Close Advises

Close Amplifiers

Close Apogee

Close Awesome system

Close CD Players

Close DAC

Close Dematerialization

Close Discs-Music

Close Fascinating components

Close Fidetech

Close Headphones

Close Historical

Close Interconnect cables

Close Loudspeakers

Close Optimization

Close Power cable

Close Preamplifiers

Close Receivers

Close Reference systems

Close Stereophile

Close Threshold

Close Tube amp list

Close Tubes and others

Close Tuners

Close Turntables


Close Accessories

Close Connectors

Close Interconnect cables

Close Power cable

Close Repairing your equipment

Close Speakers cables

Close Threshold amps optimization

To receive news about this website, consider subscribing to our Newsletter.
27 Subscribers
Webmaster - Infos
Dematerialization - The basics you must know


Over the last few years, the « dematerialization » of the music became more and more important, both in terms of market share and in terms of interest for the masses. At the very beginning, most of the people interested by the dematerialization were not HiFi enthusiasts, but mainly iPod addicts: they did not look at the best musical quality but at the ease of use.
But the things are changing, and the dematerialization is in the eye of a growing number of audiophiles, and we receive at Threshold Lovers, an increasing number of requests to comment and inform about this new approach, as well as to review various devices more or less dedicated to the dematerialized music.

But first of all, what do we mean by dematerialization ?

By dematerialization, we mean listening to the music that is not stored on a physical support like CD, DVD audio, SACD, Bluray audio or ... vinyl (see note 1) ! Most of the time, it will be stored on a file form, which can be stored on a computer disk, USB key or memory card (or on the Internet).

[note 1] : to not make a confusion between "digital" and "dematerialized"...


Our well known CD is more than 20 years old...(oh my god...). For those of you who were already able to compare the musical quality of the first CDs with a excellent vinyl turntable (see note 2), you probably remember that the first CD players were simply : a disaster. They were a disaster even for those of us who were not familiar with the high end systems...

It took two or three years for the CD to acquire an acceptable musical quality, or more exactly for the CD players to reach an acceptable level of musical quality for the masses.

When this first step has been reached, then the audiophiles were stll saying the same thing: « the CD is not as musical as the Vinyl » (and note that they are still a lot to say that, but this is not the purpose of this article).
Then we saw a new kind of (very important) devices coming on the market: the DACs. The Digital to Analog Converter, that emphasizes on the importance of the digital conversion to achieve a highly musical system. I am not 100% sure that WADIA was the first to offer a high quality DAC, but I remember that the WADIA 2000 DAC has marked our memories as being one of the most famous step forward in the digital music reproduction, accompanied with a few other well engineered products like the THETA for example. Most of these high end DACs used proprietary chips or algorithms, and were a « tour de force » at this period. The price of these DACs in France at this time were often higher than the price of big car !


Many years after, the DACs have made a lot of progresses, and they deliver a high musical quality at a low price. More exactly the chips involved in the DACs have (really) largely improved (see note 3), the manufacturers have perfectly understood the design traps to avoid, they know how to design a good DAC, the jitter is better and better managed, and a lot of people can afford a very good DAC at a very reasonable price (this was still more difficult 5 years ago...).

So we could believe that all is becoming perfect in the digital world. This could be the end of this article, but unfortunately Steve JOBS has made his iPod so popular, and the Majors have been so lazy (see note 4) in taking care about their customers, that the dematerialized music has become indredibly successful. Looking for a higher musical quality than a iPod (we can easily understand), the HiFi enthusiasts began to use another kind of DACs to convert the digital flow issued from their iPod or computer into analog music. These DACs were connected to the computer via a USB cable.

And then another problem appeared: the Audiophiles pointed out that these USB based DACs were not as musical as the SP/DIF based DACs when the same file was listened. For sure, we had got another very intesting subject to discuss about, for us, the Audiophiles...

So stop watching CNN, stop eating your burger, stop kissing your girl friend, we are just going to try to understand what happens !

[note 2]: ahhh, the Linn Sondek LP12 with a Lingo power supply, the Oracle, The famous Micro Seiki, the Well Tempered, etc...
[note 3] : for example the quality improvement between a CS-8412 chip input receiver and his pin-to-pin compatible brother CS-8414 is simply incredible....
[note 4]: pyracy is there but the materialized music supports, such as the DVD-Audio or SACD, could have been the successor of the CD if they had been correctly marketed and if it had been possible to copy them (yes copy by making the high resolution digital flow fully accessible) in order to lesson to them in our car or second house. But this is not the subject of our article...


