Hi-res audio explained – formats, technology, advantages & practical tips for the best sound

Today, Hi-res audio is a collective term for digital audio files and services that work with sampling rates and word widths that are higher than the historical CD standard of 16 bit and 44.1 kHz. The technical discussion is not only about higher numbers – 24 bit or 32 bit, 48 kHz, 96 kHz, 192 kHz and even up to 768 kHz or DSD formats – but also about the question of what acoustic information is actually retained, how it is created and how reliably it can be transported through the chain from the master to the loudspeaker.

It is important for consumers to understand that Hi-res audio is not a single construct, but an ecosystem of recording processes, file formats, conversion techniques, measurement methods and playback devices. In fact, even the room acoustics in the listening room can have a much greater impact than a higher resolution recording. This description looks at the physical principles, common formats, practical production aspects, measurement and listening tests as well as the specific influencing factors in the playback chain.

What exactly does “Hi-res” mean?

The term “Hi-res” is loosely defined; it is usually understood to mean lossless digital files with a higher resolution than that used as a reference for audio CDs. Two parameters are technically relevant: the sampling rate (samples per second), together with the anti-alias filter used, determines the frequency spectrum that can theoretically be represented, and the word width (bit depth) defines the dynamic range and the resolution of the amplitude measurement.

Higher sampling rates shift the Nyquist limit upwards, higher bit depths increase the possible signal-to-noise ratio and the dynamic range. However, it is important to note that these figures alone do not guarantee that more musical information, more details are preserved or audible – the decisive factors are noise behaviour, linearity, alias artefacts, dither strategies and the actual signal chain during recording and playback.

The audio CD was based on 16 bits and 44.1 kHz, so in principle anything above this is referred to as Hi-res audio – in practice this means at least recordings with 24 bits and 44.1 kHz or more.

Technical basics: sampling rate, bit depth and Nyquist

According to the aforementioned Nyquist-Shannon theorem, sampling at twice the highest contained frequency is sufficient to represent a band-limited signal without loss. In practical terms, this means that 44.1 kHz theoretically captures a frequency up to around 22 kHz and therefore theoretically covers all frequency ranges that the human ear can detect. You might therefore think that this is completely sufficient. But this is not the case.

Higher sampling rates such as 96 kHz or 192 kHz, 384 kHz or even 768 kHz allow the representation of higher frequencies and facilitate digital filtering, reduce phase distortions in the audible ranges through softer filters and reduce alias artefacts.

The so-called bit depth influences the quantisation-related noise and thus the possible dynamic range. While 16 bits allow a theoretical dynamic range of around 96 dB, 24 bits offer a reserve signal-to-noise ratio far beyond this, which is very useful in studio chains – especially in multi-track recording when signals are summed or heavily processed. Whether this also has decisive advantages at the other end of the chain is not quite so clear, however, as opinions sometimes differ, but more on that later.

Formats and codecs: PCM, FLAC, ALAC, DSD, MQA

Content in Hi-res audio is used in different file formats and with different encodings. Firstly, a distinction must be made between two basic technologies, namely content in Linear PCM and DSD.

Pulse Code Modulation or PCM is clearly the most widely used format. PCM-based containers such as WAV and AIFF are uncompressed, FLAC and ALAC offer lossless compression and are therefore popular for distribution and storage.

DSD stands for Direct Stream Digital and represents a completely different concept. It utilises 1-bit sigma-delta modulation with a very high pulse rate of at least 2.8224 MHz (DSD64), while higher-resolution DSD versions even use up to 1024-fold sampling (DSD128, DSD256, DSD512 and DSD1024). Its sound characteristics differ technically from PCM and require special DAC implementations.

MQA is a proprietary approach that promises to transport a ‘studio master’ quality in a smaller file size and at the same time offer an authenticity check. However, MQA is technically and philosophically controversial because it relies on compression, convolution layers and proprietary decoding steps.

The following applies to all formats: lossless formats retain the original bit pattern (after conversion), while lossy formats such as MP3 or AAC and also MQA cause irreversible data loss.

