Jump to content
  • austinpop
    austinpop

    Review: Ayre QX-5 Twenty USB2 Upgrade

     

     

    Introduction

     

    image2.jpgAs regular readers of my DAC reviews know, I’ve been a fan of Ayre DACs for quite some time. When I got back into audio a few years ago after a long hiatus, my first serious DAC of choice was the small-in-stature but mighty-in-sound-quality Ayre Codex, which made my beloved HD800s sing. Over the years, I’ve had extended stints with the QX-8 and the QX-5 Twenty, with the latter being one of my long-time reference DACs.

     

    Why? Well, because it does so much right. The sound quality of a DAC is about much more than its digital filters and D/A conversion engine. Religious battles have raged for years between proponents of SDM (sigma delta modulation) and R2R (resistor ladder) converters, chips and FPGA (field programmable gate array), linear and minimum phase filters. Yet in my experience, these differences pale in comparison to the quality of the “meat and potatoes” basics of DAC design: the power supplies, the clocks, the analog output stage, and the preamp. And this is where the strength of the QX-5 Twenty lies.

     

     


    The QX-5 Twenty USB2 Upgrade Board

     

    When Ayre announced last April that they were releasing a USB board upgrade, my ears perked up, and I immediately requested it for my long-term review unit. Here are the features of the new board:

     

    • Better incoming noise isolation
    • New Ultra-low-noise voltage regulation
    • Ayre re-clocking technology
    • More powerful XMOS 2 chip (XU208)
    • Support for higher sample rates:
      • up to 768 kHz PCM
      • up to DSD512 Native
      • up to DSD256 DoP.

     


    Why This Interested Me So Much

     

    On the face of it, the features above are “motherhood and apple pie” in the world of DAC design. Most savvy digital audiophiles know of the benefits of galvanic isolation, better power supply regulation, and better clocks. But embedded in this announcement were some additional exciting implications.

     

    First up is the USB controller, which is a hardware module that presents a USB interface to the upstream music streamer/player, while delivering data to the downstream DAC circuitry as an I2S stream. There are multiple OEM suppliers for these controllers, including XMOS, Amanero, and ComTrue. The relative quality and impact of these USB controllers has been an area of research by Emile Bok of Taiko Audio, creator of the Taiko Extreme music server. I am both an owner of the Extreme as well as an admirer of Emile’s findings and insights, which periodically result in further increases in SQ that he delivers to Extreme customers as software updates. Emile has been a big fan of the XMOS2 chipset, claiming that its extremely low latency characteristics, which when exploited by the Extreme software tunings, deliver superior sound quality. The fact that this new Ayre USB2 board uses one of the XMOS2 chipsets opened up the potential of better SQ with the Extreme. 

     

    The other area of interest in the USB2 upgrade was the support for sample rates up to 16FS (705.6/768 kHz) as well as native DSD512. To understand why this is interesting, let’s step back and revisit the data path with the original USB board, which supports sample rates up to 8FS (352.8/384 kHz). A PCM music stream that arrives at the USB input is first sent to an FPGA (field programmable gate array), where it is converted to 16FS with Ayre’s proprietary MP (minimum phase) oversampling filter (OSF), after which it is passed on to the ESS9038Pro chip for D/A conversion. 

     

    What’s the relevance here? When discussing this with Ariel Brown, Ayre’s chief designer, he confirmed that with the USB2 board and the latest firmware, when the FPGA logic is presented with 16FS at its input, it bypasses the Ayre MP filter and passes the stream directly to the ESS chip for D/A conversion. This means that music files that were upsampled upstream by HQPlayer or PGGB, for example, pass unmolested to the DAC stage. This now opens the possibility of further SQ gain, if PGGB or HQPlayer  reconstruction filters sound better than the Ayre MP filter.

     

    The conversion path with DSD is different, in that the input stream does not undergo the OSF in the FPGA, but is passed through to the ESS chip for processing. On some DACs, there is a benefit to upsampling native streams to DSD256 or DSD512, but this is DAC dependent. The support for native DSD512 in the USB2 board allows users to experiment with this path as well.

