Future Simplicity - Mac and Windows are incredibly flexible and can be installed on relatively inexpensive hardware. However, this is a double-edged sword. Incredibly flexible can also mean instability and complexity. Inexpensive hardware is nice, but it isn't the height of living for audiophiles. Does anyone remember the MacBook Pros with the keyboard and specific USB ports on the same bus, enabling a nice pop and click once in a while while trying to listen and type at the same time? What a feature :~)
The future of TrueHD Atmos playback can follow the template already set by high end music server manufacturers, for high resolution two channel audio. For example, an Aurender could have a built-in software Dolby TrueHD Atmos decoder for playing play lossless TrueHD Atmos music files, outputting via Ravenna to systems with 16 channels. Trinnov already does software TrueHD decoding, so it isn't a reinvention of the wheel. Plus, the only changes to existing servers are all in software. This could mean many units in existing audio systems are upgradable for TrueHD Atmos playback.
Let me make clear, this is my vision of how I hope music servers progress, not related to anything manufactures and I have discussed. I hope they all implement it tomorrow though.
Audio Over Ethernet
Tying my music servers and digital audio hardware together is Ravenna. It's an AES67 compatible technology for sending audio over Ethernet. An in-depth discussion of AES67 and Ravenna can be heard in my 2017 interview with Merging Technologies' Dominique Brulhart here.
Ravenna equals freedom and quality when it comes to audio playback. Sending audio via Ravenna around the network, and even around the world, is rock solid. Ravenna was designed to handle high PCM and DSD sample rates, extreme channel counts, and very low latency. It's the standard that should've been used in audiophile products years ago, rather than the most non-standard standard that is UPnP/DLNA.
Ravenna is used to record live performances of full symphony orchestras, where there are no second chances. I think it's robust enough for even the highest of audiophile expectations. It works with macOS, Windows, and Linux, which means almost any music server is capable of sending audio via Ravenna. Most high end music server manufacturers have yet to enable Ravenna, but Aurender has supported it for years.
Note: I frequently play 12 channel DXD music in 24 bit / 352.8 kHz WAV files, without a single hiccup. Morten Lindberg at 2L recently started selling his DXD discreet 12 channel WAV files, and there's absolutely nothing like them. No decoder is necessary to play them as they aren't encoded like TrueHD 24/48 files for Atmos.
Using Ravenna I completely avoid using HDMI in/output anywhere in my system. I also gain the ability to use different dedicated DACs rather than route my audio through an AV processor. Even if a DAC doesn't support Ravenna with an Ethernet input, devices are available that convert Ravenna to AES/EBU for output to a more traditional DAC or group of DACs depending on how many channels are needed.
On Windows, Ravenna uses ASIO and works just like any two channel audio device that supports ASIO. On macOS it uses Core Audio and is seen in Audio MIDI Setup, just like any two channel audio device connected to a Mac.
Ravenna is an open standard without a proprietary license, but manufacturers who've implemented it usually require their hardware for it to work or they sell their implementation of Ravenna on its own. For example, Merging Technologies sells its Virtual Audio Driver / AES76 / Ravenna for Mac here - https://www.merging.com/products/aes67-vad
Recapping the digital part of my Atmos system thus far, I use a MacBook Pro or Windows 11 computer, running Apple Music, Audirvana, and JRiver, with room correction and Dolby TrueHD Atmos decoding, outputting audio over Ethernet via Ravenna. Now for the digital audio hardware.
Digital Audio Hardware
Anubis - I use the Anubis like a digital to digital preamp. Audio is sent from my computer via Ravenna to the Anubis. I setup two 7.1.4 12 channel sources on the Anubis, and I can switch between them just like switching inputs on a preamp. One for my Windows PC and the other for Mac. The Anubis is extremely flexible, as software defines everything other than the physical in/outputs visible on the hardware. The only wire going to my Anubis is a single Ethernet cable that simultaneously power the unit and sends/receives Ravenna audio over Ethernet.
My favorite part of the Anubis is its volume control. The unit sites right next to my listening chair, giving me easy access to switch inputs and spin the large volume dial for fine r coarse adjustments. There's even a nice mute button on the top for immediate silencing of the music.
Music goes into the Anubis, then out from the Anubis to the Merging HAPI MKII. Again, all over Ethernet using Ravenna, without any loss of quality and very low latency. The HAPI MKII is setup as a DAC only, with its volume dial set to maximum. HAPI MKII uses ESS ES9028PRO DAC chips and has dynamic range of 125 dB (> 127 dB A-weighted), THD+N in excess of - 116 dB, with new local low phase noise oscillator circuitry.
The system can play 16 channels up through DSD256 and PCM at 32 bit / 384 kHz. The HAPI MKII also has a built-in 8 channel AES/EBU output, for people who want to use other DACs to do conversion. A four channel system with a couple dCS DACs could be amazing. In that case, the Anubis isn't even necessary, as audio could route directly from a computer to the HAPI MKII, and out from there to the DACs with volume control. The HAPI MKII features its own volume control, but it's only via the physical front panel knob. A bit inconvenient for us audiophiles.
After the HAPI MKII converts the audio to analog, it's off to the amps, and an article for another time.
At first blush this system appears complicated. However, I've done all the legwork and know what works and what doesn't. Most people can also use a single music server and an iPad for playback. I have to use everything because I have to understand how it all works and test it, so I can write about it.
The easiest route to Atmos music is to use something like an NVIDIA SHIELD with a hard drive full of MKV files, connected to a processor like a Lyngorf MP-40 or Trinnov Altitude16. For most people, this is all they'll need and it can double as a home theater processor. I'll write more about this option at a later date.
As an audiophile who loves music and understands technology, I had to implement the best system I could put together. This involved a mix of consumer and professional software, state of the art digital signal processing, and professional hardware. A system like this doesn't exist as a package or from a traditional high end audio manufacturer. Perhaps soon we'll see music servers and 12 channel DACs from our favorite companies, to play immersive audio.
To view the complete system visit https://audiophile.style/system
To view my articles about immersive sound visit https://audiophile.style/immersive