[Jawed]

Thanks, very interesting.

The pulse array modulator is the digital circuit that converts the data output (6 bits) from the noise shaper into an array of PWM signals. The output from the pulse array modulator is then fed back into the noise shaper to provide the feedback for the noise shaper. The element outputs then go to the OP flip flops, which via resistors give the analogue output that is then fed to the output stage. A flip flop and resistor is hence one element (1e). We need to use a flip flop so that the jitter from the FPGA is eliminated.

So this seems to imply that pulses have 64 different widths and that the elements take it in turn to add a unit-pulse to the sequence of existing unit-pulses that all join together to form a modulated "super-pulse".

e.g. if there were 2 elements, A and B, in the pulse array and the desired width of the pulse were "12", then A would fire, then B, then A, then B ... until A and B had both fired 6 times.

As pulse A lasts for a finite amount of time, B starts its pulse just as A finishes its pulse. So this is how you get "constant switching energy": as A turns off, B turns on. Although it seems to me that each full modulated super-pulse (e.g. the 12-wide pulse) starts and ends with pulse switching events that don't have counter-acting opposite unit-pulses.

With 4 or 16 or 20 elements, the longest pulse that's 64-wide (or 63-wide?) will require most elements to fire multiple times.

What I'm failing to understand is why quality improves with the greater set of elements. And, consequently, I'm missing how that gives DAVE better headroom at max/min output.

Is the number of elements directly related to the number of noise shaper output bits? It seems not as I think you're saying that the noise shaper output is 6-bit (both at 104MHz?) in both DAVE and Mojo (and Hugo?).

Does an increase in elements improve linearity due to power supply sizing/consistency of pulse energy? Do the pulse elements take time to recover from their previous "on"? Is that effectively a power supply noise effect?

 

[Rob Watts]

Thanks, very interesting.
So this seems to imply that pulses have 64 different widths and that the elements take it in turn to add a unit-pulse to the sequence of existing unit-pulses that all join together to form a modulated "super-pulse".

e.g. if there were 2 elements, A and B, in the pulse array and the desired width of the pulse were "12", then A would fire, then B, then A, then B ... until A and B had both fired 6 times.

As pulse A lasts for a finite amount of time, B starts its pulse just as A finishes its pulse. So this is how you get "constant switching energy": as A turns off, B turns on. Although it seems to me that each full modulated super-pulse (e.g. the 12-wide pulse) starts and ends with pulse switching events that don't have counter-acting opposite unit-pulses.

With 4 or 16 or 20 elements, the longest pulse that's 64-wide (or 63-wide?) will require most elements to fire multiple times.

What I'm failing to understand is why quality improves with the greater set of elements. And, consequently, I'm missing how that gives DAVE better headroom at max/min output.

Is the number of elements directly related to the number of noise shaper output bits? It seems not as I think you're saying that the noise shaper output is 6-bit (both at 104MHz?) in both DAVE and Mojo (and Hugo?).

Does an increase in elements improve linearity due to power supply sizing/consistency of pulse energy? Do the pulse elements take time to recover from their previous "on"? Is that effectively a power supply noise effect?

Hi,
 
Yes the output is 6 bits, but that includes overload bits too as the noise shaper needs that. So for 20 elements, its a range of 0 to 20, with 21 to 31 clamped to 20, and negative values set to 0. So with that any value from 0 to 20 can be encoded with all elements switching with net cancellation of edges - so common master clock jitter is eliminated for constant values.
 
Now the usual range is 17 values (16 elements on, and all off), so I need an extra overhead to ensure individual elements do not stop switching, and provide the noise shaper with headroom, so hence the extra 4. This means no distortion from pulse array as it approaches clipping.
 
The 4e arrangement is different.
 
Hope that makes sense, Rob

 

[romaz]

Hi Rob,
After a good two-week loan experience I have now ordered a Dave for use initially as a headphone-only system - AK380/Dave/HE1000 most of the time but I also have HD800 and will shortly be getting a MacBook Pro. I am pleased by all the posts commenting on how unnecessary expensive interconnect cables will be with Dave but I am not sure if that includes mains cables and mains conditioners. Do you have any advice to offer on whether mains cables and conditioners can affect the sound quality of Dave? Thanks in advance

I can share with you my own experience.
 
