FAQ’s

General Questions:

How much do your products cost?

The INVICTA sells for $4,995.00 (USD)
The MIRUS sells for $4,995.00 (USD)
The CONCERO HD sells for $850 (USD)
The CONCERO HP sells for $850 (USD)
The HERUS sells for $350 (USD)
The HERUS+ sells for $425 (USD)
These amount are due on order either via Paypal or Wire Transfer. Once received, your Resonessence system will be constructed and tested in Canad

Why does the courier (UPS for example) demand that I pay taxes on my Resonessence product?

Summary: We don’t collect taxes for countries other than Canada, the shipping company will collect them upon delivery.

Details: The Resonessence company is registered in Canada, (in BC actually) and we do not have operations (ie offices or design centres) in any other country. Our WWW site and on-line store at Resonessence.com operate out of Canada (from the town of Kelowna in BC Canada). Our obligations to collect taxes are therefore limited to Canadian taxes. If you buy from our on-line store and you are a resident of Canada, that on-line store will calculate the appropriate taxes for your Province or Territory, but if you reside in any other country, we are not obligated to collect taxes for that country, and so we don’t do it. You will note that the site will say zero (0) for the tax payable if you reside anywhere other than Canada. (If we had an operation in, for example, California USA, we would be obligated to collect USA taxes for USA shipping addresses – we have no such office, so we don’t collect the taxes we don’t need to collect.)

But, just because we do not collect taxes for your government does not mean that you are free from tax. Various countries have different ways of ensuring that the tax gets collected when the company from which you purchased the product does not collect tax for them. For example, the way the USA does it is to obligate the shipper to collect the taxes and duties that are due. When you buy from our Canadian WWW site, we do not collect the tax due to the US government, but the shipper is obliged to do so. The shipper (UPS for example) will bring our product to your door and hand you a bill for the taxes, which are payable at that time, and he will only release the product to you when you pay that bill. It can be quite a surprise, and you may be left with a bad impression of Resonessence. Please be assured that we are in no way benefiting from this arrangement. All the same taxes apply when you buy from Amazon for example, but because Amazon has operations in just about every country in the world, they are obliged to collect tax at the point of sale: Amazon requires you to specify where you are resident, and then adds the appropriate taxes.

One final note: you may well ask “Why do you have us select our country when we go to check out from the Resonessence Canadian WWW site?” This is nothing to do with taxes, it is because we have made arrangements with dealers in certain countries to direct WWW traffic to their local site. If you are a resident of the UK for example, you will be directed to Unilet, and they WILL collect all the appropriate taxes (VAT in the case of the UK).

Where can I purchase your products?

Right here on the Web Site, or you may visit one of our Dealers and Representatives.

What is the MIRUS PRO?

Summary: The MIRUS has been upgraded to use the next generation ESS PRO series DACs – the ES9028PRO — and it sounds better. The technical specifications have not changed, but the audio experience has improved.

Details: ESS announced a next generation of audio DACS. They are an improvement upon even the remarkable SABRE DACS of the past. The new PRO series DACS have detailed improvements to the noise shaping modulator, less artifact correlation to the signal and all reviewers agree sound superior. Now you can experience the PRO series in MIRUS.

If you own a MIRUS we offer an upgrade to MIRUS PRO.

t.

What is the difference between the INVICTA and INVICTA MIRUS?

Summary: The MIRUS variant of INVICTA is designed for those customers who prefer to listen to music without headphones. It replaces the Headphone module with another ESS Sabre ES9018 module that is made to work in parallel with the first ES9018 module. Having two ES9018 improves performance even further. We can now guarantee a remarkable 130dB of Dynamic range and better than 114dB (0.0002%) Distortion + Noise. See the MIRUS Specifications.

Details: The internal construction of the INVICTA product line is modular, in fact we make use of physically separate modules to increase electrical isolation. The INVICTA consists of two primary analog modules: the DAC output module that drives the rear panel audio outputs, and the headphone driver module. The DAC output module has an ES9018 Sabre DAC and uses four channels in parallel for each stereo output. The INVICTA Mirus variant substitutes a second DAC output module for the headphone module. Hence it cannot provide a headphone output, but it can provide a total of eight ES9018 DAC channels for each stereo output. (It uses eight DAC channels to drive both the XLR (balanced) and the RCA (unbalanced) analog outputs.) You will read of other manufacturers who have also discovered that every time the Sabre DAC output channels are paralleled together performance improves. It is a unique feature of the Sabre DAC that its outputs can be added in this way (without the need for potentially noisy intermediate amplifiers). Doubling the number of channels reduces the noise by 3db. However, Resonessence does a little more. Firstly, we use a circuit level implementation that distributes the left and right channels equally between the Sabre chips, this ensures perfect channel to channel matching. And secondly, we actually also parallel the AD797 output amplifiers so that they also have their noise reduced by 3db at the same time. This second level of connectivity in parallel extends to the regulators as well: their noise goes down too.

The second generation INVICTA and INVICTA Mirus are optimized for 384kS/s. They differ from the first generation in that the USB underflow/overflow indicators are removed and we have added additional LEDs to show 352.8kS/S and 384kS/s.

A major upgrade to the INVICTA is also implemented in these second generation machines: the headphone module replaces the ES9016 with the higher performance ES9018.

The good news is that we can provide a software update for all existing generations of INVICTA that also provides these higher sample rates, so if you own an INVICTA you can get both DSD(64 and 128), DXD and higher samples rates in those machines. (Earlier generation INVICTA with the firmware update are only missing the high sample rate indicators).

INVICTA MIRUS ships with the Apple Remote Control and software revision 5.0.0 or later is available to allow all INVICTA and MIRUS product (shipped over the past two years or shipping now) to operate with the Apple remote as well as the original Resonessence remote.

text.

Why does Invicta and Mirus have unused connectors on the back panel?

INVICTA and MIRUS: Our USB firmware/software implements ASYNC 1.0 and 2.0 compliant USB Audio. The operating system of the host computer is detected and the DAC will switch between USB 2.0 if connected to the Mac, and USB 1.0 if connected to Windows. Note: After release 2.0.1 (approximately January 2012) we added the ability to manually set the USB Protocol for those customers who had purchased third-party USB Audio 2.0 drivers for Windows. Furthermore, we now provide at no cost, a customized Thesycon USB Audio 2.0 driver for Windows. See the FAQ below.

CONCERO and HERUS Support USB Audio ASYNC 2.0

What Warranty do you offer?

We warrantly the Invicta and Mirus hardware to be free from defects in workmanship and materials for three years. The details may be found here

We warrantly the Concero hardware to be free from defects in workmanship and materials for one year. The details may be found here

Please note: we may include other elements in the shipment to you, for example the USB cable. These we do not manufacture and we do not warranty, we include them in the shipment as a courtesy.

INVICTA/MIRUS Specific Questions:

Can you please say more about the use of DAC’s per channel in MIRUS?

