Why do passive components, inductors and ferrites in the signal path affect sound even when they measure perfectly?
Short answer
Watts's stated reason is that the effect sits below what is normally measured. He says metal-to-metal contacts, solder joints and some passive parts add tiny small-signal non-linearities that degrade depth and detail resolution Head-Fi, Oct 2015Head-Fi, Mar 2017. He says ferrite and cored inductors change inductance with signal current, so phase and transient timing shift with level Head-Fi, Feb 2018Head-Fi, Apr 2021. In March 2024 he told a CanJam audience that in a lot of cases passive components do not measure perfectly, but a lot measure perfectly and still change the sound MFG Audio Reviews 1:32. He reports measuring contact distortion only into low-impedance loudspeaker loads (a corroded load switch, crimped connections) and failing to measure it on interconnects Head-Fi, Mar 2017Head-Fi, Apr 2018Head-Fi, Jun 2021. For analog circuits he said no number can be put on it Head-Fi, Apr 2022. His design answer is to keep the analog path as simple as possible Head-Fi, Apr 2014Head-Fi, Apr 2022.
Where he starts (2014 to 2018)
In April 2014, describing the Hugo, Watts wrote "You really can hear every solder joint, every passive component, and every active stage", and that Hugo has "two resistors and two capacitors in the direct signal path" Head-Fi, Apr 2014. In July 2015, in the DAVE thread, he wrote "every passive component is audible" and that metal-to-metal interfaces, solder joints included, have an impact Head-Fi, Jul 2015.
In October 2015 he gave a mechanism. He put it down to a layer of oxides and impurities on metal contacts: small signals have to break through it and larger ones do not, so very small signals are attenuated Head-Fi, Oct 2015. In the same post he said he had "only ever measured this problem once" (a digital power amp whose load-switch contacts had corroded) and that "the brain can detect levels that no measuring equipment can detect" Head-Fi, Oct 2015. In December 2015 he listed RF characteristics, dielectric absorption, small-signal non-linearity, resistor non-linearity and resistor thermal non-linearity as all influencing sound, and said Chord products are reflow soldered in an inert nitrogen atmosphere because soldering makes a big difference Head-Fi, Dec 2015.
In March 2017 he said his cable-design work in the past showed "small signal non-linearity was crucial for depth and detail resolution". He added "for interconnects I have failed to measure any kind of distortion, but nonetheless it is very audible", and admitted "I do not know why this is, and its a puzzle as to why the ear/brain is so sensitive" Head-Fi, Mar 2017.
In April 2018 he separated the cases. With 8 ohm loads, he said, "dirty contacts are easily measurable in terms of distortion" Head-Fi, Apr 2018, but he had "not seen directly any measurement issues with interconnects when it is just a milli-amp or so of signal current flowing" Head-Fi, Apr 2018. At the Nintronics open day talk of November 2018 (uploaded December 2018) he said "it doesn't matter how small the error is, it's audible", and applied the same reasoning to cables that show no measurable difference Pursuit Perfect System 29:21.
Inductors and ferrites
His case against inductors and ferrite cores in the analog signal path rests chiefly on one mechanism, which he states from 2016 to 2024. In April 2016 he said a delay that changes with signal level "can occur with cored inductors", calls the result phase intermodulation distortion, and said reducing it "makes it sound smoother and darker, with better instrument separation and focus" Head-Fi, Apr 2016. In July 2017 he said "amplitude phase non-linearity is highly audible - this occurs for example with inductors, whose inductance value changes with current", noted that passive loudspeaker crossovers have this problem, and said replacing passive components with digital direct drive was on his road-map Head-Fi, Jul 2017. In February 2018 he wrote that "All ferrites and inductors are non-linear", so phase shift changes with signal level, which he called phase intermodulation distortion and said is audible because transient timing becomes amplitude dependent Head-Fi, Feb 2018. In the same post he said he treats RF on the output "without using inductors or ferrites, as these in the past has had a problem with SQ" Head-Fi, Feb 2018.
In April 2021 he wrote "It's a very bad idea to use ferrites and or inductors in the analogue signal path", citing inductance that varies with current, the resulting level-dependent phase shift, and hysteresis, and saying instrument separation, focus, timbre variation and soundstage get degraded Head-Fi, Apr 2021. In October 2021, discussing Zobel networks on power-amp outputs, he said an air-cored inductor "doesn't degrade the SQ too much" while a ferrite core "definitely degrades performance" Head-Fi, Oct 2021. In April 2022 he added that for inductors the issue is high signal currents having a different phase shift from small ones Head-Fi, Apr 2022. In March 2024 he said inductors are devices "you shouldn't use in the audio signal path at all, particularly ferrite core inductors" MFG Audio Reviews 52:45, and "you should never use inductors in the signal path or ferrite beads in the signal path" MFG Audio Reviews 53:30. Asked about phase shift, he said "Air-core inductors are okay because they're perfectly linear" MFG Audio Reviews 55:28. That answer sits beside his February 2018 wording that "All ferrites and inductors are non-linear" Head-Fi, Feb 2018; in October 2021 he had put air-cored inductors in between, as not degrading sound "too much" Head-Fi, Oct 2021.