If you've read my previous posts, you know that I have not found any significant difference from one source to another.  If you recall, as a worse case scenario, I used a dirty 12-core Mac Pro with its switching PSU (which is a much dirtier source than any MacBook Pro as laptops contains lighter hardware with less RF and can run off batteries) connected to mains without line conditioning and with its standard 18g mains cable and to DAVE with a generic $10 USB cable.  This was compared against a variety of purpose-built music servers ranging from an $8,000 Aurender N10, $3,500 Aurender X100L + $1,200 custom Kenneth Lau linear PSU, $3,000 custom built PC with Windows Server 2012 R2 core + Audiophile Optimizer + linear PSU, and a $300 Sonore Sonicorbiter SE (Roon endpoint) + linear PSU.  While blinded, a group of 3 listeners could not state a preference for one source against another even though the optimized sources had the further advantage of being connected to a Son of Q balanced power supply (isolation transformer + RF filter) and benefited from expensive mains cables (Audience AU24SE or Digital Dynamics Challenger AE15 with active RF shield).  
 
Well, I have compared one more source because I couldn't leave well enough alone.  I was convinced I could find some source that the DAVE couldn't work its magic on.  I currently have on hand an Oppo BDP-105D which is an excellent CD/SACD transport.  Unlike typical CD/SACD transports, this unit is capable of Tidal streaming although those of you who have an Oppo know that Tidal streaming with the Oppo is utterly horrible.  While CD/SACD playback is reference quality, Tidal streaming sounds hollow and distant, worse than highly compressed MP3 playback from an iPhone. It is probably the worst implementation of Tidal streaming I have heard and I'm not sure why because the SQ I get from Netflix streaming from the Oppo is very good.  Well, I connected this Oppo unit to DAVE via my optical cable and as expected, CD and SACD playback is excellent, as good as any of the other sources I previously mentioned.  To my surprise (but maybe by now, I shouldn't be surprised), Tidal playback from the Oppo through the DAVE is also utterly spectacular and as good as any of the sources listed above.  If there was maybe a small iota of doubt I had before, that doubt is now gone.  Just feed the DAVE a bit-perfect file and find yourself rewarded.  I suspect you could stream a bit-perfect file wirelessly to the DAVE through Bluetooth and it would sound just as good.  Perhaps a feature to consider for the future?
 
Now what about DAVE itself?  As I shared with you, mechanical isolation matters.  What about the quality of mains power to the DAVE?  Here is the standard 18g US mains cable that came with my DAVE.  I had not used it on the DAVE until now:
 

 
Here is the $3,500 Dynamic Design Challenger AE15 mains cable I have been using with the DAVE.  
 

 
This mains cable is specifically designed for low power digital front ends like a DAC and has an active battery-operated RF shield and I can verify this shield works.  As you turn the shield from on to off with a switch, the music becomes less illuminated and probably has a similar impact as Shunyata's or Isotek's digital mains cables with RF shields.  Could the cheap 18g mains cable that comes with the DAVE sound as good as this expensive specialty cable?  Furthermore, with the DAVE, does a dedicated line conditioner/isolation transformer make a difference?
 
Well, here is my testing methodology:
 
1.  DAVE with cheap 18g mains cable > direct to wall VS expensive mains cable > direct to wall.
2.  DAVE with cheap 18g mains cable > direct to wall VS expensive mains cable > Son of Q (isolation transformer + RF conditioner) > wall
3.  DAVE with cheap 18g mains cable > Son of Q > wall VS expensive mains cable > Son of Q > wall
 
My listening room / home office has a dedicated AC line with good earth ground although this line also powers LED and incandescent lights in the room, a large LED television, several computers, computer monitor, scanner and printer, speaker amp and powered subwoofer.  My HE-1000 was used and connected to DAVE via DHC cable.  Two-channel was also evaluated.  Three tracks were chosen for their wide dynamic range, complexity and because I know them well.  I used (1) a DXD recording by 2L of Mozart's Requiem, a complex passage that contains amazing depth and a large chamber choir, (2) an organ improvisation of Deilig er Jorden (Fairest Lord Jesus), another DXD recording by 2L that I chose for its complex and rich bass dimensionality and (3) Beethoven's Piano Sonata No 29, a 16/44 recording performed by Alessio Bax, a young up-and-coming pianist I saw live a few months ago.
 