Since the weaknesses  of the digital conversion have been discovered, one term has become popular: « jitter », and the promise of "the perfect sound issued from any CD player" has turned into a lot of confused explanations on the reason(s) for which we do not get this famous perfect sound... But the engineers have increased their understanding of the problems and what exactly happens between the file reading and the analog conversion. This is what we are going to study in the following sections.

In order to simplify, we will limit our study to the audio usage of the USB. The USB standard has been designed to provide an easy to use way of exchanging data from a computer. Although there are 4 possible modes available (see note 5) in which a USB connection can be used, we will limit our study to the one that interests our audiophile community: the Isochronous mode.

Contrarily to the Bulk transfer mode (see note 5), the Isochronous transfer mode is designed for the applications flowing the data on a continuous basis, and where time is critical. The bandwith is guaranted and the latency is limited. It is the mode for our audio applications. The data integrity is not the main purpose of this mode: the error are detected using the classical CRC method, but the data is not sent a second time in case of error: you certainly understand why you have been told about « errors detection » and why they should be avoided as much as possible (the data is not resent) ...

Contrarily to the Control transfer mode, the flow is unidirectional. The supported speeds of this mode are « Full speed » and « High speed » only (see note 6) : this is not a news, like we all know that the very first low speed USB was not suitable for audio...

Like we will see later, a major problem with the Isochronous mode is that if the data is corrupted for whatever reason (bus failure, interference, etc...) the data will not be sent a second time: there is no acknowledgement signal (or receipt) given by the receiving device. Consequently, if the data is corrupted, you can only trust the error detection system to * hope * the data to be recovered...

What is the read/send sequence in the Isochronous mode ?

The sequence can be described as follows:

  • the USB bus uses a dedicated clock (i.e. a quartz used only by the USB bus) @ 1kHz
  • the computer reads the data from the music file, which can be stored on a hard disk, a memory card or any other file storage device
  • when it has read the data, it stores the blocks of data in its memory and pushes (we say spools) them to the USB output in a continous flow at 1 kHz frequency (i.e every 1ms)
  • the rythm of this continuous flow is permanent and does not depend on the value of the data: the blocks are sent whereas they are empty or not

So you have understood that this guarantees that there is no timing problem.

Unfortunately it does not... ! mad  Even if the clock is dedicated to the USB bus, the actions are not only clock dependant: they also rely on the processor, which can be busy by many other processes than the USB actions ...

When it comes to the synchronization between the 2 devices (reader and converter), the Isochronous mode can be handled by 3 different methods:
  • Synchronous
  • Adaptive
  • Asynchronous
that we are going to compare herebelow.

Be aware that a very few companies do not use the USB audio in the classical Isochronous mode, but they adopt the bulk transfer mode and their own device: we will update this article soon on this specific approach (see note 5).


IV-1.The Synchronous method.
This is obviously the simplest one, and probably the first that appeared in USB audio. With this method, the clock that rules the DAC is made from the 1 kHz signal coming from the USB cable.
The main problem is not only that it is limited in the highest frequency that this method can handle (see note 7) but is mainly that it is very sensitive to the jitter. And the jitter is not our friend at all from a musical quality viewpoint...

We could say that the DAC is slaved to USB. And this is a very bad idea because the D/A conversion will be hugely impacted by the jitter and therefore we can just expect a poor quality...

IV-2.The Adaptive method.
  The jitter generated by the USB bus being our ennemy, what could be done to avoid it ? We can use an external clock instead of using the 1kHz USB clock.
  • In this approach, a separated clock, -meaning that the 1kHz coming from the USB cable is no longer  the master clock at the USB audio interface level-, is used. This separated clock is located in the USB audio controller interface.
  • It is of course a better approach than the Synchronous one, because it is far less sensistive to the jitter. Nevertheless, it can still be affected by a problem coming from the USB bus.
  • In addition, the use of a Frequency Synthesizer presents some disadvantages, because it can generate its own jitter...
  • A dedicated circuit looks at the data transfer rate on the USB bus and slightly modify the clock to be perfectly  in line with the transfer rate. The dedicated circuit is a classical Phase Locked Loop known as « PLL ».
Synoptic of the way it works:

The first chips were limited to 16 bits such as the Texas Instruments PCM2702 for example: datasheet is available here.
Nowdays, most chips offer the support of 24bits/96kHz. As always when talking about computers, drivers are needed and the USB Audio drivers class 1 cannot go beyond 96kHz: you need USB audio drivers class 2 to reach the 192kHz. See note 8.