Yes we can. We have tried to tread a fine line here between describing what we are doing and not giving away a technical aspect, but we see that we have sown some confusion in our customers minds so let’s just explain clearly what’s going on.

We use a total of two ESS 8 Channel DACs, we connect four outputs of one DAC to four outputs of the other DAC to make one stereo output. So each of our outputs comes from both DACs. We do that so that any mismatch between the two DAC chips cancels out.

Bottom line: there are 8 DACs contributing to each of the two outputs. That’s 16 DACs total, 8 to each output

.

Is there any reliability issue running Mirus at 0db all the time?

The MIRUS will have no issue running at 0dB output for the whole time, this is what it was designed to do. There will be no degradation of lifetime or reduction in audio quality.ext.

Are any relays used within the Mirus?

Yes, the relays are used to prevent power on/off pops and clicks. DONT worry about them affecting audio quality though since they are not inline in the signal path. They are only engaged during the first second of powerup and during powerdown. The rest of the time they are not in the audio circuitry. We obsessed about this detail of the design since we do not believe in having relays inline with the audio path.

(Note: if you are a designer of audio equipment you may be puzzled by our answer: how can the relays accomplish anything if they are “not in the signal path” when music is playing? The relays are not in the configuration you may expect, audio does not flow through the relays, the relays open to activate the DAC output and when closed force the output to be at zero volts. Hence as we state above, they are only on for the first second or so as the power stabilizes, and again for a second or so after receiving the shutdown command. Even if you yank out the power cord, there is sufficient time (and available charge) to activate the shutdown sequence.)

Is there any sound quality loss in playing FLAC from an SD card, rather than WAV?

No there is no sound quality loss. The FLAC is decoded into RAM where it becomes normal lossless PCM data, it then is sent to a FIFO for buffering between the RAM and the DAC. Then it is transmitted to the DAC via I2S interface. All this is done for the lowest jitter. The FLAC file is decoded in chucks in a streaming fashion. This is VERY difference from a regular CPU, where the bursty processing can have audible effects. The FLAC decoder is actually a hybrid hardware decoder and software. We did this so the CPU loading is actually very constant (smooth) during the stream decoding. This is only possible since we did not use a general purpose CPU, but rather made a custom decoder in the FPGA to accomplish this. I think the results speak for themselves. I have not heard of one person claiming they can tell the difference on the INVICTA (MIRUS), where its fairly well accepted that on a computer that is decoding on the fly, audible differences can be perceived.

What fuses should I use in the INVICTA and MIRUS?

For operation at 110V We recommend 250V rated fuses 500mA slow blow types. For operation at 230V we recommend 250V rated fuses 250mA slow blow types. The fuses are 5mm by 20mm.

How do I change or swap the fuses in the Invicta?

We have posted a video outlining the procedure for changing the fuses in the Invicta.

How do I enable or disable the HDMI output?

The HDMI output may be turned on (enabled) or off (disabled). You may choose to turn it off if you are concerned that the high frequency signals to the monitor may interfere with the exceptional audio output – although this is mitigated by good cables and the careful design of the INVICTA/MIRUS. You will need to turn off HDMI output at least temporarily when updating the firmware to version 6.2.1 or higher. To do so take the following step: Press the left of the four buttons and rotate to “Options”, press the rotary knob. Rotate to HDMI Output, press the rotary knob. Rotate to Disabled (or Enabled to turn it back on) and press the rotary knob.

CONCERO Specific Questions:

How do I setup CONCERO on an Apple computer?

We have prepared a full tutorial outlining how to connect a CONCERO via USB 2.0 on an Apple computer.

How does the Volume Control work on the CONCERO HP?

The operation of the volume control on the CONCERO HP changes the volume on the player application, but that application then sends the the volume control level back to the CONCERO and CONCERO implements a volume control in the 32 bit digital domain of the Sabre DAC. This is the best of both worlds: you can adjust the volume either on the CONCERO HP front panel knob, or on the music player software control, in both cases you are getting the high quality internal volume control of the Sabre DAC chip. You are not using an inferior software based volume control. To give you all the detail and also explain how volume control works in the DSD modes, here is what our senior engineer had to say on the matter:

USB DAC mode. Volume up and down commands are sent over a HID interface to the computer. Its the same thing as having a volume up or down button on a keyboard. Its up to the computer to decide how to handle these. This has both plus and minus.

Plus is the computer is in control of the volume of the DAC, and the volume can also be changed by using the volume sliders in the program. The minus is that some programs don’t handle what they do with volume control command very well. Those volume up and down commands are then sent from the computer back to the CONCERO HP where it applies a 32bit digital volume control in the Sabre chip!!. This is NOT a software volume control. It appears to work the same as a software volume control, since the computer tries to hide this level of detail from the user. Since the CONCERO HP supports hardware volume control, the computer sends the commands to the hardware for it to process. If there is no support for hardware volume control (which must be defined in the USB Audio descriptors) then the computer will automatically scale the data in software.

Now here is an issue I am aware of: what to do for DSD? Since the computer has no way of scaling the DSD amplitude in order to have a volume control, you MUST have a DAC with hardware volume. Since many DAC do have a volume knob, but are not sending commands to the computer saying volume up please or volume down please, some programs cannot support volume in DSD streaming mode.

From our tests, some DSD capable music players seem to ignore volume commands in DSD mode. However if you go to the windows volume control panel, you can still change it. Why? Because even if the application does not pass on the CONCERO HP volume commands in DSD mode, the computer is really the master of the volume, so MS Windows can change it. (That is, it can send the volume commands to the CONCERO HP – and you get the internal Sabre DAC volume control).

Again from our tests, other DSD capable music players seem to accept volume control commands in DSD mode, and if the DAC supports hardware volume control, it will pass it along. Thus in these players that do pass on the volume requests, the CONCERO HP volume control works just fine.

SPDIF DAC mode. Since there is no computer attached to the Concero_HP in the mode, the volume is handled internally….

HERUS Specific Questions:

Why is the volume very loud when I swap headphones?

Short answer: turn off the music when you swap headphones.

Detailed answer: We have had reports that HERUS can go to a full volume setting when the user swaps, or un-plugs and then re-plugs, the headphones. This behavior was unexpected and we have worked to discover why this is happening. We find that it is because music is playing while the headphones are being swapped or re-inserted. The act of removing or inserting the 1/4inch headphone jack causes a momentary short on the driver output; this is unavoidable and is a consequence of the design of the jack. During this time, if music is playing, the HERUS will attempt to drive the signal into this short-circuit. The output impedance of the HERUS is so low (0.2 ohms) that it dumps considerable current into this momentary short circuit, and that high current causes the USB power source to collapse. When the power source collapses, the HERUS electronics executes a hardware reset sequence, one consequence of which is that the ESS Sabre DAC comes up in maximum volume configuration. The USB chip resets to a default state and starts once again to send data (music) to the DAC, this music then comes out at full volume. Restarting a track or otherwise causing the software to communicate with HERUS will clear the problem, since the software will interact with the Sabre DAC and return it to the correct state. The solution to the problem is to ensure that there is silence in the headphones before removing or inserting them from the jack. [You may be inclined to ask why a hardware reset sets maximum volume in the Sabre DAC. The reason appears to be that the DAC must come up in maximum volume mode for those hardware configurations that do not use its internal volume control and have a very minimum software controller. The HERUS is a quite sophisticated software control system and makes no use of this “feature”.]