     


    Review Hardware Chain

     

    My system hardware for this review is shown in the picture below. It consists of a chain of audiophile switches, feeding my music server, the Taiko SGM Extreme, equipped with the Taiko USB card upgrade. Audio data is output over USB to the Ayre QX-5 Twenty. My reference headphones are the Meze Elite, the Abyss AB-1266 Phi CC, and a modded Sennheiser HD-800 (SD mod), driven by the Cavalli Liquid Gold headphone amp. The switches are powered by independent DC rails supplied by a Paul Hynes SR7MR3DRXLFC10 power supply. AC power is delivered via a 6AWG dedicated circuit to a Sound Application TT-7 Reference power conditioner, to which the amps, the Extreme, and the Paul Hynes PSU are directly connected. Additional details are available in the System Details section below.

     

     

    Review-topology.png

     


    Software

     

    I did the bulk of my listening with local files resident on my Extreme server. Both native and PGGB-upsampled files were used, and were played back using the TAS (Taiko Audio Server). Only for the cases that I used HQPlayer upsampling, I used HQPlayer 4 Desktop as the music player as well.

     

     

    Review Playlist

     

    Fall 2021 Review Playlist on Qobuz (US)

     

     

    To enable you to listen to the same tracks that I did, I have created a public playlist on Qobuz USA. This playlist includes the tracks mentioned in this review, as well as some of the others I listened to in the course of this evaluation. Please note that in some cases, the Qobuz track will not match the mastering I listened to, especially since all my listening was with PGGB-upsampled files. Still, this gives you a sense for the music I listened to for evaluation..

     

     

    Installation and Setup

     

    The new USB2 card is a small circuit board that replaces the original USB board. It comes with a 1-page instruction sheet and is very easy to replace, as long as you’re comfortable wielding a screwdriver, and take the usual precautions with power and grounding. I am not particularly handy, but had zero issues doing the switch myself. I’m sure most users will be able to get their Ayre dealers to do it for them, if they prefer.

     

    image3.jpg

     

     

    My board also came with a USB flash drive to update the QX-5 firmware to recognize the new card. I was also able to download the latest Windows ASIO driver from the Ayre website. This driver is only needed for music streamers/players running Windows, like the Extreme.

     

     

    Listening Comparisons

     

    I conducted my listening tests in 3 phases, to evaluate the different aspects of the USB2 upgrade:

     

    1. Improvement due to the hardware alone (all users will benefit):
      • Compare USB2 vs. USB1 hardware (native content, Ayre ASIO driver)
    2. Improvement due to Taiko’s exploitation/optimization of XMOS2 (only benefits Taiko Extreme users)
      • Compare Ayre ASIO driver with Taiko ASIO driver (native content)
    3. Improvement due to USB2’s 16FS support and OSF bypass
      • Compare native content that undergoes Ayre’s MP filter vs. upsampled (PGGB or HQPlayer) content that bypassess the MP filter.

     

     

    Improvement of USB2 over USB1

     

    This comparison is really the crux of this review. Without any further exploitations in software/drivers/OS or upsampling, how much of an SQ uplift does the swapping of the older USB1 board with the new USB2 board provide?

     

    To evaluate, I listened to several tracks in their native format, without any upsampling, before and after the board swap, keeping the ASIO driver (Ayre’s own) unchanged. Note that these were not ideal A/B comparisons, where I could play back A then B then A in quick succession. Rather, I listened first with the USB 1 board, then swapped the board, which took a few minutes, and then listened again. As QX-5 users know, it takes the unit some time to achieve full SQ after a full power cycle, so this was a confounding factor I was prepared for.

     

    Let me just cut to the chase. Despite the challenging A/B logistics, the improvement with USB2 was laugh-out-loud obvious from the first note. Do you recall the improvement going from a vanilla USB cable to your current favorite? Well, this was bigger than that. Much bigger.