This is what I found.  While unfiltered power to components pre-DAVE (ie music server, CD transport) don't seem to impact the DAVE, good power to the DAVE does make a noticeable difference and while this difference is not as pronounced as I have heard with my other DACs, it is noticeable even with blind testing.
 
In each scenario, when DAVE was connected to the expensive mains cable, the volume seemed elevated and this is how I could so easily pick it out even when blinded.  I have noticed this phenomenon in the past with my other DACs and it isn't because the SPL is actually higher but more likely because the noise floor is lower and the dynamic contrasts are greater.  The highs are more extended, easily noticeable at the 45 second mark on Mozart's Requiem.  There is also more definition to the bass with the bass layers more finely delineated on the organ track.  On the Alessio Bax track, when he is hammering on the upper register keys of the piano, with the expensive mains cable, there is clearer delineation of the keys whereas with the cheap mains cable, there is more noticeable smearing.  On 5/5 tries, when blinded, I was able to accurately pick out the more expensive mains cable. 
 
When the cheap mains cable was kept plugged into the wall and the expensive cable was plugged into the Son of Q conditioner, this delta widened further although the benefit was less pronounced compared to the expensive cable by itself.  When the cheap mains cable itself was plugged into the Son of Q, the gap became narrower.  Both with and without the Son of Q, there was no noticeable compromise in dynamics, a claim made by some against line conditioners.  It became clear to me that both the isolation transformer/RF conditioner and the more expensive mains cable resulted in additive improvements essentially leading to a darker, more dynamic and refined presentation.  The effect was more noticeable on headphones than it was on 2-channel and I surmise that this is because "direct to DAC" is more capable of revealing improvements.
 
These are my personal observations about the DAVE so far:
 
1.  The digital file matters.  There is a clear improvement with my 24/192 and DXD recordings over 16/44 (I'm not talking about upsampled music but native recordings).  While PCM is technically superior to DSD, DSD and especially DSD128 can sound very good.  Unfortunately, since Roon cannot currently play DSD256 via DoP, I am unable to play my native DSD256 recordings and I am left wondering how good this sounds through the DAVE.  Regardless, I am pleased the DAVE is so versatile.
 
2.  While the digital file matters, the digital file server does not (and I never thought I would ever be saying that).  Simple bit-perfect delivery even with a cheap USB or optical cable is all that is required.
 
3.  The DAVE benefits from mechanic isolation, especially with 2 channel.
 
4.  The DAVE benefits from clean mains power.  While Rob has done a splendid job of isolating the DAVE against its own switching PSU, a mains cable with good RF shielding can make a noticeable difference.  While some believe connecting a DAC to a dedicated line conditioner can potentially rob it of some of its dynamic potential, with either the Audience aR6-TSSOX passive conditioner that I have on hand as well as the Son of Q transformer-based conditioner that I have decided to replace the Audience unit with, I have not found this to be the case.  With these types of devices, YMMV and will depend on how much noise is in your mains line and so try before you buy.  In my system, the DAVE with the cheap mains cable by no means sounds harsh or lifeless but my better mains cable + isolation transformer / RF filter clearly adds a desirable sense of refinement.
 
6.  While the quality of digital interconnects don't seem to matter, the quality of analog interconnects, headphone cables and speaker cables absolutely matter as the DAVE has the potential to reveal the best qualities but also the limitations and flaws of such cables.
 
7.  With the DAVE, headphones potentially will sound better than 2 channel or any scenario where you have to add an outboard preamp or amplifier.  At the present time, because most DACs do not have the ability to connect directly to a headphone without an intermediary amplifier, it is my belief that the DAVE is the best headphone DAC in the world for most headphones. The immediacy and clarity that this direct connection provides cannot be overstated and while I cannot speak for the preference of others, personally, I would try to find another headphone before I considered adding a headphone amplifier.  As for 2-channel, for me it is a forgone conclusion that I will be adding Chord's new amp to my system once it becomes available if it provides the same level of transparency I am getting now with my HE-1000 direct to DAVE.