Recording and mastering: How Hi-res audio is created (and sometimes quality is lost)

The benefits of Hi-res audio begin in the studio: miking, preamplifiers, A/D converters, clocking and digital signal processing as a whole determine what information is captured for the first time.

A high-quality A/D converter with clean clocking, low jitter and good linearity is crucial for digitising the analogue signal true to the original. This can also mean that an outstanding A/D converter with a lower resolution sometimes delivers better results than one where the developers were only looking for the most impressive figures possible in the data sheet.

In mastering, there is also the fact that processing such as equalising, dynamic processing or sample rate conversion (SRC) can result in information loss or artefacts.

Many producers prefer to work in 24 bits and higher sampling rates in a Hi-res chain in order to have sufficient headroom and as little interpolation artefact formation as possible when cutting, editing and rendering.

In the end, the mastering target (for streaming, CD or vinyl) often decides how the file is finally reduced or adapted – and this is where dynamics and detailed information are often lost, regardless of the original resolution.

The playback chain: DAC, clocking, analogue stages and speakers or headphones

Even if a Hi-res file contains all the information perfectly, the playback chain must convert the data precisely back into an analogue signal. The digital-to-analogue converter is at the heart of this; design decisions regarding oversampling filters, modulatory architectures (R-2R, delta-sigma), clock systems and analogue output stages influence linearity, noise behaviour and distortion.

Jitter – i.e. temporal discontinuities in the sampling clock – can distort fine phase information and introduce audible artefacts in poor implementations. Analogue amplification stages, the quality of the power supply, cables and finally room acoustics or headphone attenuation also shape the end result. In other words: No link in the chain is dispensable; a ‘lossless’ Hi-res audio file is of little use if the DAC or analogue stage blurs the benefits.

Perception, psychoacoustics and measurement methodology

Whether and under what conditions differences between standard resolutions for CDs and Hi-res audio are audible depends on numerous factors: Signal level, content (e.g. transient-rich percussion vs. quiet voices), listening room, loudspeakers, headphones, listening position, hearing health of the listener…

Laboratory tests sometimes show that trained listeners can recognise differences under ideal conditions, while in real living room conditions and at loud listening levels the advantage often disappears. Measured variables such as THD (total harmonic distortion), IMD (intermodulation distortion), signal-to-noise ratio, dynamic range, frequency response and jitter are objective indicators of the technical quality of a reproduction chain. Sensible evaluation methods combine measurements, controlled double-blind tests and musically relevant listening samples in order to link technical data with subjective perception.

Streaming, distribution and practical limits

Distribution of Hi-res audio has become more widespread with higher bandwidth streaming services, but practical limitations remain: Network stability, client implementation, API limitations, and crypto containers can affect bit-accurate playback.

Some streaming services deliver true bit-perfect Hi-res audio, others transcode or use loudness levelling, which can alter the master. For end users, the question remains as to whether local playback from a high-quality file is preferable to streamed Hi-res audio and how useful resampling or upsampling is on the end device. Rights management and licence models also play a role in the choice of format and distribution.

Is Hi-res audio available for streaming?

Yes, of course you can also use content in Hi-res audio for streaming in your own network. All you need is a high-performance network, suitable streaming clients and, of course, appropriate content that can be made available on a central storage solution in your own network or via a streaming service.

The only actual industry standard in the streaming sector, UPnP, naturally supports content in Hi-res audio, theoretically in all relevant formats, whether Linear PCM or DSD. In practice, however, the streaming client used determines which content with which resolution is actually supported.

Interestingly, Hi-res audio has so far been completely excluded from Apple AirPlay and even Apple AirPlay 2. This streaming technology platform from Apple Corporation is limited solely to content in ‘CD quality’ and thus 16 bit and 44.1 kHz, although Apple Lossless is actually used, which would be capable of much more.

The situation is different with the competition from Google Inc., whose streaming technology platform Google Cast is indeed capable of transmitting content in Hi-res audio, although it relies solely on Linear PCM and has reached the end of the line at 24 bit and 96 kHz.

