If the data reaching the DAC is bit perfect, can the source (laptop, phone, streamer, CD transport or network gear) still change the sound?

His argument in one place

In May 2018, in the DAVE thread, he set the argument out for "bit perfect data being delivered to one of my DAC's via USB - and here there is only one mechanism that can possibly alter the sound - and that is RF noise injected into Dave and the rest of one's system" Head-Fi, May 2018. In the same post he wrote that "Jitter has absolutely no measured or subjective influence on my DAC's", that RF noise "for sure changes the SQ" by generating noise floor modulation, and that the result "can easily be confused with improved transparency" Head-Fi, May 2018.

He has made the same claim across products and years. In April 2014, of Qute and Hugo, he said the only explanation he could find for reported source differences was RF noise, "as it's not jitter and it is bit perfect" Head-Fi, Apr 2014. In January 2016 he wrote of Mojo, 2Qute, Hugo, Hugo TT and DAVE that "Any SQ differences in sources is not down to jitter but source RF noise and correlated noise" Head-Fi, Jan 2016. In February 2019 (Qutest thread) he first conceded "Technically USB is better from the clock POV; but in practice I have had optical sound identical to USB", then wrote: "assuming bit perfect data, the source can only change the sound by adding RF noise into the DAC" Head-Fi, Feb 2019. In May 2021 (Hugo 2 thread): "It is only noise floor modulation that can affect sound quality with bit perfect sources" Head-Fi, May 2021.

In March 2021 (Watts Up thread), answering a DAVE owner, he listed the possible causes with bit-perfect sources as jitter levels (which he said are not an issue with his DACs), random and signal-correlated RF noise, and audio-bandwidth signal-correlated noise, adding: "I have never encountered a situation that has not been explained by these effects" Head-Fi, Mar 2021. In his March 2024 CanJam NYC seminar he opened by calling it a myth that digital cables and sources cannot change sound quality because the data is bit perfect, then said "Except that, of course, they can", and named the same three issues: jitter, random RF noise and correlated RF noise MFG Audio Reviews 0MFG Audio Reviews 2:33.

The mechanism he describes

In January 2016 he described two errors: RF noise causing intermodulation distortion in the analog electronics, heard "as a brightness to the sound quality", and signal-related current noise in the ground plane, "thus degrading small signal resolution" and upsetting depth perception Head-Fi, Jan 2016. In February 2019 he stressed that the effect needs current flowing through the DAC's ground plane: "No current flow, no problems" Head-Fi, Feb 2019. In the 2024 seminar he described this as current flowing in a loop from the mains through the source, the cable ground and the DAC and back to the mains MFG Audio Reviews 17:23MFG Audio Reviews 18:03, and in a podcast recorded in October 2025 (published March 2026) he again described circulating RF currents in the ground loop between a computer's mains and the DAC's mains over USB ecoustics 1:07:21.

He ties the size of the effect to the source's processing. In the 2024 seminar he said processor activity produces bursts of correlated gigahertz noise that is worst for small signals, and that "a mobile phone is going to be better than a big processor in a large computer" MFG Audio Reviews 20:22MFG Audio Reviews 22:11. On why galvanic isolation does not remove it completely, he pointed in 2016 to the isolator's 2 pF coupling capacitance, which he put at nearly 80 ohms at 1 GHz Head-Fi, Jun 2016Head-Fi, Jul 2016.

How he says to judge a source

May 2016, DAVE: "When you reduce RF noise, things sound softer and warmer and smoother", and with a bit-perfect source "the only thing that can make a difference to the SQ is RF noise, so go for the warmer and softer sound" Head-Fi, May 2016. May 2017, Hugo 2: "assuming the source is bit perfect, then if it sounds brighter it most definitely is worse" Head-Fi, May 2017. May 2018, Hugo 2: "the source that gets closer to the sound of optical is the best one" Head-Fi, May 2018. September 2020: use a phone playing from local memory as the reference and choose a streamer that comes close to it Head-Fi, Sep 2020. December 2022: "Use a mobile phone, or a battery powered laptop as your reference source, with of course an app that provides bit perfect data" Head-Fi, Dec 2022.

2014 to 2015: Hugo, before galvanic isolation

In May 2014 he wrote of Hugo "I prefer optical, as it normally sounds smoother and warmer" Head-Fi, May 2014; in July 2016 he said "Before galvanic isolation, optical was so much better than USB" Head-Fi, Jul 2016. In June 2014 he wrote: "I think Hugo works well whatever the source - so long as it is bit perfect" Head-Fi, Jun 2014. In October 2014, of Hugo: "The source is not too important - if you use optical, as RF noise and signal correlated ground currents are eliminated - assuming the source is bit perfect" Head-Fi, Oct 2014, and that on the digital inputs "the only mechanism that can upset the sound is RF noise and correlated noise being injected into Hugo's ground plane" Head-Fi, Oct 2014. In an April 2015 StereoNET interview he said phones were less of a problem than PCs and laptops: "I was hearing huge differences between laptops as a source. Now with galvanic isolation you can't hear the difference" article, Apr 2015. That paragraph does not name the product with galvanic isolation.