 

[romaz]

If any of you are wondering how I'm playing my SACDs from my Oppo through the DAVE, this is how:
 

I posted this previously on this thread as it worked well with my TotalDac but I just recently began to test it with my DAVE and I can tell you that it works flawlessly.  It is also adequately powered by my Oppo's HDMI port and so you won't need to plug it into the wall.  As you know, an SACD stream is encrypted via any digital connection except HDMI and so if you try to play a track from your SACD player to DAVE via optical or coax out, all you get is silence.  This $20 gadget I purchased from Ebay (made in China) will grab the SACD stream from HDMI and convert it to optical without altering the stream.  In other words, it remains bit-perfect and sounds very good, just as good as any other source you have.  DAVE will show a PCM sample rate of 88.2 kHz on its screen which is the appropriate DoP sample rate for DSD64.  If you own an SACD player with HDMI out (like an Oppo or select Cambridge or Arcam player) and you have a large SACD collection, this little gadget is highly recommended.

 

[romaz]

As the DAVE has the potential to sound outstanding in almost any situation, even without expensive mains cables or line conditioning, the focus quickly shifts to the quality of the digital file.  With a poor recording or poorly implemented digital file, the DAVE probably won't sound much better than any other DAC and so I am left to wonder if this is why certain people don't notice much difference between the DAVE and other DACs that it should easily outpace.  Moreover, there are vinyl aficionados who continue to believe that digital has yet to catch up to vinyl because digital still sounds, well...digital.  Much has been made of the flaws of DSD but as I hear it, 16/44, while better than DSD64, remains a chokehold on the DAVE.  As as been mentioned, certain DXD recordings sound especially good on the DAVE and I would have to concur with this.
 
While there is much about DAC technology that is beyond my current understanding, I believe I have a better grasp of the science of hearing and so I will lay out here what I understand and I would appreciate it if Rob or anyone else could correct my understanding and my logic if it is flawed.
 
As anyone who has spent quality time with the DAVE knows, this DAC excels in time resolution resulting in a very real sense of depth and the appreciation of even the most subtle spatial cues.  It is well established that the adult human ear has an auditory time resolution capability of somewhere between 5-10µs.  This means that if two sounds occur >10µs apart, most of us will hear 2 discrete sounds whereas if two sounds occur <5µs apart, most will hear only one sound.  It is this ability that tells us a tree is crashing quickly towards us and that we should run for safety.  It is this same ability that gives a sound a certain tonal quality and dimension.  It is also this ability that tells us we are listening to a CD which contains discrete bits of data interposed with blank gaps as opposed to analog media which contains a continuous data stream.
 
When we are presented with a digital file, all we are given are its bit rate (from 16-24 bits) and its sampling rate (from 44 to 384 kHz) and as we know, a CD or Redbook represents 16/44.  I have always considered sampling rate as the more important of the two parameters because this is what gives us the time resolution of the file.  With Redbook, at a sampling rate of 44.1 kHz, auditory time resolution is 20.8µs, meaning our ears can clearly discern that there are gaps and that there is time information missing.  As you increase the sampling rate to 96 kHz, the auditory time resolution improves to 10.4µs which is improved although compared to vinyl, most of us can still tell there are gaps.  It is only when you boost the sampling rate to 192 kHz that you bring down this auditory time resolution to 5.2µs.  At this sampling rate, most of our ears are unable to discern a difference against vinyl.  If we are to saturate our ears so that music sounds continuous rather than being made up of digital bits, it would seem we need to target a sampling rate of 192 kHz.  At the same time, anything beyond this would seem beyond the threshold of our ability to discern and seem unnecessary.  I understand the DAVE and other DACs that upsample will go beyond this sampling rate to address ringing issues introduced by the ADC but it would seem to me that this is a different issue altogether.
 