Can I use HERUS to drive an Amplifier?

Yes you can. HERUS is a DAC connected to the USB interface that drives a headphone power amplifier with a very low output impedance (about 0.2 Ohms) into as little as 32 Ohms. This means that it will be no problem driving the input of a power amplifier, which is commonly a much higher input impedance than the load presented by the headphones. Also, the output level of 2.5V RMS maximum from the HERUS is more than sufficient to meet the typical 2V RMS requirement.

The only issue is that the HERUS does not have a standard RCA or XLR output connector. It has the TRS 1/4in Stereo Headphone Socket. You will therefore need to build a Headphone to male RCA connector.

Does HERUS work with the IPhone?

Yes. We have tested the HERUS when connected to the IPhone 5 running iOS7.0 using the “Lightning to USB Camera Adapter” and it works perfectly. See this page for more details.

Audio Technical Questions:

What are the volume settings to get line-level (2VRMS) output?

You may want to use the Resonessence Products to drive into an amplifier and to do this you will need to set the output to be nominally 2v RMS. To do this set the volume controls as follows:

HERUS and HERUS+ -1.5dB or 84% of output makes 2V RMS (because on HERUS and HERUS+ 0db or 100% is 2.4v RMS)

CONCERO_HP -5dB or 56% of output makes 2V RMS (because on CONCERO_HP 0dB or 100% makes 3.5v RMS)

To set up MIRUS or INVICTA to generate line-level output refer to the document analog output level versus volume control for details (basically it is -2dB to generate 2VRMS output on RCA and -8dB to generate 2VRMS differential on the XLR)

How do I set Foobar for optimum audio quality?

Many of our customers use the excellent Foobar player, but we occasionally get reports of poor sound quality from the Foobar player with our products. These poor quality situations arise due to the different ways that the player can be configured on Windows: some configurations appear to work (in the sense that sound comes out of our products from the USB connection to the Window’s operating system), but these configurations are not optimum because of the way Windows and Foobar interact to deliver USB Audio data.

We will begin by describing the recommended way to configure Foobar and Windows, and then explain two of the configurations that are not optimum, but may appear to work, and the reasons why they are not optimum.

Please note that none of this discussion is specific to Resonessence: the matters discussed here apply to any DAC that you may care to use with Foobar on Windows.

Recommended Configuration: Configure the default Windows Sound output device to be something other than the Resonessence USB Audio, then enable Foobar’s advanced Kernel Streaming (or WASAPI) mode to connect directly to the USB Output. In the Foobar configuration window do this:
image[21]

And in the Windows Control Panel do this:
image[11]

Note what we are recommending: the Resonessence device is NOT the default output device for Windows. It IS the output device for the advanced Kernel Streaming mode of Foobar. This configuration prevents the Windows OS from interfering in any way with the pure data transfer from Foobar to the Resonessence device. Data is delivered perfectly from the Foobar player to the Resonessence DAC with no disruptive data compression, no attempt to change bit rate, and no loss of fidelity. You will hear the finest detail of your digital music. Note also that any other “incidental” sounds from the computer, such as beeps and so forth, and any sounds generated by, for example, your web browser, will not be routed to the Resonessence DAC – the DAC is the exclusive output of the Foobar player.

Commonly found, but not recommended configuration 2: When we assist our customers in their set up of the Resonessence products we sometimes find this configuration. Knowing that Kernel Streaming (or WASAPI) is the best way to configure Foobar, knowledgeable audiophiles activate the Kernel Streaming mode, but they also leave the default Windows device as the Resonessence device as well. That is, the Foobar configuration is as shown above, and the Windows configuration is as shown here:
image[18]x

Although music will play from the Foobar player and also from any other sound source on the computer, there will be times when the sound quality is really quite poor. The reason being that the Foobar player is configured to think that it has control of the sound output, but Windows also believes it has control at the same time. This will result in erroneous configurations of the USB Audio output channel, and although some sounds may emerge, they can be quite intolerably bad.

Foobar and Windows simultaneously able to access the Resonessence device: In order to allow both Foobar and Windows to have simultaneous access to the Resonessence device you must not activate Foobar Kernel Streaming mode. That is, to leave the Resonessence product as both the output for Foobar music and the output for any other incidental (ie web, beep, or other Window’s sound) output you must activate the DS (Direct Sound) output of Foobar.

Please note that this will work tolerably well, but you are NOT getting the full fidelity possible from the Foobar player! The reason being that Window’s will engage a sample rate conversion when it tries to “merge” the sound source from the other programs with the Foobar output. Basically, Window’s will reduce the USB configuration to the lowest common denominator of the sources of the sound, and since that is commonly 44.1k 16bit, all sources will be reduced with its (that is with Window’s) internal software sample and bit rate converter to this level. This will not meet the standards of the audiophile. (Within Resonessence we judge the quality to be no better than decent MP3).

So to summarize: we recommend with the excellent Foobar player that you activate Kernel Streaming mode and leave the default Windows Sound Output as something other than the Resonessence device. You will now get the benefit of excellent fidelity in high sample rate and DSD modes.

However, if you desire to have the Resonessence device as the output from Foobar and other Windows incidental sound sources at the same time, then configure Foobar like this:
image[9]

Then configure the Windows Control panel like this:
image[18]

How do I set up my media player to use DSD?

We have produced a guide that we will update form time to time showing how to set up popular music players to use DSD with our products. The examples in the guide use CONCERO HD as an example, but what is shown applies to any of our DSD capable DAC products. Download the guide here: RLabs_DSD_Setup

How do I wire up Differential Headphones?

diff-hp2

When using differential headphone mode on INVICTA the “A” output becomes the differential drive to Left Headphone, the “B” output becomes the differential drive to the Right Headphone. The TIP is the positive drive signal and the RING is the negative drive signal. The SLEEVE is the ground. Note: to active the differential drive mode go to “Invicta Options” and select “Differential HP” – see the user guide for more details.

The INVICTA Clock Architecture

The internal time control of an audio DAC is critical to the quality of the audio signal. Imperfections in the time of events in the system translate to perceptible artifacts in the sound stage. Here is what one of our senior engineers had to say in response to a customer question:

The clocking in the MIRUS is the same as the INVICTA, it uses a 50MHz CCHD-950.  The reason 50MHz was chosen is in Audio its important to have the lower phase noise in the Audio Band which is 20Hz to 20kHz. Notice how the CCHD-950 50MHz has –90dB at 10Hz!!! This is one of the key components in getting the great bass definition the INVICTA family has. The 44.1k and 48k clocks are derived from this same 50MHz clock using a PLL inside the FPGA. You would at first think the PLL is a bad way to do this, and normally it would be. The reason we do it this way is because we are referencing all the timing to the same 50MHz clock. The Sabre then uses the internal sample rate converter and BECAUSE all clocks are derived from 1 single source, the sample rate converter operates in a way when it mathematically removes ALL the jitter down to 0.001Hz or so. It is all about knowing how the clocking in the Sabre chip works that leads us to do this method, and it has paid off. The Audio resolution from the INVICTA family is amazing.