     

    image1.jpgWhile the Mendelssohn Violin Concerto, Kyung Wha Chung, Charles Dutoit, Orchestre symphonique de Montréal (Decca, 16/44.1) is a good recording, the USB2 made it sound great. Chung’s violin took on a richer, silkier tone, without a hint of stridency. The somewhat thin sound on the USB1 of this 1981 recording was now more substantial and fleshed out on the USB2, with a larger sound stage. There was a clear reduction in noise floor, as it was now much easier to hear and appreciate the harmonic lines of other instruments under the melody of the violin. All in all, listening with the USB2 was a much more immersive and emotionally engaging experience.

     

    I could end the review right here and say - if you’re a QX-5 Twenty owner, run, don’t walk, to your phone/computer and place an order.

     

    But wait, there’s more!

     

     

    Improvement due to Taiko USB Driver

     

    Apologies, this section describes improvements only owners of  the Taiko Extreme can get. However, do check with your specific server vendor and ask if they have any XMOS-specific optimizations in software, as these may well give you similar SQ gains.

     

    In the case of the Extreme, Emile Bok and his team have decided the best way to ensure correct exploitation of specific USB chipsets, along with the optimal tuning parameters, is to deliver a customized ASIO driver package of their own. The Taiko USB Driver is an omnibus, a collection of drivers for all supported DACs. When a DAC is detected by the OS, the correct driver in the Taiko USB package gets loaded. 

     

    When I first installed the USB2 board, I ran a utility Emile provided to read the USB board’s descriptors, which I sent him. Within a few hours, he installed an updated Taiko USB driver package to support the Ayre USB2 board. This driver could now be selected as an alternative to the Ayre ASIO driver.

     

    image5.jpgYou may think a simple USB driver wouldn’t improve SQ, but you’d be surprised. On Ayub Ogada’s haunting song Kothbiro, the Taiko driver rendered Ayub’s voice with more emotion, soaring larger and with longer decays. The plucks of the nyatili lyre, the accompanying string instrument, had more texture and detail than with the Ayre ASIO driver. Remarkable! 

     

    Of course, I should caveat this by saying that the benefit of the tuned Taiko driver should benefit the USB1 board too, since it is still XMOS-based, albeit with an older controller. Due to the order I tried things, I didn’t test the benefit of the Taiko driver on the USB1 board, so all I can report is that the Taiko driver delivers a significant SQ bump with the USB2 board.

     

     

    Improvement due to Upsampling

     

    We’re not done yet. I next looked at the effect of upsampling. For PCM upsampling, my tool of choice is PGGB, although many prefer HQPlayer.
     
    When I had tried using upsampled files in the past with the USB1 board, I had to create 8FS (352.8/384 kHz) upsampled files, as this was the maximum sample rate the USB1 board allowed. This meant the Ayre MP filter was still involved in the processing path, taking the upsampled 8FS input to 16FS.

     

    On the USB1 board, the SQ improvement going from native files to 8FS PGGB files was marginal to modest at best. Since using PGGB upsampled files demands many times the storage space of native files, I would not recommend upsampling for the QX-5 with the USB1 board. I came to the same conclusion with the QX-8 in my review a few years ago.

     

    But the USB2 — ah, that’s a different story! Knowing from Ariel’s technical explanation that the QX-5’s FPGA logic bypasses the Ayre MP filter with 16FS streams, I listened to tracks, both at native and 16FS sample rates, the latter created by PGGB. This was, in effect, a comparison of the native Ayre MP filter with PGGB’s extremely long windowed Sinc filter.

     

    image7.jpgOn the first movement of Lalo’s Symphonie Espagnole, Tianwa Yang (violin), Darrell Ang, Barcelona Symphony Orchestra (Naxos, 16/44.1), this lovely recording and performance sounded impeccable at the native 16/44.1 sample rate, using Ayre’s built-in MP filter. But the PGGB-upsampled file revealed even more. Instrument separation was much better, and easier to locate in the soundstage, especially in the dense orchestral crescendos. Yang’s violin had more texture and crisper transients. I heard similar textural improvements with percussion, and better bass articulation in the double basses and cellos. These improvements were consistent across a variety of recordings, which meant that the USB2 board opened the door to additional SQ gain through the use of upstream upsampling tools like PGGB.

