I’m not sure if my LLM is better, but I use it for work (GPT-5 Pro).
I showed it screenshots of Perplexity’s response in the OP and prompted it in the following way:
“Why does Tidal with my audio setup sound much better than Qobuz even though there seems to be a preference for Qobuz on the dCS forum? In fact, it’s the exact opposite with my system than what Perplexity says here: „Qobuz generally does not apply loudness normalization within its own app, so track volumes can vary more, preserving more of the original dynamic range. Users often find Qobuz streams sound more detailed, alive, and with more natural bass response compared to Tidal, which applies more compression and loudness normalization by default, potentially making Tidal sound slightly more compressed and less dynamic.“
Below is a description of my audio system. Respond in continuous text in as many paragraphs as necessary in order to be comprehensive and cohesive and provide in-text citations of the sources used.
My audiophile setup is centered around a dCS Rossini digital front end — comprising the Rossini DAC and the Rossini Clock — feeding a WestminsterLab Quest preamplifier equipped with the carbon-fiber shielding option for superior noise isolation. The signal then drives a pair of Avantgarde Acoustic ZERO iTRON fully active loudspeakers, known for their highly neutral, studio-monitor-like frequency response and internal amplification, eliminating the need for external power amps.
All signal paths are wired with Cardas Clear Beyond XLR interconnects, used both between the DAC and preamp and between the preamp and speakers, ensuring a consistent and top-tier analog connection. Power delivery is equally meticulous: five Cardas Clear Beyond Power XL cables provide current to the Rossini DAC, Rossini Clock, the preamp, and each of the two active speakers, minimizing electrical noise and maintaining tonal purity.
The digital chain is optimized with an Innuos Phoenix Net audiophile network switch and a Roon Nucleus serving as the playback and library control hub, guaranteeing low-jitter, bit-perfect streaming and seamless integration within the Roon ecosystem. Power conditioning and stabilization are handled by an Audes 3000DC, ensuring clean, stable current across all components.
The system is placed in a living/dining room environment with no formal acoustic treatment apart from a rug positioned in front of the speakers, yet it is optimized for low-volume listening with exceptional coherence and clarity — a setup designed to deliver studio-grade precision and musical engagement even at modest sound levels.”
Below the response
It’s absolutely plausible that Tidal sounds better than Qobuz on your system. What reaches your Rossini isn’t “the service” in the abstract but the exact master, the app/transport settings, and how the Rossini is configured at the sample rate that’s actually playing—small differences that add up audibly on a highly resolving chain like Rossini + Clock → WestminsterLab Quest → Avantgarde ZERO iTRON.
Different program material and recent format changes. Many titles are not identical across services, and in mid‑2024 Tidal removed MQA and moved to Hi‑Res FLAC across its apps and integrations. If you’d compared MQA versions on Tidal to FLAC on Qobuz in the past, or if the two services carry different editions of an album today, you’ll hear differences that have nothing to do with “streaming quality” per se. Tidal confirms that MQA playback was discontinued on July 24, 2024. Tidal Support
In a Roon setup, Roon—not the service app—decides loudness handling. Roon’s Volume Leveling is based on the EBU R128 standard and can be switched Off/Auto/Track/Album per zone. When it’s on, Roon applies gain changes using its DSP engine; when it’s off, playback can be completely unaltered. Roon also uses loudness metadata supplied by Tidal and Qobuz when available, which means the services’ own app defaults don’t apply once Roon is in the loop. In practice, even a couple of dB of leveling can shift punch, space, and bass weight—especially at the modest listening levels you prefer. Roon Labs Help Center+2Roon Labs Help Center+2
How to tell if anything is touching the bits. Roon’s Signal Path shows a purple Lossless indicator when the stream is unmodified end‑to‑end; a blue Enhanced indicator appears as soon as you enable processes like Volume Leveling, upsampling, or other DSP. That visual check is the quickest way to ensure you’re doing apples‑to‑apples comparisons between services. Roon Labs Help Center
Tidal’s own app settings only matter if you use the Tidal app. If you ever play through the desktop app instead of Roon, enable Exclusive Mode (and typically Force Volume) so the app takes exclusive control of the audio device—this avoids the operating system altering sample rate or gain and helps ensure bit‑perfect output to your DAC. Tidal Support
Low‑level listening accentuates small tonal/punch differences. At low volumes our ears are less sensitive to bass and treble than to midrange (equal‑loudness contours, ISO 226). A master that’s 1–2 dB hotter overall or a touch fuller in the low end will sound more “complete” at the levels you favor—so the Tidal version can legitimately feel more alive even when both are technically lossless. Wikipedia
Rossini behavior can differ with sample‑rate families—and that’s audible. The Rossini keeps your last‑used PCM filter per sample rate, and the available filters aren’t identical across all rates (e.g., filters 1–4 work across the full rate range, while others are rate‑specific). If Tidal serves 44.1/48 kHz and Qobuz serves 96/192 kHz for a given title, the DAC may be using different filters (and even different Mapper choices, if you’ve experimented), subtly changing transient character and staging. That alone can tilt preference toward one service in your room at your volume. ManualsLib+1
RAAT is designed for bit‑perfect delivery—unless you ask Roon to process. Roon’s transport (RAAT) can deliver bit‑perfect streams to Roon Ready endpoints; again, watch the Signal Path color to verify nothing is being changed unintentionally. Roon Labs Help Center+1
Room interaction often outweighs tiny electronic differences. Research led by Floyd Toole shows that the loudspeaker–room system is a dominant factor in what we perceive, particularly in small domestic rooms and especially in the bass where room modes prevail. In a lively living/dining room with minimal treatment, masters with slightly stronger bass or presence will tend to integrate more convincingly at low level—which can make one service’s version consistently more satisfying. Cieri
Territorial catalogs can complicate “same album” comparisons. Qobuz ties accounts to a country and operates in specific regions; availability and versions can vary by territory due to licensing. That means you may not always be hearing the exact same edition from both services (an inference from Qobuz’s own region‑linked account model). Qobuz Help Center
Practical checks for your setup. In Roon, confirm per zone that Volume Leveling and Crossfade are off when you’re comparing, and that the Signal Path is purple Lossless. Then compare the same release/edition on both services and level‑match carefully. Watch the sample rate on each service; if they differ, try setting Rossini’s PCM filters intentionally for the 44.1/48 family versus the 88.2/96/176.4/192 family to see which combination best preserves low‑level punch in your room. If you sometimes use the Tidal desktop app directly, enable Exclusive Mode/Force Volume for bit‑perfect output. Tidal Support+3Roon Labs Help Center+3Roon Labs Help Center+3
Finally, about the oft‑repeated claim that “Tidal is more normalized, Qobuz is not”: Tidal did make headlines years ago for adopting album‑based normalization around −14 LUFS, but in a Roon‑based system it’s Roon’s EBU‑R128 implementation that governs loudness unless you turn it off. That’s why experiences like yours can legitimately run counter to forum preferences—even tiny differences in master/level/filtering become obvious on a Rossini‑class front end at modest SPLs. audioXpress+1
In short, your observation isn’t a contradiction of the dCS‑forum zeitgeist; it’s a consequence of program‑to‑program and setting‑to‑setting variability interacting with your DAC’s filter behavior and your low‑volume, real‑room listening context. When you equalize those variables for an apples‑to‑apples test, you’ll quickly hear whether Tidal is winning because its masters suit your room and preferences—or whether a small tweak in Roon or Rossini lets Qobuz catch up.