Bit rate, as we know, represents dynamic range with each bit representing 6 dB of DR.  This means a CD has 96 dB of DR while a 24-bit file has 144 dB of DR, an SPL that is well beyond what is desirable or safe for human hearing.  The DAVE lists a DR of 127 dBA or roughly 21 bits of DR and while Rob has told me he could have tweaked the DAVE to have a higher DR, it would possibly come at some compromise elsewhere and that it was useless to shoot for a higher value since it wouldn't translate to any improvement that could be appreciated.   Indeed, in the pit of a symphony orchestra, rarely do sound pressure levels exceed 100dB and so it would appear that 16 bits of DR or 96 dB is probably very adequate and that 21 bits of DR is more than is necessary.
 
If my logic is sound, then it would make sense for each of us to target recordings that are at a minimum of 16 bits of DR and 192 kHz of sampling.  Since these ratios don't exist, then we would want to target 24/192 recordings and indeed, this is what I try to do.  If a DXD recording is available, great, but if we have to pay more for DXD than 24/192, it would seem that it would be money wasted.  Rob, or anyone else, would you disagree?  With DAVINA, is there audible benefit to going beyond 24/192?

 

[Rob Watts]

One of the good things about the Davina project is that I will have clear answers to these problems.
 
Firstly, timing. The problem that Dave is solving, and its a very important problem only due to sampling the music, is the reconstruction of the timing of transients. Now a bandwidth limited signal (that is zero output at 22.05 kHz and above), if you use an infinite tap FIR filter, with a sinc function for the coefficients, would perfectly recover the missing waveform that was within the ADC before it was sampled. So if we have a DAC that has an interpolation filter that was "good enough" - that is double the taps and you hear no difference, and halve the time from one OP to the next and you still hear no change - then we will be left with a perfect reconstruction filter, and the DAC will re-create the signal effectively perfectly before it was sampled. What we will hear is the bandwidth limited signal. Now my question is - will bandwidth limiting within the ADC change the SQ? This I will find out from Davina, and I can test this without using decimation, so I will know this aspect for sure.
 
The second issue is amplitude accuracy. Now depth perception requires zero error in small signal accuracy - the smallest error in amplitude, no matter how small, seems to confuse the brain, and so it can't calculate the depth correctly, and we then see a degradation in the perceived depth. Now with Dave the small signal performance of the noise shaper allows a -301dB signal to be reproduced perfectly - that's way better than 50 bits, and actually more like 64 bit accuracy. So how do I encode 64 bit amplitude linearity within a 16 bit system at 44.1? Will triangular dither do it? In principle it will. Normally I use noise shapers to guarantee 64 bit audio performance, but although this works at 768 kHz, it won't work effectively at 44.1 kHz. Again, this is an aspect that I will find out from the Davina project.
 
Rob

 

[Jawed]

romaz, I'd like to rock your world slightly:

http://people.xiph.org/~xiphmont/demo/1kHz-105-dither.wav

This is a -105dB 1Khz sine wave encoded in 16-bit WAV format at 48KHz sampling rate. On my Hugo TT with HD 800 S I have to go to almost maximum volume to be able to hear the tone. But there it is

How is the -105dB tone still clearly audible above a -96dB noise floor?

The answer: Our -96dB noise floor figure is effectively wrong; we're using an inappropriate definition of dynamic range. (6*bits)dB gives us the RMS noise of the entire broadband signal, but each hair cell in the ear is sensitive to only a narrow fraction of the total bandwidth. As each hair cell hears only a fraction of the total noise floor energy, the noise floor at that hair cell will be much lower than the broadband figure of -96dB.

 

[Christer]

  As the DAVE has the potential to sound outstanding in almost any situation, even without expensive mains cables or line conditioning, the focus quickly shifts to the quality of the digital file.  With a poor recording or poorly implemented digital file, the DAVE probably won't sound much better than any other DAC and so I am left to wonder if this is why certain people don't notice much difference between the DAVE and other DACs that it should easily outpace.  Moreover, there are vinyl aficionados who continue to believe that digital has yet to catch up to vinyl because digital still sounds, well...digital.  Much has been made of the flaws of DSD but as I hear it, 16/44, while better than DSD64, remains a chokehold on the DAVE.  As as been mentioned, certain DXD recordings sound especially good on the DAVE and I would have to concur with this.
 