How does the 32-bit volume control work with a 24-bit signal?

The Invicta and Mirus volume control is done completely in the 32-bit domain, meaning that the input source is first extended to 32-bits and 32-bit precision is used all of the way through. What that means is if you set the volume control anywhere from 0 to -48dB, all the original dynamic range of the 24bit input is preserved. Once the volume is lowered below -48dB, you will start to throw away some of the LSB’s. Note that for 16 bit signals its -96dB where the original bits start to get discarded.

32bit – 24bits = 8 bits. 8 bits = 20*Log10(1/256) = -48.16479931dB
32bit – 16bits = 16 bits. 16 bits = 20*Log10(1/65536) = -96.32959861dB

xt.

Is the Invicta volume control digital or analog?

It is digital. We agree that ideally volume control should be analog because generally speaking, only an analog volume control can reduce the noise and the signal at the same time, so preserving the Signal to Noise (S/N) ratio. However, an analysis of the specifications shows that the digital volume control is at least as good as a well designed analog volume control because the signal to noise is so high – typically -132dB.

Specifically, an analog volume control would have to have less than -132dB performance relative to full scale before it would beat the Invicta/Mirus digital volume control. This is not easy to do, because a single 8k resistor has about this much noise, and by the time the output driver is added it is far from trivial to maintain a -132db noise floor.

Resonessence engineers concluded that they could, with great care, make an analog volume control at the -136dB level including the XLR driver amplifier etc, and this would just exceed (by about 4dB) the performance of the digital volume control. However, the difficulty of preventing x-talk, of implementing volume level matching, of maintaining THD, of preserving long term reliability, of lack of remote software control and so forth, all mitigate against the use of analog control and the decision was made to use the very high performance digital control.

To learn more about the Digital vs Analog volume control issue see a Digital and Analog Volume Control Example.

Can you clarify why BNC connectors are used on the rear panel?

BNC1 and BNC2 are two selectable input sources of digital music. Each is following the SMPTE 276M (also called the AES-3id) specification. That specification mandates BNC connectors. For more in formation see for example, http://www.rane.com/note149.html.

Much more information about S/PDIF can be found here: http://www.tech-faq.com/spdif.html.

The S/PDIF Specification is defined by IEC standard 60958-3, and is documented in the German patent EP000000811295B1.The SPDIF Audio Data Format is more recently part of a larger collection of the IEC-60958 standards (also known as the AES/EBU standard, and designated IEC-958 type II).

The firmware will correctly decode AES3 data sent into the BNC connectors, and it will correctly decode S/PDIF data sent into the BNC connector. (The non-professional user will typically connect a BNC-to-RCA converter on the BNC if he does not have a BNC Female connector on his S/PDIF source)

How is quantization noise calculated from number of bits?

The expression is: S/N(dB) = 6.02*N + 1.76 where N is the number of bits.

  • A 16 bit quantization can therefore achieve, in the very best case, S/N = 98.08dB
  • A 20 bit quantization can therefore achieve, in the very best case, S/N = 122.16dB
  • A 24 bit quantization can therefore achieve, in the very best case, S/N = 146.24dB

Engineers commonly neglect the 1.76 term and tend to approximate the 6.02 to 6 and simply use S/N = 6*N which can now be seen to be an approximation.

The first paper to show in detail how to calculate noise for a quantized signal is probably “Spectra of Quantized Signals” by W. R. Bennett in the Bell Systems Technical Journal 1948. In this paper Bennett develops his equation 1.4, namely D = 10*log(3r^2/2) where I have neglected the practical limitation term k as no longer applying in today’s technology. In this expression, r is the resolution of the quatization as defined by Bennett, and all log expressions are base 10.

Simple log relationships allow us to define an equation similar to Bennett’s 1.5, but omitting his need for the K term and substituting SN for D:

SN = 20 * log ( sqrt(3/2)*r )
from which
SN = 20 * log(r) + log(sqrt(3/2))
or
SN = 20 * log(r) + 1.76

Furthermore, if r is defined as 2^N where N is the binary bit count, then we have

SN = 20 * log (2^N) + 1.76
and using the logarithmic identity log-base-a (b) = log-base-c (b) / log-base-c (a) for any a, b and c we find
SN = 20 * log(2) * log-base-2(2^N) + 1.76
which is just
SN = 20*log(2)*N + 1.76
evaluating the constant term leaves
SN = 6.0206003*N + 1.76.

or to a reasonable degree

SN(dB) = 6.02*N + 1.76

A spreadsheet which runs a simulation and compares the theoretical S/N to the actual S/N of a real signal may be downloaded here.

Notes on Galvanic Isolation

Managing ground noise is key to high quality operation of any DAC. The first line of attack (used by all manufacturers) is to make multiple power supplies in an effort to decouple any that have noise from those that are sensitive. Each power supply lifts its potential relative to a common ground, and great care is taken in that common ground to make a uniform boundary condition of zero volts throughout the design.

Resonessence goes one step further than this: we galvanically isolate the power supplies. Galvanic isolation is more than just separate supplies: it is separate grounds as well, and it prevents any unwanted current flow causing a ground drop. Signals pass over specific isolation devices between these galvanic domains and noise in one domain simply cannot affect another domain because the return current is in the domain where the noise is generated. This is better than a well managed ground because there is no ground current to manage – not even the zero volt reference node is shared.

.

Why do you provide so many different digital filters?

Two of the filters are not Resonessence designs: they are the ones that come pre-programmed in the Sabre DAC. The other five, we have designed based on user feedback. Basically, these other filters are to allow our audiophile listeners to optimize the listening experience, and even customise the filters to the particular program, if desired. A unique feature of the INVICTA and MIRUS filters is that they can be selected while music is playing. You will be able to audit different filters and find the one that best matches your music program, your listening room and your taste. You may appreciate the precision of the very high performance filters, or you may prefer the subtlety of the slightly dispersive but “softer” filters. Much more information about the filters is available here.

What are the characteristics of the filters?

Graphs of the Frequency Response, Step and Impulse Response are shown here.

Ground Connections on the XLR Cable to Invicta

We were setting up a demo recently and were surprised to find that the Invicta, when connected to a power amplifier via a balanced XLR cable, sounded really bad. An investigation exposed the fact that the XLR cables had only one end grounded. This we find is not uncommon: an XLR cable may have only one of its ends connected to ground so as to avoid the problem of a “ground loop”. A ground loop is a situation where an unexpected current is flowing in a ground path and “hum” will be heard in the system due to a small voltage induced by this current.