The VIA EnvyVT1730 or the GFEC Tenor 8802L are two examples of more modern USB 2 Audio controllers.


USB Interface
» USB 2.0 High-Speed (480 Mbps) Device Compliant
» USB 2.0 Full-Speed (12 Mbps) Device Compliant
» USB 2.0 High-Speed PHY Integrated
» USB Audio Class v1.0/ v2.0 Device Compliant
» USB Device Firmware Upgrade v1.1 Compliant

High Resolution, Multi-Channel PCM Audio Bit Streams
» Up to 24-bit, 192kHz sample rate
» Up to 5 Input Streams and 6 Output Streams

I2S Audio Interface
» 8 I2S Input
» 8 I2S Output
   - Mono/ 2-Ch
   - 16/20/24-bit
   - 32 / 44.1 / 48 / 88.2 / 96 / 176.4 / 192 kHz

S/PDIF Interface
» 2 Stereo SPDIF Output,
» 1 Stereo SPDIF Input
   - Mono / 2-Ch
   - 16 / 20 / 24-bit
   - SPIDIF In: 32 / 44.1 / 48 / 88.2 / 96 kHz
   - SPIDIF Out: 44.1 / 48 / 88.2 / 96 / 192 kHz

Blu-ray Disc™ Audio Content Protection
» 24-bit/192kHz audio output delivering cinema-level audio experience

MIDI Interface
» 3 MIDI In
» 3 MIDI Out
» Any MIDI port can be used as UART

More information on the VIA VT1730 can be found on the manufacturer site.

or the GFEC Tenor8802L:

  •  USB2.0 High-Speed device supporting Audio Class v2.0
  •  USB2.0 Full-Speed device supporting Audio Class v1.0
  •  Audio Features
  •  2-Inputs support by one I2S pairs with 128/256 Fs
  •  2-Output support by one I2S pairs with 128/256 Fs
  •  Adaptive/Asynchronous Mode supported
  •  High-Speed mode support Adaptive/Asynchronous
  •  Full-Speed mode support Adaptive only
  •  Resolutions support 16/ 24-Bit with sampling rates support 44.1/48/88.2/96/176.4/192KHz
  •  Built in one IEC60958 professional 24 bit/96KHz S/PDIF RX
  •  Built in one IEC60958 professional 24 bit/192KHz S/PDIF TX
  •  One set of master I2C ports
  •  Two sets of standard MPU-401 compliant MIDI Tx/Rx pair by UART with pass-thru.
  •  Offer customized driver for Windows OS
  •  PLL integrated to support single 12MHz crystal operation
  •  1.8V core, 3.3V analog and I/O (3.3V tolerant)
  •  LQFP-128 package
  •  PB-Free

More detail and evaluation boards related to the GFEC Tenor8802L can be found on the manufacturer page.

But let us back to our subject: the synchronization methods.

IV-3.Asynchronous method, or « Host, DAC, Slave, Master, and...Success ».
This is the last one, and probably the one currently generating the biggest buzz on the internet.

This method differs from the adaptive mode in 2 points:
  • an external clock, meaning external to the USB audio controller, is used and can be chosen to have a very high precision. This clock is used to pilot the flow coming out of the buffer and going into the DAC
  • a returning flow (coming from the USB audio interface to the host) inform the host to adjust the sending rate of the data, in order to avoid a buffer overflow/underflow (the buffer is located in the USB audio controller).
We could say that the host is slaved to the DAC.