     

    What about DSD upsampling? My Taiko Extreme music server is not designed for CPU-intensive upsampling, so I had a friend with a HQPlayer Pro license upsample a couple of albums to DSD512 for me to test. First was the Lalo album above (16/44.1 PCM to DSD512). The USB2 board played the DSD512 stream with no issues, and this DSD512 version sounded better than the original native version. There was a richer tonality, and more body and density to the instruments. I still preferred the PGGB 16FS PCM-upsampled version, but this becomes a matter of individual taste.

     

    image4.jpgThe second album upsampled (or rather, remodulated) to DSD512 was Mahler Symphony No. 2, Gilbert Kaplan, Wiener Philharmoniker (Deutsche Gramophon, DSD64). I heard similar improvements in instrument separation, soundstage size, and tonal texture, when comparing the native DSD64.   

     

    The main takeaway here is that the USB2 board enables users to experiment with upstream upsamplers like PGGB or HQPlayer, should they choose. I encourage anyone going down this path to listen and decide for yourself if you prefer the built-in Ayre MP filter, PCM upsampling with PGGB or HQPlayer, or DSD upsampling with HQPlayer.

     

     

     

    Comparison with Reference System

     

    To recap: the SQ I was now hearing with the QX-5 Twenty with the USB2 board was much improved compared to the USB1 board, due to the combination of the hardware improvements in the USB2 board, software exploitation by the Taiko USB driver, and improvements due to upsampling, enabled by the USB2’s support of 16FS sample rates.

     

    I was now curious how the QX-5 Twenty with the USB2 board would compare to my reference setup shown below:

     

    Baseline-topology.png

     

     

    The QX-5 Twenty has a Pono-style dual 3.5mm balanced headphone output. Shout out to the folks at Transparent Audio for providing me with the correct adapter for their Ultra headphone cable system. This enabled me to use the Ultra cables for all the comparisons. With the USB2 upgrade in place, would the direct headphone out provide any advantage over my external Cavalli headphone amp, as it does on my DAVE?

     

    Short answer: no. The Cavalli amp improved SQ comprehensively over the direct headphone output. On the DAVE setup though, the best SQ is through the direct headphone output. So I compared the QX-5 Twenty through the Cavalli to the DAVE direct.

     

    image6.jpgListening to this wonderful new recording and performance of the Brahms Symphony No. 4, Manfred Honeck, Pittsburgh Symphony Orchestra (Reference Recordings, 24/352.8), there was hardly any daylight between the two setups, suggesting the QX-5 Twenty improvements had closed the gap considerably. The DAVE/SRCDX/DC3 still led on transparency, pinpoint imaging, and layering. It was easier to follow multiple melodic and harmonic lines simultaneously. The QX-5 Twenty’s strengths were in the density and physicality of instruments, a richness of tone, and better bass articulation.

     

    A few days before the end of my comparisons, I received a well-burned-in Sean Jacobs DC4 ARC6 PSU (MSRP $10k) for the DAVE for evaluation. With this PSU in place of my DC3, the DAVE setup pulled ahead significantly, especially making up for the deficiencies in the areas of deep bass and physicality. But at what cost? This had now become a David vs. Goliath comparison! The DC4 ARC6 PSU price alone is comparable to the QX-5 Twenty’s MSRP. Add in the DAVE, the SRC-DX, and let’s not forget the dual S/PDIF cables, and this combo is well over double the price of the QX-5 Twenty. Taking this into account, the QX-5 Twenty acquitted itself very well, and remains an outstanding option at its price point.

     


    Conclusions

     

    The USB2 board upgrade for the Ayre QX-5 Twenty is an absolute no-brainer. If you’re a QX-5 owner that uses USB as your primary input, the SQ improvement you will hear with the USB2 board will more than justify the modest $850 price. Highly recommended!

