New 28-bit DAC coming out.

[GXAlan]

This came across my Facebook feed. It's a company based in California that's claiming 28-bit true performance. The link just leads to a landing page with a few quotes from professional engineers. The fact they are announcing/showing it at AES instead of launching it at an audiophile show suggests that it's really backed by science. I hope they consider sending one to @amirm

Home - Imersiv

Delivering noise, linearity, and dynamic range performance so advanced that modern test equipment cannot measure it. imersiv HDR-A is an entirely new audio architecture, poised to re-imagine the way we capture and deliver sound, from microphone to power amplifier, and everything in-between.

imersiv.com

 

[Chrispy]

For production purposes I assume?

 

[polmuaddib]

For production purposes I assume?

Wouldn't it then be AD-DA? If it is just DAC, I am afraid it is just for audiophiles and SINAD chasers.
I could be wrong.

 

[boXem]

130 dB SNR is somewhere between 21 (126) and 22 (132) bits. 28 bits would be 168 dB.
Or somebody is playing with numbers, or my maths are really rusted.

 

[phofman]

The method is described in https://audioxpress.com/news/imersi...analog-converter-completes-32-bit-audio-chain . Two DACs in parallel, the first one converts the lower 19bits (i.e. real-world) resolution at full level, then attenuated by analog divider (i.e. low-noise) to merge with the high-level signal of the remaining upper 8 bits converted by the second DAC.

 

[GXAlan]

Doing some internet sleuthing. It’s John La Grou’s company.

John La Grou is founder and chair of POW-R, the world's leading audio bit-length reduction algorithms. Roughly one-third of all CD and downloaded music is processed with POW-R. He is also founder and CEO of Millennia Media, a design leader in critical audio recording, live sound, postproduction, mastering, and archiving. Millennia is the world's most popular front-end for film scoring and classical music recording, while Millennia's phono preamplifiers are used by the Library of Congress to archive their collection of three million historic audio recordings.

https://www.linkedin.com/in/lagrou

 

[GXAlan]

The method is described in https://audioxpress.com/news/imersi...analog-converter-completes-32-bit-audio-chain . Two DACs in parallel, the first one converts the lower 19bits (i.e. real-world) resolution at full level, then attenuated by analog divider (i.e. low-noise) to merge with the high-level signal of the remaining upper 8 bits converted by the second DAC.

Super cool! It went from vaporware to likely shipping this year!

This is what it is doing.

 

[phofman]

Not likely to DIY E.g.

US9871530B1 - Multi-path analog-to-digital and digital-to-analog conversion of PDM signals - Google Patents

An analog-to-digital and digital-to-analog conversion system using pulse-density-modulation (PDM) digital signals which minimize noise and optimize dynamic range by dividing a signal into multiple parallel pathways by apportioning a least significant range portion of an incoming signal to a...

patents.google.com

US9590648B2 - Multi-path digital-to-analog converter - Google Patents

A digital-to-analog converter which improves dynamic range by apportioning a least significant bits portion of an incoming digital signal to a low-path circuit and a most significant bits portion of the incoming digital signal to a high-path circuit. The low-path circuit has a low-path...

patents.google.com

US10256782B2 - Multi-path power amplifier - Google Patents

A multi-path power amplifier that includes two or more amplifier paths where the amplifier path having the most favorable dynamic resolution of the momentary signal is utilized. The amplifier paths are adaptively controlled by a processor, DSP, or other suitable digital processing. This results...

patents.google.com

 

[GXAlan]

Those patents are owned by Millenia, so that’s legit instead of being a no name company.

Their pro gear isn’t crazy priced

https://www.sweetwater.com/store/manufacturer/Millennia?_queryID=fe7d6242bbda6d7fb55ae36f8c7d7b03&_index=production_products

But they do have a $14.5K phono stage!

Millennia Products - Millennia Music & Media Systems

Millennia ~ Fine Audio Recording Products. For posterity; as great art. Arguably the finest music ever written; the Brandenburgs, Beethoven's major works, the 20th century masterworks (Le Sacre, R. Vaughn Williams..), the angelic vocalists (Ella..), the god-like players (Bird..) - this level...

mil-media.com

LOC’s front-end phono stage is based on the HV-3 double-balanced preamplifier used for recording nearly half of all Hollywood feature films scores (Titanic, Star Wars 1-3, Tarzan, Schindler’s List, Forrest Gump, Contact, AI, Armageddon, etc..). Leading acoustic music recording companies, such as Telarc, Sony Classical, Harmonia Mundi, DMP, Nimbus, Five-Four, Chesky, BMG, Koch, Klavier, and scores of others rely on HV-3 preamplifiers. The major symphony orchestras and operas of Los Angeles, San Francisco, Berlin, Cleveland, Seattle, Dallas, Israel, Prague, Baltimore, New York, and National (DC) employ HV-3 preamps, as do scores of others. Top live performers like Barbra Streisand, Celine Dion, Andrea Bocelli, Rolling Stones, and Bruce Springsteen use hundreds of HV-3 preamp channels in concert, as do the Academy Awards and other top global broadcasts.

Someone needs to get the company to send one to @amirm!

 

[tw99]

Can the AP actually measure to that level anyway?

 

[Scytales]

Super cool! It went from vaporware to likely shipping this year!

