I have been playing with computer-based audio playback for 2 years already, which excluded using music streamer like Sony HAP-Z1ES which I bought earlier. Computer audio playback refers to using general purpose computer as player for digital music files, either stored locally, external USB drive, or network drives.
My initial exploration started with using Windows laptop connected to USB DAC. I tried Foobar, JRiver, and HYSOLID. I fell in love with HYSOLID due to its genius design – the player is a Windows service (runs in background, no need for Windows login), and playback control is done by phone/tablet. However, I don’t like having a laptop on the hi-fi shelf.
Later I bought Intel NUC and added 19V LPSU from China. After trying various software including Daphile, Volumio, and Euphony, I selected Euphony and bought my first Euphony license in July 2019, a Linux-based paid software. For all these Linux-based software I have tried, they are truly headless since even for Linux kernel upgrades, there is no need to have keyboard, mouse and connection to monitor/tv – everything is controlled from phone/tablet. However, Euphony wins hands down for best sound quality and ease of use.
J5040-ITX motherboard with multi-rail LPSU
The past year however, I build my own linear PSU for low-powered motherboard and its accessories (5V for SSDs, optical drive, and USB Audio output). With good low-noise linear LPSU, the USB Audio Output sound quality into USB DAC is now at different performance level. It demands better playback software since each software sounds different, not to mention Linux versus Windows.
For me, the best Linux headless playback software remains Euphony. In fact I bought the second Euphony license in June 2021. As I tweaked my streamer to perform at higher level, the limitations of Linux-based playback become more apparent. Linux has its own sound since it uses generic USB Audio driver and ALSA mixer. Windows, on the other hand, requires USB audio driver from DAC manufacturer, and it typically includes ASIO and WASAPI / kernel streaming drivers as well. Compared to the previous versions of Windows, Windows 10 has been optimized for low-latency audio playback. When I compared HYSOLID running on Windows 10 (optimized for background tasks only without any other excess baggage), I was very surprised it sounds more musical with finer details (higher resolution) and darker background compared to Euphony. Due to HYSOLID limitations (no longer supported, phone/tablet app can only run on old Android/IOS devices), I was eager to explore another Windows headless software.
JPLAY is the only software I have not tried. When I found out that Taiko Audio SGM Extreme, the reference-grade music server/streamer with starting price of 24k Euro, uses JPLAY, I feel bummer, why did not I try this earlier? Today I downloaded and installed JPLAY FEMTO into the same SSD I installed HYSOLID. It has 21 days trial period. License cost is 149 Euro. I signed out HYSOLID first, and then during JPLAY FEMTO installation, I selected installation for Intel Core and for single machine, that means optimized for playing music files from local drive. I have all my music files inside a 2TB Samsung SSD, so it is pretty easy to point the location of all music files into the D:\Music directory. I also use JPLAY interface to select ASIO driver for my Marantz SA-10 used as USB DAC. I kept the others as default. With ASIO driver, native DSD playback is possible up to DSD 4x/512. With the default KS (kernel streaming) driver, DSD playback is by DoP (DSD over PCM), and limited to DSD 2x/256. I restarted the streamer so that Windows starts but remain not logged in.
For playback control, BubbleUPnP app is required for Android. The BubbleUPnP app needs to be setup where the library points to JPLAY femtoServer and the renderer is JPLAY FEMTO. By default, gapless playback is disabled, so I enabled gapless playback and also disabled FFmpeg audio decoding as recommended by JPLAY manual.
Wow, JPLAY FEMTO brings my streamer to different performance level never heard before. That’s the ‘aha moment’. I regretted not trying JPLAY earlier. It sounds more dynamics and lively, with tighter bass. If HYSOLID is better than Euphony, JPLAY is better than HYSOLID. Now I know why Taiko Audio selected JPLAY as their playback software, definitely it is the best out there.
Will JPLAY works for everyone? A few things to note. During installation of JPLAY FEMTO, the streamer needs keyboard, mouse, and connection to monitor/tv. JPLAY settings also can only be done through Windows. After all these settings have been made, the streamer can run headless. Any upgrades to JPLAY software can only be done with keyboard, mouse, and connection to monitor/tv. I am ok with this since I have a 65″ tv in my system. Getting keyboard and mouse connected to the streamer is not a big issue. The main challenge for those who are not PC literate is to do the settings in optimizing Windows 10 for audio playback, for example, disabling all background apps and set processor scheduling to Background Services. This is best done with fresh Windows install with only USB audio drivers and JPLAY FEMTO installed.
If one needs a system with a lot of functionality, including Roon integration, ease of use with friendly user interface and setup, Euphony may be a better solution. However, for those who want the best sound quality possible (me included), and not bothered about the needs to occasionally connect keyboard, mouse and monitor/tv to the streamer, JPLAY FEMTO running on Windows is the best software available. JPLAY is cheaper than Euphony too, 149 Euro for a perpetual license which can used on 2 PCs, compared to USD299/249 Euro for single Euphony license. Both JPLAY and Euphony have free trials, so why not try it yourself and share your findings.
UPDATE: BubbleUPnP free version has limited capabilities, e.g. max 16 songs in playlist. I bought it for RM15.80 which is reasonable.
To get the best sound quality from Windows 10, either with HYSOLID or JPLAY FEMTO, please create Windows password, then reboot. Let the login screen remains, no need to login. The audio related background services automatically run upon Windows startup. Without Windows login, many unnecessary apps and tasks will not run, hence CPU can handle the audio threads better.