It seems that certain XLR cables avoid the ground loop problem by having only one end at ground so preventing the current flow. This will be a successful strategy when the audio elements connected by that XLR cable have another common ground path elsewhere in the system.

However, recall that Invicta and Mirus have numerous internal galvanically isolated domains (that is, section of electronics that do not have any electrical connection between them), and one such isolated domain is the entire XLR output electronics. That means that the XLR output domain does not have a ground connected to the chassis or any other input port. In fact, the XLR output section is expecting to share the ground of the device to which it is connected: that is, it will adopt the ground potential of the power amplifier or similar to which it is connected. No power supply currents flow in this ground connection, there cannot be any ground hum. Furthermore, the precision output voltages are delivered relative to whatever the power amplifier uses as its reference ground, so minimizing common mode induced errors in even the best differential input.

This connectivity is technically the best, but as you can see, it presumes the XLR cable has ground connected at both ends. If you find “muddy” or otherwise far from perfect reproduction via a quality differential input power amplifier, take a moment to confirm that the ground is continuous across the XLR cable.

One further comment: you may be inclined to ask “How can the Invicta and Mirus provide galvanic isolation at the output when all terminals are on the same metal back panel?” The answer is that although all inputs and outputs are indeed on the same metal back panel, each is carefully electrically isolated from that panel. This means that if you are using any connectors with a metallic outer shell, those shell voltages may be different. [It depends on the detail of how your cable is constructed]. Consequently, avoid allowing any metal work of the external connectors to touch each other because that may defeat the isolation.

Why does Resonessence use the AD797 amplifier?

This topic has generated a few questions and comments that we did not anticipate. The pragmatic answer is that the design of the Invicta went through many revisions and the AD797 design worked the best in listening tests and in lab measurements. It certainly is not the lowest cost solution, but it performs the best as far we can determine. Really we should have left it at that and said no more, but our engineers have some thoughts as to why the AD797 works well and we shared them in communications with our reviewers – so for the record we will state them again here. As time permits we may write a great deal more, because this topic touches on many fascinating aspects of electronic design, and of audio design in particular – many too profound to go into one FAQ response, but here is a start.

Our discussion here will focus on the issue of feedback, and of FET input vs. bipolar input.

The idea of feedback as a catch-all solution to a less-than-optimum design is very tempting: with negative feedback a circuit that has only the virtue of high gain while neglecting linearity and so forth, may be made to look good. Negative feedback is a powerful technique that within a given bandwidth, and below a certain closed loop gain, will indeed suppress artifacts in the open-loop circuit design. With feedback we need not worry too much about the linearity of the circuit, negative feedback will fix it up for us.

But audio engineers and informed audiophiles know better: our ears are remarkable instruments that can expose what we may believe to be adequate performance in the closed loop configuration as being inferior to a better designed open-loop circuit with less feedback. This has led to a rule-of-thumb for some audio designers: “negative feedback does not seem to work its magic as advertized, always minimize negative feedback and use as close to open loop circuits as possible”. Not a bad conclusion, since following this rule will force the engineer to pay attention to the open loop linearity of a design, and that is always a good thing.

As soon as we challenge ourselves to build a linear open-loop circuit we begin to see things very differently: that bipolar device with its low voltage noise, high transconductance and reasonable current gain was the mainstay of many designs, but it is horribly non-linear in the open-loop configuration. Its transconductance as a function of collector current changes very rapidly, and its adherence a logarithmic characteristic makes its large signal response badly non-linear. Plus its finite (and not very low) base current is a source of noise over and above the voltage noise; if we are not careful the base current noise exceeds the voltage noise.

It is interesting to note that discrete designs evolved in the days before operational amplifiers when very high gain configurations exploiting folded cascodes, balanced current mirrors, base current cancellation techniques, and so forth, were not viable. Devices did not match, and in any case too many of them were needed. Hence high gain with lots of negative feedback did not commonly appear in discrete designs. Performance demanded inherently linear, closer to open loop designs, and for this reason it is not too far from the truth to say that a search for a linear open-loop design will lead one to appreciate the benefits of discrete designs.

One of the known means to improve open-loop linearity is to use a FET in the input stage. A FET is an inferior device in terms of transconductance per unit drain current, and it has a higher voltage noise for a given transconductance, but its approximately square law relationship increases its large signal linearity when compared to a bipolar device. Linearity in the presence of a high slew rate disturbance is also superior, since the input pair requires a much larger overdrive before the differential signal current saturates. And, what is more, a FET has a gate current near zero, so we can neglect gate current noise. If we are not to rely on negative feedback to linearize our design, the FET is clearly a better choice. Hence the second rule-of-thumb: “in audio designs a FET input amplifier works best”.

At this point in our thought process we seem about to conclude that a FET based discrete design is what we need. Why then does Resonessence use the AD797 op-amp which Analog Devices shows as a bipolar input design?

This drawing is from the AD797 data sheet – note the devices Q1 and Q2 – clearly shown as bipolar devices.

We use the AD797 because when we prototyped the Invicta, the AD797 measured and sounded the best. If that was not the case we would not have used it.

So all this commentary that follows is an “after the fact” attempt to understand why this sounds the best and to reassure ourselves that we have at least some understanding of the electronics involved to guide our design efforts in future products.

Negative feedback seems to work in this op-amp: why is that? Because, as described in the data sheet, this op-amp is using symmetry and bootstrapping for high gain, and a clever technique to enhance open-loop linearity (ADI calls this “cancellation of distortion due to the output stage”). It approaches the mathematical ideal of achieving all the gain in one stage because of its use of symmetry, bootstrapping and device matching (ADI says “This matching benefits not just dc precision, but, because it holds up dynamically, both distortion and settling time are also reduced”) . In summary we can see that this op-amp is a high linearity open loop design. It is what we would strive to do in a discrete design, but it has the benefit of matched, complementary, on chip, bipolar devices. And it is a clever design.

If you are going to use negative feedback at all, this may be the op-amp that can do it. But, no amount of matching and symmetrical cancellation can reduce base current noise from Q1 and Q2: it should still be inferior to a FET in that aspect – how can ADI claim ultra-low noise and how can Resonessence achieve -130dB of noise with this op-amp?

We need to look a little deeper to see how this is done, and a close reading of the data sheet gives us a clue. If the input were actually as shown in the drawing on the left, we may, if the bipolar device were exceptional, have a current gain (beta) of say 500. That is, I1 would be about 250uA at most. [The current nominally splits in two, 125uA per input device. Current gain (beta) equal to 500 would give about the 250nA of input current specified in the data sheet.] But this analysis is contradicted by the statement on page 12 “.. special input transistors running at nearly 1 mA of collector current.” Apparently, we must conclude that “nearly 1mA” of collector current gives rise to about 250nA of base current, or a current gain (beta) of 4,000. This is clearly no ordinary bipolar device. In fact it approximates a Junction FET (JFET) device, but still has a voltage noise of 0.9nV/RtHz and for compound impedances (noise source impedances) up to 1k ohms the current noise may be neglected. Analog Devices appear to have the ideal input device (at least for low impedance sources) that has the very low voltage noise of a bipolar device, and the very low input current (hence low input current noise) of a FET.