This mode is very interesting because:
  • a very high precision external clock can be chosen by the manufacturer to minimize the jitter (of course if the clock is of poor quality, do not expect a godd result...).
  • the high precision clock is as close as possible to the DAC (or even in the DAC chip itself)
  • the flow of the data sent from the buffer to the DAC is not affected by anything happening on the USB bus between the host and the USB controller
  • the flow of the data sent from the host to the USB controller is piloted by the status of the buffer (empty enough / close to overflow / close to underflow)
  • the DAC works in the best possible conditions
and seems to currently be the favorite one amongst the audiophiles. This approach allows to  use the best available clocks and a more flexible design.

Synoptic of the way it works:

This design is much more fexible and allow the manufacturers to use the best available clocks. Although we all know that in the HiFi world, the price does not guarantee at all the final quality, when it comes to the DAC, you have to be even more cautious !... Any manufacturer can write "Asynchronous" on a DAC, but the quality of the design is everything.

A very few manufacturers do have a lot of expertise in the clocks design and in the design of DACs. Amongst them, I would just like to emphasize on ANTELOPE, which adresses the profesionnal world (music sutdios, etc...). It is one of the very rare HiFi manufacturers that has a very strong position in the pro world, and design/sell really high precision Master clocks, such as their atomic clock ISOCHRONE 10M. This Master clock is of course designed for the music studios and its average price tag, 47,000€ (!), clearly keeps it out of the consummer target. Nevertheless, they have other models of Master clocks available here.
Their HiFi consumer DACs are designed to be connected to an external Master clock, such as their ZODIAC line of DACs: I have not (yet) reviewed them, and I do not know if a high end product such as the Zodiac Gold could be improved by using one of these external Master clocks, but the design seems very serious.

These are the main concepts of the USB audio used in the dematerialized world. Threshold Lovers will continue to update this section with further information and reviews.

Functional block diagram of the VIA VT1730:

Finally, the last few months/years have seen a new kind of « ready to use USB receiver modules », pushed on the market by a few companies that have made digital their core business, with a real and deep knowledge of the digital techniques. Amonst them, the XMOS company is well known in the audio world, because their XMOS chips/kits/boards are largely used by various HiFi manufacturers, some of them being the most famous. The use of a XMOS chip is close to become a sign of well sounding products in the DACs arena. We encourage you to learn about their USB Audio modules.

[note 5]: these 4 modes are:
  • the Bulk transfer mode, used for transferring large files or large amount of data. It is designed for applications where time is not critical but data integrity is; this includes for example storage devices or printers. Time being not critical, the bandwith is not dedicated and the transfers are performed using the « available bandwith ». Therefore, the delivery of the data is guaranted (it can be resent in case of error) but the latency is not.
  • the Interrupt transfer mode, used for transferring small amounts of data periodically, such as keyboard or mouse for example. Look to this link if you are looking for more information on this mode.
  • the Control transfer mode, which is designed to control devices, to establish communication with them and to access to their descriptors. A specific packet (setup packet) allows to get the vendor information and some informations specific to the connected device. Bi-directional flows.
  • the Isochronous transfer mode is the one described in this article.

[note 6]: Low speed, Full speed and High speed modes are defined as follow:
  • Low speed mode: 1.5 Mbit/s, the tolerance is +/- 15,000 ppm meaning +/-1.5%, its main purposes were keyboards, mouses or joysticks...
  • Full speed mode: 12Mbit/s, the tolerance is +/-2,500 ppm meaning +/-0.25%
  • High speed mode: 480Mbit/s, the tolerance is +/-500 ppm
  • Super speed is the USB 3.0, but as far as I know there is no audio devices currently using this mode. Up to 5Gbits/s full duplex.

as you can see, a huge improvement as been achieved.

[note 7]: The synchronous method is limited to 48kHz.

[note 8]: To date, a third party driver (i.e a non Microsoft driver) is required to run USB Audio class 2 on Windows, even Windows 7. This is not really a problem, but you just depends on the manufacturer to get driver updates in the long run. As far as I know, it is natively supported by Mac OS-X and Linux.

This article has been written by Nounours - January 2013.


Creation date : 18/01/2013 @ 21:39
Last update : 12/01/2016 @ 19:57
Category : Dematerialization
Page read 6463 times

Print the article Print the article

Reactions to this article

Nobody gave a comment yet.
Be the first to do so!

Members List Members : 693

Your nickname:


[ Password lost ? ]

[ Become a member ]

  Member online:
  Anonymous online: 15

Total visits Total visits: 1525578  




   visitors online

^ Top ^