Practical tips: How to use and evaluate Hi-res audio sensibly

In practical terms, it is advisable not to consider Hi-res audio in isolation, but as part of an optimised workflow: first and foremost, it is important to pay attention to genuine Hi-res audio content; all that glitters is not always gold. Some providers lure you in with high resolutions, while others offer the same album in a lower resolution or not at all as a Hi-res audio download. A clear indication to exercise caution here. It is also important to pay attention to high-quality A/D converters and clean, accurate clocking.

However, it is worth taking a look at the entire signal path for the best quality playback: a clean, low-jitter DAC is only half the battle, the rest of the chain must also be right and even the listening room is crucial. In addition to the best possible signal processing at the digital level, the analogue part of the HiFi chain is also crucial, from the amplifier to the speakers. From a sober point of view, an optimised listening environment can also achieve substantially more than simply selecting the highest sampling rate. Listening tests should be carried out blind and with suitable music passages; above all, parameters such as volume are crucial in order to be able to make a serious comparison. Louder alone is not a real quality advantage…The most important questions and answers about Hi-res audio in a nutshell

Glossary around Hi-res Audio

• Sampling rate – number of measuring points per second; determines which frequency spectrum can be digitally represented.
• Bit depth (word width) – Number of bits per sample; influences the theoretical dynamic range and amplitude resolution.
• Nyquist frequency – Half of the sampling rate; maximum displayable frequency without aliasing.
• PCM (Pulse Code Modulation) – The common linear digital format that encodes samples with a fixed word width.
• DSD (Direct Stream Digital) – An alternative 1-bit model with a very high pulse rate, used in some Hi-res workflows.
• FLAC / ALAC – Lossless compression formats that allow bit-perfect restoration of the original data.
• MQA (Master Quality Authenticated) – Proprietary approach to providing ‘studio-authentic’ files with compression and authentication; technically controversial.
• Dither – Purposefully added noise before quantisation in order to reduce audible quantisation artefacts.
• Jitter – Temporal instability of the sampling clock; can distort phase information and cause distortion.
• THD / IMD – Metrics for the distortion of a system; important for assessing the linearity and purity of playback.
• Oversampling / Filters – Digital measures in DACs and ADCs that avoid aliasing and shape the signal; filter design influences phase behaviour and impulse response.
• Dynamic range – Difference between the quietest and loudest signal that can be reproduced; depends on bit depth and noise.
• Lossless vs. Lossy – Lossless means: complete recoverability of the original data; lossy sacrifices data for smaller files.
• Sample rate conversion (SRC) – Conversion between sample rates; if poorly implemented, SRC can introduce artefacts.
• Bit-perfect – Bit-accurate output guarantees that the same unaltered bits from the lossless music file (WAV, FLAC, ALAC, …) are retained during signal processing, ensuring accurate performance

Prices and availability

As already mentioned, content in Hi-res audio has long been standard with numerous streaming providers and download portals. At least with streaming, there is only a small surcharge – if any – compared to standard offers, especially as there are also other advantages available. However, the situation is different for downloads, which are sometimes significantly more expensive than lossy formats such as MP3 or AAC, and even content in “CD quality”, such as FLAC or ALAC with 16 bit and 44.1 kHz. In addition, you should pay very close attention to where you specifically purchase downloads in Hi-res audio, as many providers, and not least the content suppliers themselves, continue to commit a lot of fraud in this area. Unfortunately, simply upscaled content that offers no advantages over data with a lower resolution is not uncommon.

Conclusion

Hi-res audio is technically attractive and – with careful recording, mastering and playback – offers the potential to preserve additional fine details, transients and dynamic reserves. However, simply increasing the sampling rate or bit depth does not guarantee an audible advantage; the quality of the entire signal chain and the production methodology are decisive. For recording studios, archivists and critical listeners, Hi-res audio is a useful tool for maximising information density.
For the average listener in the home, the audible gains are contextual and often subtle: better converters, clean clocking solutions, correct mastering and a good playback environment usually prove to be more effective levers than blindly upgrading to the highest possible sampling rate.