But, as I mentioned, this is all reasoning “after the fact” : the point is that tests confirm the performance of this amplifier. We may conclude that in the limit of an exceptional design, an integrated device has shown good performance. The open-loop gain is sufficiently linear and the input current low enough to perform very well in the Invicta application. The input devices behave like FETS in terms of the current noise, yet the amplifier benefits from the bipolar-like low voltage noise and, due to good design, does not suffer from open-loop non-linearity that one may expect of a bipolar input amplifier.

Can you please provide a table of absolute voltage output on the XLR and RCA connectors vs. volume level?

Yes, we’ve made a table available of analog output level versus volume control.

Why does Resonessence state that the Invicta is designed to drive a power amplifier and not a pre-amplifier?

When designing the Invicta and Mirus we had in mind a power amplifier (with XLR inputs) as the typical device that would be used to drive loudspeakers and so zero dB is designed to drive about 5V RMS. [Production units are set to drive 4.6v RMS with a 5% tolerance]. This can be much too high if connected to the pre-amplifier input of a consumer power amplifier. We do not want our customers to inadvertently damage their equipment so we suggest that -30dB is the volume setting when connecting to a pre-amplfier. -30dB creates less than 200mV RMS output and this is safe in all circumstances. [For example, AVI units such as the Denon AVR 951 and similar have 200mV input sensitivity].

However, -30dB will perhaps not create sufficient volume for more typical input sensitivities of 1.0v RMS.
We can calculate the safe volume setting for a pre-amplifier as follows:

Volume RMS output
0 4.6
-1 4.1
-2 3.654
-3 3.257
-4 2.902
-5 2.587
-6 2.305
-7 2.055
-8 1.831
-9 1.632
-10 1.455
-11 1.296
-12 1.155
-13 1.03
-14 917.8m
-15 818.0m
-16 729.1m
-17 649.8m
-18 579.1m
-19 516.1m
-20 460.0m
-21 410.0m
-22 365.4m
-23 325.7m
-24 290.2m
-25 258.7m
-26 230.5m
-27 205.5m
-28 183.1m
-29 163.2m
-30 145.5m

Consequently, if you know your pre-amplfier has about 1V RMS input level then -13dB is a good choice, if you know your pre-amplfier has a 2V RMS input then -7dB is a good choice.

Please note that the table above is for the XLR output ports. If you are using the RCA outputs, the levels are half of what is shown above.
For example, an RCA output connected to a pre-amplifier would deliver 1.0VRMS with a volume setting of -7dB, and would deliver just over 2.0VRMS with a volume setting of -1dB.

Note that many consumer products follow a Dolby recommended specification which calls for 2.0vRMS output, which would be -7dB if XLR is used and -1dB if RCA is used.

Technically speaking, the very best performance will result from the Invicta or Mirus XLR connected to your power amplifier XLR input, and your use of the volume control as the playback volume control, but we understand that many applications will want to provide a fixed level into a complete pre-amplifier – hence this note to help out.

[Actually, as many audiophiles are aware, and as documented in our a Digital and Analog Volume Control Example, the very best performance would result if the XLR output were set to zero dB, and a carefully designed, precision, analog volume control were inserted between the XLR output and the power amplifier XLR input. The very high dynamic range of the Invicta and Mirus would justify this if the digital source had SNR numbers in the mid-120’s. The reality today is that the performance difference for 16 bit sources is truly negligible, but as 24 bit music encoded at the studio with, for example, the new ESS Sabre ADC, becomes the norm, the audiophile may wish to consider this for the ultimate system. ]

If it turns out that many of our customers are using the Invicta to drive pre-amplifiers and perhaps only using the volume control when Headphones are connected, we will issue a software update that allows separation of XLR/RCA output levels from the volume control and sets them to a pre-set level. Provide feedback to us if you think this is a good idea. Note: Software release 2.1.0, August 2012, added this ability to turn off the XLR and RCA outputs in Headphone mode.

What amplitude and power does the Invicta provide to the headphone outputs?

Refer to the Specifications page for more details, but basically the headphone outputs drive up to 5.1v RMS and maintain an output impedance of 1.0 ohm until current limit activates. Take care to ensure the volume level is low before activating the headphone output: the maximum available power is very loud in the headphone.
(Note that in the Invicta Settings each output has an independent trim level so that the perceived volume can be the same if two headphones with different sensitivity are connected at the same time.)

The power and distortion+noise levels of the headphone drivers are as follows:

Headphone nominal impedance Peak Voltage Drive Maximum RMS Power Delivered THD+Noise (or better)
30 +/-3.9v* 260mW* 100dB
50 +/-5.8v** 280mW** 100dB
300 +/-7.1v 85mW 110dB
600 +/-7.2v 43mW 110dB

*With a 30 ohm load the headphone driver will start to limit the current (which introduces distortion) at a volume setting greater than -4.5dB. The specified output levels are therefore those for a -4.5dB volume setting. **With a 60 ohm load the headphone driver will start to limit the current (which introduces distortion) at a volume setting greater than -1.5dB. The specified output levels are therefore those for a -1.5dB volume setting. Current limit does not activate for any volume setting with 300 or 600 ohm load.

Do the Invicta and Concero upsample/oversample a 44.1kHz source up to 192kHz?

Invicta, Mirus and Concero may apply optional (and user selectable) Resonessence up-sampling filters, you may read more about them here: Details of the Resonessence Filters. This first filter step, which is completed in our DSP engine, then delivers 176.4kS/s or 192kS/s to the Sabre DAC.

This first filter is key to the quality and audio ambiance of the signal, since experienced audiophiles can easily detect differences in filter performance. Our customers tell us that the Apodizing Filter is often their preferred choice. Once into the 4x samples rates (ie into the 176.4 and 192 sample domain), further filtering has very little effect (in fact no effect that we can detect) on the audio experience.

Jitter at this 176/192 point in the signal path however, does have profound effect on quality; for which reason Resonessence utilizes the jitter rejection aspect of the Sabre DAC (activating it’s Asynchronous Sample Rate Convertor and Jitter Removal Circuits – read more here). This results in a further up-sampling to the precision master clock rate, which in all Resonessence products is an ultra-low jitter fixed-frequency 50Mhz clock.

Note: as of release 5.0.0 of the INVICTA and MIRUS software, you may provide data at up to 384kS/S from SD Card and USB sources.

USB Questions:

Can you say something about the Invicta USB Interface?

Our USB interface supports those drivers native to the Mac and Windows operating system. (As of September 2013 we have updated the USB to also operate with Linux). No special drivers need be installed. As a convenience for our users, the Invicta USB code determines the host type: if we detect a Mac operating system we switch to Asynchronous USB Audio 2.0; if we detect a Windows operating system we switch to Asynchronous USB Audio 1.0.