This is what it is doing.
View attachment 400435

I find the claim of 27 bits or 162 dB of dynamic range a bit misleading after having read Jan Didden's article in AudioXpress, where the whole system is described.

J.D. explains that the upper (7 bits) path and the lower path are only combined if the predicted output signal level reach a threshold above -26 dBu analog. From that threshold onward, the combined theoretical analog noise floor after the summing point is no better than the analog noise floor of the least good path, ie the analog noise floor of the "high path", which is -106 dBu (presummably over a 20 kHz audio bandwidth) from the text.

Hence, my belief is that if the signal barely reaches -26 dBu, the available signal-to-noise ratio at that level would be just 80 dB.

Even if the exact threshold and noise levels of a particular implementation would be different that the figures quoted by J.D., the logic of the system would remain the same.

So, what happens is signal dependent noise floor modulation. The rational behind that trade-off is, according to J.D., that noise is masked by sufficiently high signal, thus subjectively non-intrusive.

OK. Fine. That makes sense.

But I fail to see what purpose this system is intended to achieve beyond the possibility to claim to make use of all the 32 nowadays available bits in the digital domain.

What enjoyable musical sound needs 162 dB dynamic range? Here I like to quote Prof. Jamie Angus, who said about that very topic that we are no longer dealing with sound, but rather explosion, and more particularly the effect of thermobaric munitions.

 

[GXAlan]

What enjoyable musical sound needs 162 dB dynamic range? Here I like to quote Prof. Jamie Angus, who said about that very topic that we are no longer dealing with sound, but rather explosion, and more particularly the effect of thermobaric munitions.

https://www.sciencedirect.com/science/article/abs/pii/S0003682X21005284

Since it’s targeted at music production, I suspect it is for the convenience of recording. Taiko drums hit 120 dB at 5 meters, so if you are trying to mic the drum more closely, maybe you would want to have more dynamic range esp if it’s a mix of taiko drum and something else.

 

[solderdude]

Sooo... if the maximum SPL peaks my ears can shortly tolerate = 125dB SPL and a 0dB sound pressure can only be detected when having acclimated in an acoustically dead room why would one need a DAC that can do 162dB dynamic range ?
What amp can match that ?
How many recordings actually have a S/N ratio >120dB anyway ?

It's just a numbers game and has no benefit to audio.

 

[Sokel]

Sooo... if the maximum SPL peaks my ears can shortly tolerate = 125dB SPL and a 0dB sound pressure can only be detected when having acclimated in an acoustically dead room why would one need a DAC that can do 162dB dynamic range ?
What amp can match that ?
How many recordings actually have a S/N ration >120dB anyway ?

It's just a numbers game and has no benefit to audio.

My crystal ball shows me people listening to it with speakers that compress as early as 100dB and claim divine improved SQ,just like with the ones changing from a 115dB SINAD DAC to a 120dB one

 

[DVDdoug]

This doesn't excite me but I don't see how it's a bad thing (if/when it becomes economical).

Zoom makes some 32-bit floating-point ADCs and I WONDER if floating-point DACs are coming?

As far as I know, Zoom doesn't publish the digital headroom/clipping-point but (as expected) the nose-limited dynamic range is less than 24-bits. (They specify an EIN of -127dBu.)

 

[Scytales]

Those interested can read this short AES presentation by Mr La Grou :

E-library page - AES

AES E-Library ← Back to search

aes2.org

It is from this paper that Jan Didden took his numerical examples in his AudioXpress article. Those figures are thus La Grou's own when describing his systems. Mr La Grou effectively assumes a noise taken in the 20 Hz to 20 kHz bandwidth and a dynamic range from 77 nV RMS to 4.9 V RMS (or 6.9 V peak) at the analog output.

 

[kemmler3D]

This seems like a convenience thing (listen to very low level signals in the studio by turning up the gain on the DAC instead of in the DAW for some reason) and/or a "because you can" thing, sort of like having a commuter sedan that can do 0-60 in 2.5 seconds.

 

[AnalogSteph]

If I'm not totally wrong the developer has been on here. I'll just have to find the thread(s) again....
EDIT:

Theoretical minimum noise levels?

The author submitted several patent applications starting in 2016, wonder why it took so long to make it into an AES presentation? Patents don't seem to have been granted yet, though... and I would certainly assume some clever instrumentation guys came up with the idea already decades ago...

www.audiosciencereview.com

 

[Koeitje]

130 dB SNR is somewhere between 21 (126) and 22 (132) bits. 28 bits would be 168 dB.
Or somebody is playing with numbers, or my maths are really rusted.

Sounds impossible to be honest, must be close to some kind of theoretical limit for electronics at room temperature.

 

[KSTR]

Sounds impossible to be honest, must be close to some kind of theoretical limit for electronics at room temperature.

It's not impossible because a clever real-time range switching is done (actually, a cross-fade). Once the signal level is low enough, you switch to a heavily padded down output and re-adjust the digital input.

With ADC it's easier and it's been done for many years, if not decades. Any better guitar multi-effect pedal does this, these days.

The benefit of it is that you don't need to care about signal level ranges, Connect and forget. You can use one ADC and one DAC unit to cater for any signal from uV/mV range to several ten's of volts. For AC signal signals only, of course.