He sets limits on where this applies. In February 2018 he wrote "Ferrites or inductors on digital signals have no downsides" Head-Fi, Feb 2018. In April 2021 he said ferrites and inductors on a power supply or mains "is not in the direct signal path, so can be used without worries of distortion" Head-Fi, Apr 2021, and that ferrites and RF filters on the digital side are there to stop RF noise creating noise floor modulation in the analog sections Head-Fi, Apr 2021. In May 2019 he said Hugo 2's USB ground runs through a ferrite bead Head-Fi, May 2019 that "has no significant effect at audio frequencies, but will reduce RF currents flowing as the impedance is higher at RF" and "may or may not have a SQ benefit" Head-Fi, May 2019. For BNC cables carrying digital data from the M Scaler he recommended solid ferrite cores with a resonant frequency above 2GHz in April 2020 Head-Fi, Apr 2020. In April 2014, about RF pickup on a Hugo headphone lead, he suggested trying "cylindrical ferrite cores" Head-Fi, Apr 2014; in February 2018 his advice on RF treatment for headphone cables was "to be extremely careful" Head-Fi, Feb 2018. These are different situations, and the posts do not say he changed his view.
Capacitors and resistors
On capacitors he points to dielectric absorption. In January 2018, about Hugo 2, he said AC coupling capacitors "can add considerable bloom to bass, particularly electrolytic capacitors", and that Hugo 2 has "no coupling caps at all (with the digital DC servo)" Head-Fi, Jan 2018. In July 2017 he attributed Mojo's fuller bass to "the use of a single coupling capacitor" Head-Fi, Jul 2017; in January 2022 he said removing it in Mojo 2 was the big change and that it "gave Mojo it's distinctive warmth" Head-Fi, Jan 2022. In August 2018 he said "The coupling caps make a huge difference to SQ" and also said switching the digital DC servo on and off made no difference to sound Head-Fi, Aug 2018.
The March 2024 CanJam talk speaks most directly to measuring perfectly. On dielectric absorption in tantalum and electrolytic capacitors he said "You're not going to measure it because it's a dynamic issue", noting measurements use a steady sine wave MFG Audio Reviews 51:15. On resistors, he said surface mount types "with everything else being equal, sounds better than leaded resistors" MFG Audio Reviews 49:42, and that resistor self-heating modulating the signal is "surprisingly quite audible" MFG Audio Reviews 49:42MFG Audio Reviews 50:30. In March 2022 he called MELF resistors "hideously undesirable" because of internal pressure contacts Head-Fi, Mar 2022. In February 2023 he said "Every passive component, a resistor, a capacitor, will degrade the sound quality" Soundnews 13:58.
What he does about it
His stated approach is to keep the analog path short and simple. At CanJam NYC 2022 (late February; video published March 2022) he described Mojo 2 as having "one amplifier, four components" in the direct path MFG Audio Reviews 14:49. In August 2022 he said the balanced output on TT2 measures the same as single ended but he still prefers single ended because of added passive components and soldered joints in the feedback path Head-Fi, Aug 2022. In a post dated between March 2025 and March 2026 about the TT2 gain relay he said "parallel solder joints are perfectly fine, but series soldered joints degrade depth, detail resolution and transparency" Head-Fi, Mar 2025. At CanJam NYC 2023 (late February; video published March 2023) he said "every time you have a metal to metal contact, your small signal gets degraded" The Headphone Show 48:23. In a podcast recorded in October 2025 (published March 2026) he repeated the oxide explanation for cables, saying "small signals get attenuated more than large signals" ecoustics 46:39. In an Audio Force Thailand video published March 2026 (recording date not stated) he said Mojo 2's built-in 4.4 mm connector sounds better than an adapter because "every metal-to-metal interface damages the sound quality" Audio Force 32:33; in another published May 2026 (recording date not stated) he said an adapter "upsets detail resolution and sound stage depth" Audio Force 21:35. In the Soundnews interview (recorded autumn 2026, published October 2026), about a planned Pro ADC, he described analog mixing desks as having "lots of contacts, lots of amplification, lots of capacitors, all of them degrading the sound quality" article, Oct 2026.
Limits he states
In August 2017, comparing Hugo 2 with chip-DAC designs, he said noise floor modulation, jitter sensitivity, small-signal resolution, analog complexity and timing reconstruction "are thousands of times more important than the quality of the passive components", adding that by better analog for Dave he did not mean "simply" the quality of "a handful of passives" Head-Fi, Aug 2017. In April 2022 he said "it's not possible to directly put a number onto these issues, unlike in the digital domain" Head-Fi, Apr 2022. In March 2024 he told listeners to do listening tests rather than assume MFG Audio Reviews 53:45, and said that in many cases passive components do not measure perfectly MFG Audio Reviews 1:32. In August 2022 he wrote "Now measurements are very important - they tell you if something is grossly wrong - but I will never sacrifice sound quality just to improve measurements" Head-Fi, Aug 2022.