However, feeback from our customers in late 2011 and early 2012 suggest that the automatic setting of USB Audio 1.0 or USB Audio 2.0, based on our detection of the Operating System type, was not helpful.

Many customers had installed third party USB Audio 2 drivers on their Window’s Computers. Consequently in release 2.0.1 and later we provided a manual override of the automatic section, and went one step further by providing, at no additional charge, a customized Thesycon USB Audio 2.0 driver for Windows. This is delivered to our customers on the SD Card shipped with every purchased INVICTA and MIRUS system.

[ If you purchased your system prior to the software release 2.0.1 – typically meaning prior to January 2012 – you may contact us via Email to [email protected], or call us on Canadian/US Phone Number 1-778-477-5536 and we will provide, at no charge, a Thesycon USB Audio 2 driver for Windows. We cannot make this code available for general purpose download, since each is customized to the particular unit. If you Email or call, include the serial number of your INVICTA or MIRUS – it is on a sticker on the base of the unit, or available in the “Settings” menu. ]

The Invicta and Mirus can enumerate in two fundamentally different modes: as an Audio Device and/or as a Human Interface Device (HID) . The two run together to provide audio output and feedback to the computer for Pause/Rewind/FF etc.

The USB interface code is our own – (but as mentioned, the USB Audio 2.0 Driver for Windows we provide, is a third party-product) we have not purchased USB interface code or used any open-source code. Our in-house expertise in USB software is more than sufficient to write our own code and by doing so we can provide distinctive features for our users (such as the Mac/PC detection mentioned above). USB and the asynchronous audio aspect of it are not trivial to implement. USB (and for that matter MPEG etc) are industry standard specifications and there is a tendency to imagine that since they are in the public domain, surely any company can make use of them?

The reality is that the USB specification goes into great detail as to what is and what is not available in the USB protocol, but it does not tell you how to achieve any particular higher level functionality. It is like a map of a neighbourhood: it describes each house and all aspects of each house, but it does not say the order in which you should visit them, or the relative positions of one house to the next. It certainly does not tell you the shortest trip through the neighbourhood, nor the route to be taken by the mail delivery or anything like that. FedEx could have one route and UPS another. Each thinks his is the best route, and the specification will not tell him otherwise.

The specification is written this way to allow many different implementations and not accidentally favour any pre-existing solution from any one company. And, since companies are all driven by human beings, it is common for the engineer to validate that his design works on his particular set-up, only to find that there is a subtle incompatibility to another company’s implementation.

We have tested our implementation on the Mac and on the PC using various software to provide the audio output, and validated that our code works as expected and includes acceptable margins. For example, the USB software can silently, indefinitely and without any artifact at all, tolerate +/-1000ppm of timing error in the host relative the precision internal clock.

We would be foolish to suggest that our USB interface is perfect: software is complex and cannot be proven to be perfect. Recognizing this, we also provide a software update program which we will use to deliver updates to all our customers if the need arises. Don’t hesitate to inform us of any problems you find or suspect may be present. We learn from you and will support you.

Which USB audio protocols are supported by Invicta and Concero?

INVICTA and MIRUS: Our USB firmware/software implements ASYNC 1.0 and 2.0 compliant USB Audio. The operating system of the host computer is detected and the DAC will switch between USB 2.0 if connected to the Mac, and USB 1.0 if connected to Windows. Note: After release 2.0.1 (approximately January 2012) we added the ability to manually set the USB Protocol for those customers who had purchased third-party USB Audio 2.0 drivers for Windows. Furthermore, we now provide at no cost, a customized Thesycon USB Audio 2.0 driver for Windows. See the FAQ below.

CONCERO and HERUS Support USB Audio ASYNC 2.0

Does the Invicta support DSD over PCM (DoP)?

Yes, the Invicta, Mirus, CONCERO HD, CONCERO HP and HERUS all support version 1.0+ of the DoP open specification. We’ve provided a brief tutorial and overview on how to configure the Invicta for DSD over PCM.

When using a USB audio source does the Invicta control the computer, or vice versa?

When the Invicta or Mirus DAC is connected to a Mac or PC, we are able to control basic playback functionality such as play/pause, next track and previous track. Of course, you are also able to use the computer keyboard, remote control or mouse to control playback functionality as well.

How do I know what version of the Resonessence USB Driver I have installed?

On the software update page for INVICTA/MIRUS we ask you to uninstall any old Resonessence USB Audio 2.0 drivers.
You can find the latest version of the RUSB software here.

But how do you know if you have an old one installed in the first place? To find out go to the Control Panel Uninstall a Program option and look at the list of installed programs:

rusbver

The version number in this example is 1.67

SD Card Questions:

How large an SD Card can I use?

There is no limit applied by the INVICTA or MIRUS. You can use as large a card capacity as is available. At time of writing this note 128GB cards may be purchased and we have confirmed that these work in the system. We expect no issues up to the maximum advertized capacity of 2TB. Do bear in mind the formats of SD Card: we support the FAT32 file systems on SD, SDHC and SDXC cards.

How do I Format my SD Card?

Refer to the SD Card Association WWW page here: SD Formatter 4.0 for SD/SDHC/SDXC

Is the SD Card reader simply a feature or does it have some technical advantages over USB for example?

Updated Notes: Prior to release 3 of our firmware, we got reports from beta-testers that the SD Card source ‘sounded better’ than USB, and we sought a reason for this in our design. It is hard to see why that would be, but experience teaches us to listen to our beta-testers and trust the feedback they give us, even if our design sense tells us that there should be no difference at all.

We note that since Release 3 this difference between USB and SD Card is no longer reported. But truth to tell, we did not actually even know how to address the problem since we could only identify second order effects (as described below) which are all already removed to the maximum extent we know how to implement. USB buffering which was a possible candidate for the source of imperfection, is not in our control since we are the client, not the host, on the USB bus.

However, Release 3 of the firmware is reported to have made SD Card and USB indistinguishable, and you may be interested to know what we did in Release 3 that seems to have serendipitously achieved this.

Release 3.0.0 and later firstly removed the ‘polling’ of the user interface and replaced it with interrupt-driven user input. This means that after the DSP has set up the USB processing paths and so forth, it no longer need keep checking for user interaction: it only responds when an actual user-interface event occurs.

Perhaps having more bearing on the issue, is that the software team found a redundant timing check in the USB data path and removed it. This timing check was ensuring a precise lock to the internal low phase noise clock, but it turns out that the data was already in that clock domain, hence this step was not necessary and removed.