2016 to 2018: DAVE, Hugo 2 and the M Scaler

In January 2016: "Dave is insensitive to the digital source, assuming the data is bit perfect", with the caveat that the rest of the system may not be Head-Fi, Jan 2016; and "Now with Dave I can no longer hear which source is connected" Head-Fi, Jan 2016. In April 2016, after a USB cable comparison on headphones: "So far all of my tests have revealed that the source makes no difference to the sound, but this could be very system dependent" Head-Fi, Apr 2016. In May 2016 his advice was "to use a source that is convenient for you", having "failed to hear any significant difference between different sources", qualified with YMMV Head-Fi, May 2016.

The next day he reported that his laptop on battery sounded smoother through DAVE than on mains power: "It was a worthwhile improvement, but not a change that is large" Head-Fi, May 2016. In June 2016 he traced most of this to residual coupling through the USB isolation rather than the mains, and wrote that "Battery operation and Dave's galvanic isolation means effectively perfect RF noise isolation from the source" Head-Fi, Jun 2016Head-Fi, Jun 2016. In July 2016 he put the scale in context: before galvanic isolation "I used to get huge differences from sources if they were mains powered - now we are talking about a small change that is only audible on AB tests" Head-Fi, Jul 2016. In April 2018 he called the battery-laptop gain "too small a difference to worry about" in practice Head-Fi, Apr 2018. On USB versus optical on DAVE, in July 2016 he said USB "has zero jitter/timing issues as the clock comes from Dave" and that "Detail resolution and flow is slightly better than optical" Head-Fi, Jul 2016. In August 2016 he called the difference between optical and USB "in my bottom scale of SQ differences", said that in everyday listening "they all work identically from the musicality POV", and still credited USB with "slightly better timing" Head-Fi, Aug 2016.

On players and operating systems, in August 2016 he said Chord's Windows driver resends faulty packets, so bit-perfect data at the DAC is guaranteed, while driverless iOS, Linux and Android do not have this; with DAVE and a battery-powered, ungrounded Windows laptop he would be "very surprised if there were differences on any capable bit perfect program" Head-Fi, Aug 2016. In May 2020 he said "I don't think data integrity plays any part in this", noting that "a faulty data stream is easily audible" Head-Fi, May 2020.

With Hugo 2 he reported in 2017 and 2018 that on headphones he could not tell optical from USB with his MSI laptop, but heard a small difference once the analog output was grounded through a power amp; he warned that "one swallow does not a Summer make" Head-Fi, May 2017Head-Fi, Jul 2017Head-Fi, Dec 2017Head-Fi, May 2018. In January 2018 he wrote that he had thought USB had the edge on optical on DAVE "but I was wrong about that": the extra brightness was RF noise pick-up Head-Fi, Jan 2018; the day before he had written "I have changed my mind on this now" Head-Fi, Jan 2018. In July 2018, with an M Scaler and TT2 on headphones, he heard "no consistent difference at all" between his laptop on battery and on mains Head-Fi, Jul 2018.

On servers, in May 2018 he said they sound brighter, "classic hallmarks of RF noise generation", and that the problem is less DAVE than a power amp sensitive to RF Head-Fi, May 2018. In September 2018: "A battery powered bit perfect source supplies the ultimate digital source" Head-Fi, Sep 2018. In November 2018: "But what do servers bring to the party? Nothing - in my experience it's just convenience" Head-Fi, Nov 2018.

On CD players, in July 2018 (Qutest thread) he suggested that different-sounding players over optical most likely had "different error rates", meaning the data may not be bit perfect, and that ripping gives error-free, bit-perfect playback Head-Fi, Jul 2018. In November 2018: "If the optical source is bit perfect, in my experience it always sounds and measures the same, when you are using Hugo 2 or any other Chord DAC" Head-Fi, Nov 2018.

2019 to 2022: "not quite there yet"

In March 2019 (Qutest thread) he wrote that his job is to make the quality of a bit-perfect source irrelevant, "and I am not quite there yet as sources still do make a difference" Head-Fi, Mar 2019. In the same post he said that adding complexity to the source creates ground loops, and that the result can be mistaken for an improvement Head-Fi, Mar 2019. In October 2019, for Hugo 2 on battery into headphones: "theoretically there is no mechanism for a battery powered source to create noise floor modulation via RF noise", which he said listening tests confirmed Head-Fi, Oct 2019.