     

     

     


    Reference System Details
     
    Music Computer:          Taiko Audio SGM Extreme Music Server, Taiko USB upgrade
    Headphone Amplifier:   Cavalli Liquid Gold
    Headphones:                 Meze Empyrean, Audeze LCD-5, Sennheiser HD800 (SD mod)
    DAC:                                Chord DAVE
    USB to dual-SPDIF:      Audiowise SRC-DX bridge
    Ethernet Switches:        SOtM sNH-10G, Uptone EtherREGEN, Buffalo BS-GS2016 (modded for LPS)
    Power supplies:             Paul Hynes SR7MR3DRXL (dual regulation, 3-rail)  for switches 
                                             Sean Jacobs DC-3 for DAVE
    Power Details:               Dedicated 30A 6AWG AC circuit, Sound Application TT-7 Reference Power Conditioner
    Power Cables:               Sablon King (wall to TT-7), Sablon Prince (Extreme), Cardas Clear Beyond (DC-3, SR-7), 
                                             Cardas Clear for all other components
    USB cables:                   Sablon Reserva 2020 USB
    BNC cables:                   High Fidelity Cables CT-2 in Schroeder config, JSSG360’d (DIY)
    Ethernet cables:            Sablon 2020, SOtM dCBL-Cat7, Supra Cat 8
    DC cables:                      Neotech OCC (DC-3), Paul Hynes fine silver (SR-7)
    Interconnects:               Cardas Clear XLR balanced
    Headphone cables:       Transparent Ultra cable system
    Accessories:                  Synergistic Research Tranquility Base XL UEF with Galileo MPC, Synergistic Research MiG 2.0 footers, 

                                            Taiko Audio Daiza Isolation Platforms

     


    Acknowledgments
     
    Many thanks to the following companies for supplying cables and accessories to aid in this evaluation: 

    • Cardas Audio, for a full loom of Cardas Clear cables.
    • Transparent Audio, for the Transparent Ultra headphone cable with a full complement of headphones leads and source terminators.

     


    About the Author

     

    Rajiv Arora — a.k.a. @austinpop — is both a computer geek and a lifelong audiophile. He doesn’t work much, but when he does, it’s as a consultant in the computer industry. Having retired from a corporate career as a researcher, technologist and executive, he now combines his passion for music and audio gear with his computer skills and his love of writing to author reviews and articles about high-end audio.

     

    He  has "a special set of skills" that help him bring technical perspective to the audio hobby. No, they do not involve kicking criminal ass in exotic foreign locales! Starting with his Ph.D. research on computer networks, and extending over his professional career, his area of expertise is the performance and scalability of distributed computing systems. Tuning and optimization are in his blood. He is guided by the scientific method and robust experimental design. That said, he trusts his ears, and how a system or component sounds is always the final determinant in his findings. He does not need every audio effect to be measurable, as long as it is consistently audible.
     
    Finally, he believes in integrity, honesty, civility and community, and this is what he strives to bring to every interaction, both as an author and as a forum contributor.




    User Feedback

    Recommended Comments

    2 hours ago, Mercman said:

    I also tried upsampling everything to DSD512 with Roon and Audirvana Studio, but I found that the sound was better listening at native sampling rates with my setup.  Charley Hansen felt that upsampling didn't benefit his DAC and the USB2 board confirmed this,


    Steve,
     

    I wouldn’t use Roon upsampling as a basis for judging the value of upsampling with the USB2. HQPlayer and PGGB are much better.
     

    I recommend trying PCM 705.6/768 (16FS) with PGGB. With HQPlayer, try both DSD256 and  DSD512.

    Share this comment


    Link to comment
    Share on other sites

    As an ES9038Pro DAC, I'd be careful about the limits of DSD256 and DSD512 performance.

     

    Remember guys, bigger isn't always better. There are important nuances and unlike AKM 449X DACs, the ESS offerings do not have "DSD-Direct" modes. Not to say that the ESS doesn't sound good, just that optimal performance needs to be considered in this context.

    Share this comment


    Link to comment
    Share on other sites

    I found upsampling to DSD256 was preferred with my Playback Designs MPD-8. The QX-5 sounded best native. But I haven’t tried the HQPlayer yet as austinpop suggested.

    Share this comment


    Link to comment
    Share on other sites

    1 hour ago, Archimago said:

    As an ES9038Pro DAC, I'd be careful about the limits of DSD256 and DSD512 performance.