Finally, FLAC decoding capability was added to the system. To achieve the real-time decompression ‘on-the-fly’ at 192kS/s was a significant challenge. Thankfully, we were able to do this in all INVICTAs that we have ever shipped, due to an op-code level optimization of the processor. That is, not relying on the compiler to give optimum operation code sequences, but actually examining them and hand-optimizing where a clear improvement was possible. We even went so far as to add additional hardware units to the processor to improve its op-code utilization. (This is possible since the processors in Resonessence products are the MicroBlaze reconfigurable processors from Xilinx). All this has led to a very efficient processor architecture that needs fewer clock cycles to complete common audio tasks.

These changes are now reported to make USB and SDCard exactly the same high quality audio. But, as you can see, it is unclear exactly where the fault was (if indeed there was one at all) and why Release 3 by all accounts has improved it.

Original Notes: We all like digital music because information stored in digital form is, in principle, incorruptible and preserved. The first forays into digital music were compromised by the need to compress the data: the lowest sample rate was used and even then the data files were too large to store conveniently. Compression was invented (MP3) to make long play-lists viable on early hardware. Thankfully, as technology has moved on, lossy compression is no longer needed and emerging standards are all loss-less. All manner of digital sources can now deliver digital music to a player: USB is popular, WiFi is often mentioned, and of course SD cards can be used as well.

To the casual listener the digital data is just another source of music: bit-perfect digital data from any source is identical, whether it be delivered by USB or on an SD card. What matters is the quality and care with which that digital data was captured at the studio, because, we generally assume, once captured into the digital domain it is now inviolate and available for reproduction anywhere, hence the digital revolution that surrounds us.

To reproduce the quality captured by the digital encoding at the studio is the challenge for Resonessence and other high-end Audio DAC manufacturers. To produce the very best sound quality requires attention to a number of aspects that are well known and some that are less well-known. For example, we learn from the comments of experienced audiophiles that the listening experience can differ in detail depending on where the digital data is originating (USB, Memory stick, CD etc).

This is a surprise, why do different digital music sources sound different? Is there something about the audio engineering in the DAC that can explain this? What do we need to attend to in the DAC design to minimize unexpected issues such as this?

The first consideration is that the audio device must be the timing master – we cannot rely on the imprecise timing that a typical computer provides – that is far inferior to the audio requirement. Audiophiles are well aware of the problems of jitter and good clock management. WiFi is an even bigger challenge because data flow within a typical radio environment is exceptionally unpredictable and prone to drop-out.

At the lowest level, low jitter and precise timing means that the audio clock has to be the “master clock” to the degree that even a phase locked loop cannot achieve. [Some solutions attempt to lock the average rate of the audio clock with a phase locked loop – but this is inferior to the extreme high “Q” that a master crystal audio oscillator can achieve]. Consequently, for the highest quality audio reproduction a low phase noise oscillator defines the master clock at the audio reproduction site (that is, within the DAC system and in a managed noise environment). Well-known buffering and flow control devices then surround this audio subsystem and ensure that data is clocked in and out of the interface as the audio subsystem needs it.

In principle this solution (audio master clock, sophisticated flow control to the digital transport electronics) is all we need – we are done with the design. However, experience teaches otherwise. Audiophiles can perceive a difference between data source such as USB and SD card – how can this be since digital data is just digital data isn’t it? How can there be any distinction between data that has arrived over the USB link as opposed to data that has been extracted from the SD card since they are each a bit-perfect copy of the other?

There can be differences due to what engineers and scientists call second or sometimes “higher order” effects. We will describe one of them in order to explain what these can be, and we will simply state that the Invicta has been designed to mitigate this and other similar effects.

Consider the example of USB audio. All credible systems have solved the problem of flow control and placed the master clock at the DAC (asynchronous USB and the Audio USB standards etc) and commonly, to lessen the load on the USB host, a buffer memory of significant depth is present in the signal path to allow for unpredictable data delays. Two processes (at least, maybe more) are then running in the DAC: the first is the high precision clock and data flow to the DAC element itself to minimize phase noise, the second is a supervisory asynchronous process that is watching the buffer memory status and feeding back to the USB host to maintain long term synchronization. And, that surprisingly, is the source of a second order problem that evidence suggests may explain why one data source differs from another.

That asynchronous process of flow control is occurring with frequency characteristics right in the middle of the audio band: the flow control process kicks in and out in the millisecond timescale, and as a result of this, data is flowing in bursts that have audio band frequency products. To make the situation worse, the USB driver itself is a low impedance buffer sucking pulses of charge from the power supply over many frequencies, but with a strong audio band component. The buffer memory load and unload is again a current drain, more so if the memory is larger, and again it has audio frequency elements in it. All this means that the power supply is stressed: there will be a small (if the design is done right) modulation of the power proportional to the USB flow control. This modulation breaks into the audio stream primarily through clock phase modulation (which does not degrade DNR and does not degrade THD and so cannot be seen in the specifications, but there is no doubt at all that it is audible).

All high end manufacturers have a guideline: the more power supplies the better – precisely to mitigate effects such as these. A good DAC design will have a completely separate power supply for the USB interface to minimize this effect. However, there is next step after multiple supplies and that is galvanic isolation. Galvanic isolation truly separates the power supplies because the use of multiple power supplies alone does not isolate the ground connection. Galvanic isolation as used in the Invicta does truly isolate the grounds as well as the power. With galvanic isolation the audio system ground is not disturbed at all by the USB flow control artifact and the phase noise modulation with power supply variation is at the very minimum. [It is still not zero because there are other effects that even couple between galvanically isolated domains, but it is the very lowest achievable.]

Contrast the difficulty of achieving the absolute maximum performance in the presence of these second order effects with the USB and SD Card. In the USB interface the flow control is affected by the USB host: its performance will change the detailed operation of the flow control process in the audio DAC. The basic problem is that the USB is a serial data source and marshalling (or “serializing”) of the data into the USB “pipe” is a necessity. However in the SD Card the data is randomly accessed at a far higher rate that in the USB (in a few nano-seconds) and in any order that the controller cares to ask for it. In this case then the frequency domain characteristics of the data access process are under our control: we can ensure that access is “even” and does not exhibit a frequency characteristic in the audio band. But even more than this, the SD Card does not have a line capacitance to charge and discharge (the USB does: the wavefront propagating in the controlled environment of the cable presents a significant load to the driver). The charge disturbances are lower to begin with before we add our “evening out” procedure when using the SD Card source.

We should stress that these are indeed second order effects, far below the level of attention typically applied to a lower-cost consumer product.

Finally, there are other second order effects. You may rightly guess that the display has a similar artifact: a characteristic current draw that can be in the audio band if care is not applied and so forth. All that we can find are taken care of in the Invicta.

Which audio formats are supported by the SD card reader?

As of firmware release 5 we support FLAC, uncompressed PCM WAVE files (*.wav), AIFF Files (*.aif or *.aiff) and DSD Files in DFF and DSF format. Bit rates and sample rates up to 24-bit/384kHz are supported.

Does Invicta support reading SDXC memory cards?

We are pleased to report that as of firmware release 3.0.3 SDXC cards are now supported. Go to the download page and install the latest firmware. (All INVICTA machines shipped are compatible with this update).