In August 2020 (DAVE thread) he reported that "My MSI laptop beat it easily" against a server, the server sounding bright with false transparency Head-Fi, Aug 2020, and wrote that "bit perfect servers can only affect the sound on my DACs by changing the RF noise levels within the DAC"; the best-sounding source "may come from a streamer. Or an isolated lap-top", whichever injects the least RF noise current Head-Fi, Aug 2020. Later that month (DAVE thread) he wrote that for the amount of RF noise injected into the DAC "optical wins big time" over USB, since USB decoding involves heavy processing and an optical receiver none Head-Fi, Aug 2020. In September 2020 he guessed that a good bit-perfect streamer with optical outputs would match a phone's sound for files up to 192k Head-Fi, Sep 2020.

On network gear, in January 2020 he warned that a mains-powered USB-to-optical converter can create an RF loop, and that "Ethernet has galvanic isolation, but there will be capacitance coupling through the isolators" Head-Fi, Jan 2020. In an October 2020 post (date estimated from the post ID) he wrote: "I have to admit not having any experience of network switches" Head-Fi, Oct 2020. In March 2021: "Adding extra boxes will almost certainly degrade the sound by increasing overall RF", and "if you cannot hear a difference from network switchers then do not worry!" Head-Fi, Mar 2021.

In March 2021 he explained a DAVE owner's different-sounding CD players over optical as mains-borne RF noise picked up by the amps: "I do not think it is Dave that's causing the issue" Head-Fi, Mar 2021. In April 2021, to a Qutest owner, he wrote that optical is immune to RF but a CD player still creates RF noise and injects it into the mains, where the DAC or amps pick it up Head-Fi, Apr 2021. Also in April 2021 (TT2 thread, date estimated) he found optical much better than USB from a new high-powered desktop PC, concluding "So for sure USB can be identical to optical, but with nosier sources, optical will win out" Head-Fi, Apr 2021; he repeated this in February 2022 Head-Fi, Feb 2022. At CanJam NYC 2022 (seminar video published March 2022) he said galvanic isolation is about 99% effective, not 100%, and that Mojo and Hugo 2 do not have it MFG Audio Reviews 56:58.

In May 2021 (Hugo 2) he said USB sounds identical to optical when everything is battery powered, that a mains-powered source feeding a Hugo 2 that drives a mains-powered amp lets RF currents flow, and advised avoiding digital-to-digital converters Head-Fi, May 2021. In November 2021: "Now I am not saying that bit perfect sources don't upset/change the sound; but the perfect DAC would have no difference in SQ whatever the source of bit perfect data - and that's my focus" Head-Fi, Nov 2021. In March 2022 (DAVE) he restated the 2018 reversal: "In 2016 with Dave I did indeed prefer the brightness of USB", and "the \"benefits\" of clocking from Dave was an illusion" Head-Fi, Mar 2022. The next day he added that "This does not mean that all USB sources sound the same as optical, as the galvanic isolation on the USB is not perfect" Head-Fi, Mar 2022, and later that day: "There is no possibility of USB performing better than optical" (apart from optical's 192k limit), and making DACs immune to source noise is "a very much bigger challenge than one might think" Head-Fi, Mar 2022. In October 2022 (TT2 thread): "I too used to prefer USB than optical", and "there is no technical benefit to using USB over optical when using my DACs" Head-Fi, Oct 2022. That differs from his February 2019 remark that USB is technically better from the clock POV; both are his dated statements. In July 2022: "my goal is so that any bit perfect digital source will sound identical" Head-Fi, Jul 2022.

2023 to 2026: the goal restated

In February 2023, in Bucharest: "Currently, with my DACs, they do not sound the same. You will have different streamers sounding different" Dusty Audio 48:52, adding that nanovolts of RF noise reaching the DAC degrade performance Dusty Audio 51:22. The next day he said that with bit-perfect data "the difference in the sound is not the streamer, it's actually the sensitivity of the DAC" Soundnews 10:21.

In November 2023 (DAVE thread) he suggested optical from a streamer in place of USB: "you will either get better SQ, or the same SQ if RF noise on USB is low"; he added that he runs optical from his Windows PC with Qobuz and gets better sound quality Head-Fi, Nov 2023.

In January 2024 he wrote that he had found "you couldn't beat a battery powered lap-top with optical (or mobile phone in flight mode with locally stored data) when using bit-perfect data", and that his focus is "making the DAC immune from source generated RF noise" Head-Fi, Jan 2024. In February 2024 (Mojo 2): "A mobile phone or battery powered DAP connected to Mojo 2 should sound the same as optical", assuming bit-perfect data that has not been processed Head-Fi, Feb 2024. In March 2024: "don't discount your mobile phone as the best source you can get" MFG Audio Reviews 22:11.