What he has not said: He gives no measurement of the analog small-signal effect on an interconnect (he says he failed to measure it Head-Fi, Mar 2017), no figure for how small a non-linearity in a passive part is audible in the analog domain Head-Fi, Apr 2022, and no statement of whether his passive-component comparisons specifically were blind. In general terms, in September 2018 he said he often runs a single blind test "when the listening test reveals something extraordinary" Head-Fi, Sep 2018.
Quotes
“It's a very bad idea to use ferrites and or inductors in the analogue signal path - they have significant SQ problems.”2021-04-02 Head-Fi, Apr 2021
“A lot of passive components measure perfectly, but they still change the sound quality.”2024-03 MFG Audio Reviews 1:32
“But for interconnects I have failed to measure any kind of distortion, but nonetheless it is very audible.”2017-03-04 Head-Fi, Mar 2017
“soldered joints, and passive components all degrades small signal accuracy (impossible to measure) which subtly degrades depth and detail resolution”2022-08-26 Head-Fi, Aug 2022
Dated positions
| When | Products | Position |
|---|---|---|
| 2014-04 to 2015-12 | Hugo, DAVE | He says every solder joint and passive component is audible, that oxides at metal contacts attenuate very small signals, and that Hugo keeps only two resistors and two capacitors in the direct signal path. Head-Fi, Apr 2014Head-Fi, Jul 2015Head-Fi, Oct 2015Head-Fi, Dec 2015 |
| 2017-03 to 2018-04 | General | He says small-signal non-linearity from contacts is audible but not measurable on interconnects, while dirty contacts are measurable as distortion into 8 ohm loads. Head-Fi, Mar 2017Head-Fi, Apr 2018Head-Fi, Apr 2018 |
| 2016-04 to 2018-02 | General | He says cored inductors and ferrites make delay and phase depend on signal level, causing phase intermodulation distortion, and treats RF on the output without inductors or ferrites; ferrites on digital signals he says have no downsides. Head-Fi, Apr 2016Head-Fi, Jul 2017Head-Fi, Feb 2018 |
| 2018-01 to 2022-01 | Mojo, Mojo 2, Hugo 2 | He attributes bass bloom to coupling capacitors and dielectric absorption, and says Hugo 2 and Mojo 2 avoid coupling capacitors (Mojo 2 removed the one Mojo had). Head-Fi, Jan 2018Head-Fi, Jul 2017Head-Fi, Jan 2022 |
| 2021-04 to 2021-10 | General | He calls ferrites and inductors in the analog signal path a very bad idea, says air-cored inductors are much less harmful, and says ferrites and inductors on power supplies or digital signals are not a problem. Head-Fi, Apr 2021Head-Fi, Apr 2021Head-Fi, Oct 2021 |
| 2022-03 to 2022-08 | Mojo 2, Hugo TT 2 | He calls MELF resistors undesirable, says added passive parts and solder joints degrade small-signal accuracy in a way that is impossible to measure, and says no number can be put on the analog effect. Head-Fi, Mar 2022Head-Fi, Aug 2022Head-Fi, Apr 2022 |
| 2024-03 | General | At CanJam NYC he said a lot of passive components measure perfectly yet change the sound (while many do not measure perfectly), citing capacitor dielectric absorption, leaded resistors and ferrite-core inductors, called air-core inductors okay, and said to rely on listening tests. MFG Audio Reviews 1:32MFG Audio Reviews 51:15MFG Audio Reviews 55:28MFG Audio Reviews 53:45 |
| 2025-03 to 2026-10 | Hugo TT 2, Mojo 2 | He says series solder joints degrade depth, detail resolution and transparency while parallel ones are fine, repeats the oxide explanation, says every metal-to-metal interface (such as a headphone adapter on Mojo 2) damages sound, and lists analog contacts and capacitors as degrading sound. Head-Fi, Mar 2025ecoustics 46:39Audio Force 32:33article, Oct 2026 |
Caveats and open points
- He gives no numeric threshold or published measurement for the analog effect on interconnects or passive parts; the only figures he ties to this explanation come from digital-domain simulation of his noise shapers, which he applies by analogy.
- Head-Fi, Mar 2025 has no exact date; this archive places it after 2025-03-27 and before the 2026-03-03 scrape.
- Ferrite statements cover different situations (analog signal path, digital lines and BNC cables, power supplies, headphone cables, USB ground); they are not one rule applied everywhere.
- The two Audio Force Thailand videos (published March and May 2026) give no recording dates, and both discuss the Mojo 2 4.4 mm connector, so they may not be separate occasions. The CanJam NYC 2024 talk was given on 9 or 10 March 2024; the video was published on 20 March 2024.
Topics: Passive components and physical construction, Measurement vs listening