While there is much about DAC technology that is beyond my current understanding, I believe I have a better grasp of the science of hearing and so I will lay out here what I understand and I would appreciate it if Rob or anyone else could correct my understanding and my logic if it is flawed.
 
As anyone who has spent quality time with the DAVE knows, this DAC excels in time resolution resulting in a very real sense of depth and the appreciation of even the most subtle spatial cues.  It is well established that the adult human ear has an auditory time resolution capability of somewhere between 5-10µs.  This means that if two sounds occur >10µs apart, most of us will hear 2 discrete sounds whereas if two sounds occur <5µs apart, most will hear only one sound.  It is this ability that tells us a tree is crashing quickly towards us and that we should run for safety.  It is this same ability that gives a sound a certain tonal quality and dimension.  It is also this ability that tells us we are listening to a CD which contains discrete bits of data interposed with blank gaps as opposed to analog media which contains a continuous data stream.
 
When we are presented with a digital file, all we are given are its bit rate (from 16-24 bits) and its sampling rate (from 44 to 384 kHz) and as we know, a CD or Redbook represents 16/44.  I have always considered sampling rate as the more important of the two parameters because this is what gives us the time resolution of the file.  With Redbook, at a sampling rate of 44.1 kHz, auditory time resolution is 20.8µs, meaning our ears can clearly discern that there are gaps and that there is time information missing.  As you increase the sampling rate to 96 kHz, the auditory time resolution improves to 10.4µs which is improved although compared to vinyl, most of us can still tell there are gaps.  It is only when you boost the sampling rate to 192 kHz that you bring down this auditory time resolution to 5.2µs.  At this sampling rate, most of our ears are unable to discern a difference against vinyl.  If we are to saturate our ears so that music sounds continuous rather than being made up of digital bits, it would seem we need to target a sampling rate of 192 kHz.  At the same time, anything beyond this would seem beyond the threshold of our ability to discern and seem unnecessary.  I understand the DAVE and other DACs that upsample will go beyond this sampling rate to address ringing issues introduced by the ADC but it would seem to me that this is a different issue altogether.
 
Bit rate, as we know, represents dynamic range with each bit representing 6 dB of DR.  This means a CD has 96 dB of DR while a 24-bit file has 144 dB of DR, an SPL that is well beyond what is desirable or safe for human hearing.  The DAVE lists a DR of 127 dBA or roughly 21 bits of DR and while Rob has told me he could have tweaked the DAVE to have a higher DR, it would possibly come at some compromise elsewhere and that it was useless to shoot for a higher value since it wouldn't translate to any improvement that could be appreciated.   Indeed, in the pit of a symphony orchestra, rarely do sound pressure levels exceed 100dB and so it would appear that 16 bits of DR or 96 dB is probably very adequate and that 21 bits of DR is more than is necessary.
 
If my logic is sound, then it would make sense for each of us to target recordings that are at a minimum of 16 bits of DR and 192 kHz of sampling.  Since these ratios don't exist, then we would want to target 24/192 recordings and indeed, this is what I try to do.  If a DXD recording is available, great, but if we have to pay more for DXD than 24/192, it would seem that it would be money wasted.  Rob, or anyone else, would you disagree?  With DAVINA, is there audible benefit to going beyond 24/192?

Hello again romaz

  As the DAVE has the potential to sound outstanding in almost any situation, even without expensive mains cables or line conditioning, the focus quickly shifts to the quality of the digital file.  With a poor recording or poorly implemented digital file, the DAVE probably won't sound much better than any other DAC and so I am left to wonder if this is why certain people don't notice much difference between the DAVE and other DACs that it should easily outpace.  Moreover, there are vinyl aficionados who continue to believe that digital has yet to catch up to vinyl because digital still sounds, well...digital.  Much has been made of the flaws of DSD but as I hear it, 16/44, while better than DSD64, remains a chokehold on the DAVE.  As as been mentioned, certain DXD recordings sound especially good on the DAVE and I would have to concur with this.
 