In July 2024, on the Quartet M Scaler, he said he had further isolated it against source noise and that "Time will tell when people use it in different systems with different sources, whether I've been successful" HIFICLUB 47:23HIFICLUB 48:49. In an interview recorded at a Munich show, most likely May 2024 (published February 2025), asked whether local-network gear (router, switches, cables) affects the sound, the transcript records "So yeah, it does", followed only by his practice of testing from files on a local SSD; the recording should be checked before relying on that line audio-2G 11:52audio-2G 11:57. In the same interview he said Qobuz and his local copy sound the same in practical terms when they are the same recording, and that with a perfect DAC fed bit-perfect data the source "would make no difference to the sound quality" audio-2G 12:31audio-2G 12:43audio-2G 13:29. In an Audio Force Thailand interview published May 2026 (recording date not stated): "My intent is to make them sound the same. And this I think we can do" Audio Force 18:19. In an interview published October 2026 (its transcript places the recording before May 2026; recording date not stated) he recalled that Head-Fi reports of USB and different sources sounding different led him to study how source noise changes the sound, and to galvanic isolation on DAVE, "the first product that had galvanic isolation on it" HiFi Buys 2:22:57HiFi Buys 2:23:47. The "first product" wording sits against his own earlier statements: in September 2014 he said the new QBD "has galvanic isolation built in" Head-Fi, Sep 2014, and in April 2015 he named galvanic isolation on the USB as a change in Hugo TT article, Apr 2015, both before DAVE, whose production he said in June 2015 was scheduled for October 2015 Head-Fi, Jun 2015. See the galvanic isolation answer.

The system matters

He repeatedly says the result depends on the rest of the system. In 2016 he wrote that DAVE being insensitive "doesn't mean the rest of your system isn't" and that most power amplifiers are very RF sensitive Head-Fi, Jan 2016Head-Fi, May 2016. With Hugo 2 he found headphone-only use and a grounded output gave different results Head-Fi, Jul 2017Head-Fi, Oct 2019, and he told owners to "take my advice with a very big pinch of YMMV" Head-Fi, May 2018.

What he has not said

He gives no figures for how much a particular streamer, server or switch changes a Chord DAC's output. In May 2018 he said he had measured RF noise's effect "many times and correlated it to SQ" Head-Fi, May 2018, and in August 2020 that a newer Dell laptop "significantly degrades measurements on my APx 555 test gear" Head-Fi, Aug 2020, but his source comparisons are listening tests. He said in 2020 he had no experience of network switches, and in 2021 that "I have zero experience of using Apple myself" Head-Fi, Oct 2020Head-Fi, Mar 2021. Chord's Poly and 2go were not his projects Head-Fi, Sep 2020Head-Fi, Oct 2021. On other makers' DACs, he wrote in October 2019 that "all DACs are sensitive to ground currents flowing into the DAC ground plane" Head-Fi, Oct 2019 and in May 2020 that other DACs are likely to be jitter sensitive Head-Fi, May 2020; otherwise his statements about how sources sound are about Chord DACs.

Quotes

“assuming bit perfect data, the source can only change the sound by adding RF noise into the DAC”2019-02-09 Head-Fi, Feb 2019

“It is only noise floor modulation that can affect sound quality with bit perfect sources.”2021-05-31 Head-Fi, May 2021

“assuming the source is bit perfect, then if it sounds brighter it most definitely is worse.”2017-05-17 Head-Fi, May 2017

“Currently, with my DACs, they do not sound the same. You will have different streamers sounding different, and there are various technical reasons why that is the case.”2023-02-04 Dusty Audio 48:52

Dated positions

Caveats and open points

• In the Munich-show interview (most likely May 2024, published February 2025) the transcript's 'So yeah, it does', in answer to whether network gear affects the sound, is followed only by his local-SSD practice; it is not clear from the transcript what he was agreeing to.
• Head-Fi, Oct 2020 and Head-Fi, Apr 2021 carry dates estimated from their post IDs.
• He qualifies source findings by system: headphone-only use versus a grounded output into a power amp, mains loops, and battery power, often with YMMV.
• The April 2015 StereoNET paragraph about galvanic isolation does not name the product it refers to.
• His 2018 CD player remarks concern players that may not deliver bit-perfect data (error rates), so they sit at the edge of this question.
• Network gear is covered only by general remarks about extra boxes, Ethernet isolation capacitance and his lack of experience with switches; there is no test of specific network equipment.
• His USB versus optical statements differ by date: the 2016 USB preference on DAVE, the 2018 reversal, the February 2019 clock remark and the 2022 retractions are all quoted above rather than reduced to one view. The HiFi Buys interview (published October 2026) and the Audio Force Thailand video (published May 2026) have no stated recording dates.

Topics: Source devices, streamers and networks, RF noise: mechanisms and audibility, Ground loops, galvanic isolation, earthing and mains