La nostra filosofia sui chip DAC – MicroSound Technology audio hifi

Some audiophiles, observing our choice of the PCM1792A or PCM1794A converter chips, raise an eyebrow, pointing out the age of the chip and arguing that technology has advanced beyond it.

Let’s clarify this point, because, in our view, the confusionconfusion caused by effective marketing from major competitors is widespread.

The technology behind the PCM179x DACs

First, let’s look at the type of converter employed by Texas Instruments in their PCM179x family of chips.
It is a unique and proprietary DAC architecture that splits the conversion into two parts and treats bits differently depending on their weight.

Let’s read the description present in their datasheet:

The PCM1792A uses TI’s advanced segment DAC architecture to achieve excellent dynamic performance and improved tolerance to clock jitter.The PCM1792A provides balanced current outputs.
Digital input data via the digital filter is separated into six upper bits and 18 lower bits. . The six upper bits are converted to inverted complementary offset binary (ICOB) code. The lower 18 bits, associated with the MSB, are processed by a five-level third-order delta-sigma modulator operated at 64 fS by default.
The 1 level of the modulator is equivalent to the 1 LSB of the ICOB code converter. The data groups processed in the ICOB converter and third-order delta-sigma modulator are summed together to an up to 66-level digital code, and then processed by DWA algorithm to reduce the noise produced by element mismatch. The data of up to 66 levels from the DWA is converted to an analog output in the differential-current segment section.
This architecture has overcome the various drawbacks of conventional multibit processing and also achieves excellent dynamic performance.

This type of conversion is unique and adopted by TI in its top-end DACs.
Now let’s see the advantages explained more clearly.
The combination of an ICOB converter for the most significant bits and a five-level delta-sigma modulator for the lower bits is a distinctive architectural choice.
The idea behind this architecture is to take advantage of the strengths of both approaches:

The ICOB reduces distortion for the most significant bits.
The delta-sigma modulator improves resolution and reduces quantisation noise for the least significant bits.
The DWA algorithm helps to minimise errors due to internal component mismatch.

From an engineering viewpoint, the PCM1794A (and the PCM1792A) still have characteristics that make them competitive even today:

Signal-to-Noise Ratio (SNR) of –132 dB (mono) and Total Harmonic Distortion plus Noise (THD+N) of 0.0004% very low values, which few modern commercial DACs easily surpass.
Balanced current-out outputs,which allow a high-quality I/V implementation.
The hybrid multibit/delta-sigma architecture,designed to combine the linearity and low distortion of multibit with the noise-reduction advantages of delta-sigma.

Of course, sound quality does not depend solely on the DAC chip, but also on the overall implementation, filtering, power supply and analogue output stage.
It can therefore be said that this architecture still offers high-level performance compared to many alternatives.

Is the PCM1792 obsolete?

Texas-Instruments has replaced the PCM1792 with the slightly improved PCM1792A, which is happily still in production today.
In full production health there is the entire family: PCM1792A, PCM1794A, PCM1796A, DSD1792A, DSD1794A, etc.

But what about technological evolution?

One must distinguish between marketing and real technological improvement.
We must ask: how does a company with enormous semiconductor design resources and an annual turnover of US$15.6 billion (2024)¹ not decide to create chips similar to those of ESS or AKM (if not outright acquire those companies) which are the main players in the DAC market for hi-fi? By comparison small to Texas Instruments, for example, ESS reported revenues of only US$7.5 million in 2023. ².
For example: 32-bit DAC chips — do they make sense?
Do they sound better?

Why do we believe 32-bit DACs are unnecessary in practice?

Actual effective resolution is much lower than 32 bits.
- Even modern high-end DACs, such as the ESS ES9039PRO or the AK4499EX, provide an effective resolution of approximately 21–22 bits in real-world conditions, as determined by their SNR and THD+N performance.
- A 24-bit DAC theoretically has a maximum SNR of around 144 dB (6 dB per bit).
- In reality, the best commercial DACs reach around 125–130 dB SNR, corresponding to roughly 21–22 bits of actual resolution.
- A 32-bit DAC would require an SNR of around 192 dB… impossible to obtain with current technology.
Most recordings and audio files do not reach 32 bits.
- Most music is recorded at 16 bits (CD) or 24 bits (studio).
- Even hi-res formats (like DSD or PCM 192 kHz/24 bit) do not exploit full 24 bits because of electronic noise and recording environment — the last bits are masked.
- Reaching “32 bit” only serves to make us believe there is an improvement where in reality nothing changes.
Physical limitations of electronics.
- D/A converters are subject to thermal noise, component instability and clock jitter.
- Even in ideal conditions, a truly 32-bit DAC would need an amplification and audio system completely free of background noise up to –192 dB — something impossible even in the most advanced labs.
Marketing and inflated numbers.
- Producers know that “32 bit” sounds better than “24 bit”, so they use it as a commercial lever.
- Even if the DAC internally processes 32 bits, the analog part cannot reach that resolution.
- Many DACs use 32-bit internal oversampling for processing (filters, volume, dithering), but the analog output stage cannot achieve 32-bit resolution.
ENOB (Effective Number of Bits)
- ENOB represents the true resolution of a DAC once noise and distortion are taken into account.
- Anche 32 bit DAC con ottimi valori di SNR (120–130 dB) forniscono in pratica circa 20–22 bit effettivi.

Conclusion

A well-designed 24-bit DAC (with clean power supply, low jitter and a high-quality analog stage) is more than sufficient for any practical and even professional application. 32-bit DACs are more a marketing operation than a real technological advance.