     

    Remember guys, bigger isn't always better. There are important nuances and unlike AKM 449X DACs, the ESS offerings do not have "DSD-Direct" modes. Not to say that the ESS doesn't sound good, just that optimal performance needs to be considered in this context.

     

    Good point!

     

    I tried DSD upsampling merely for curiosity. I found upsampling PCM to 16FS to bypass the Ayre MP filter to provide more of an SQ uplift.

    Share this comment


    Link to comment
    Share on other sites

    I believe that you aren’t bypassing the DAC filter, but inserting another filter into the chain ahead of the digital filter built into the DAC.  But if it sounds better to you, that's great.

    Share this comment


    Link to comment
    Share on other sites

    Thank you very much @austinpop

     

    What would you say is the greater SQ delta:

    Going from the old to the new USB  board or (with the new board) from the Ayre to the Taiko driver?

    Thanks

     

    Matt

    Share this comment


    Link to comment
    Share on other sites

    2 hours ago, Axiom05 said:

    Very interesting review. I own a QX-5 Twenty with the USB2 board and I've been loving the new sound, the other inputs are now redundant as far as I'm concerned. I would just like to add a note of caution when comparing straight 44.1 KHz w/ the Music filter vs 16X upsampling and bypassing this filter. Since the Music filter is a slow roll-off type it begins reducing high frequencies just above 10 KHz, and according to John Atkinson's measurements it is down 1 dB at 13 KHz. Personally I find this mild high frequency reduction very audible and it results in the loss of some reverberation and a reduced sense of the recording space. It can make some classical recordings sound very dry and dull, particularly those made with a more "natural" microphone technique. I am one of the few people that actually prefers the Measure filter because it restores a flat frequency response. Consequently, if you upsample 44.1 KHz content to 705.6 KHz and bypass the Music filter you are also restoring the high frequencies compared to using the Music filter with 44.1 KHz content. The same is true for 48 KHz but to a lesser extent. Filter character aside, you may or may not find the change in high frequency response to be audible.

     

    Again, thanks for the review and the useful information that it provided. I had no idea that the Ayre filter was bypassed if fed a 16X file, a potentially very useful thing to know.


    Very interesting. I must admit, I had done the Music vs. Measure filter comparison some time ago, and decided to stick to Music. I didn’t revisit it in the context of the new USB2 upgrade.

    Share this comment


    Link to comment
    Share on other sites

    2 hours ago, matthias said:

    Thank you very much @austinpop

     

    What would you say is the greater SQ delta:

    Going from the old to the new USB  board or (with the new board) from the Ayre to the Taiko driver?

    Thanks

     

    Matt


    The USB2 board upgrade is the bigger SQ uplift.

    Share this comment


    Link to comment
    Share on other sites

    I love that Ayre gives people the option between listen and measure. Given what Rajiv has reported here, people also have the option to bypass both by using 16fs. Very nice options. 

    Share this comment


    Link to comment
    Share on other sites

    Rajiv

    I read Your review of dCS Bartok and now of Ayre QX-5.

    With USB2 upgrade how does it compare now with dCS Bartok?

    Share this comment


    Link to comment
    Share on other sites

    1 hour ago, Vule said:

    Rajiv

    I read Your review of dCS Bartok and now of Ayre QX-5.

    With USB2 upgrade how does it compare now with dCS Bartok?

     

    It has been too long since the Bartók was in my system, and a lot else has changed for me to say anything definitive. All I will say is both DACs are outstanding and deserve an audition.

    Share this comment


    Link to comment
    Share on other sites

    Does anyone know if the  USB2 is used in the QB9 Twenty upgrade??

    Share this comment


    Link to comment
    Share on other sites

    On 6/1/2022 at 4:33 PM, mr_bill said:

    Does anyone know if the  USB2 is used in the QB9 Twenty upgrade??

    Yes it is.

    Share this comment


    Link to comment
    Share on other sites



    Create an account or sign in to comment

    You need to be a member in order to leave a comment

    Create an account

    Sign up for a new account in our community. It's easy!

    Register a new account

    Sign in

    Already have an account? Sign in here.

    Sign In Now




×
×
  • Create New...