While there is much about DAC technology that is beyond my current understanding, I believe I have a better grasp of the science of hearing and so I will lay out here what I understand and I would appreciate it if Rob or anyone else could correct my understanding and my logic if it is flawed.
 
As anyone who has spent quality time with the DAVE knows, this DAC excels in time resolution resulting in a very real sense of depth and the appreciation of even the most subtle spatial cues.  It is well established that the adult human ear has an auditory time resolution capability of somewhere between 5-10µs.  This means that if two sounds occur >10µs apart, most of us will hear 2 discrete sounds whereas if two sounds occur <5µs apart, most will hear only one sound.  It is this ability that tells us a tree is crashing quickly towards us and that we should run for safety.  It is this same ability that gives a sound a certain tonal quality and dimension.  It is also this ability that tells us we are listening to a CD which contains discrete bits of data interposed with blank gaps as opposed to analog media which contains a continuous data stream.
 
When we are presented with a digital file, all we are given are its bit rate (from 16-24 bits) and its sampling rate (from 44 to 384 kHz) and as we know, a CD or Redbook represents 16/44.  I have always considered sampling rate as the more important of the two parameters because this is what gives us the time resolution of the file.  With Redbook, at a sampling rate of 44.1 kHz, auditory time resolution is 20.8µs, meaning our ears can clearly discern that there are gaps and that there is time information missing.  As you increase the sampling rate to 96 kHz, the auditory time resolution improves to 10.4µs which is improved although compared to vinyl, most of us can still tell there are gaps.  It is only when you boost the sampling rate to 192 kHz that you bring down this auditory time resolution to 5.2µs.  At this sampling rate, most of our ears are unable to discern a difference against vinyl.  If we are to saturate our ears so that music sounds continuous rather than being made up of digital bits, it would seem we need to target a sampling rate of 192 kHz.  At the same time, anything beyond this would seem beyond the threshold of our ability to discern and seem unnecessary.  I understand the DAVE and other DACs that upsample will go beyond this sampling rate to address ringing issues introduced by the ADC but it would seem to me that this is a different issue altogether.
 
Bit rate, as we know, represents dynamic range with each bit representing 6 dB of DR.  This means a CD has 96 dB of DR while a 24-bit file has 144 dB of DR, an SPL that is well beyond what is desirable or safe for human hearing.  The DAVE lists a DR of 127 dBA or roughly 21 bits of DR and while Rob has told me he could have tweaked the DAVE to have a higher DR, it would possibly come at some compromise elsewhere and that it was useless to shoot for a higher value since it wouldn't translate to any improvement that could be appreciated.   Indeed, in the pit of a symphony orchestra, rarely do sound pressure levels exceed 100dB and so it would appear that 16 bits of DR or 96 dB is probably very adequate and that 21 bits of DR is more than is necessary.
 
If my logic is sound, then it would make sense for each of us to target recordings that are at a minimum of 16 bits of DR and 192 kHz of sampling.  Since these ratios don't exist, then we would want to target 24/192 recordings and indeed, this is what I try to do.  If a DXD recording is available, great, but if we have to pay more for DXD than 24/192, it would seem that it would be money wasted.  Rob, or anyone else, would you disagree?  With DAVINA, is there audible benefit to going beyond 24/192?

Hello again romaz, your last two posts are  very interesting indeed.
Since I haven't heard 16/44.1 via DAVE I will wait with my opinion regarding DSD 64 versus 16&/44.1. and DAVE. But via every thing else I have heard including Hugo 16/44.1 sounds better than ever via Hugo but still not as  good as  well recorded DSD64 IMHO. The timing and flow of DSD wins clearly for me over even interpolated 44.1 as with Hugo.
Ok at times resolution is a bit lacking in fff with lots of instruments playing con tutta forza.
But I basically never experience the digititis  listening fatigue 16/44.1 tends to induce even via Hugo after a while, with DSD 64.
And among  all classical recording  pros and conductors  I know, and they are quite a few, I can't recall that anyone thinks 16/44.1 sounds better or even as good as DSD 64.
What I can't  understand is why  there is so much attention to adress the problems and missing timing information with 16/44.1 when basically everything since many years is recorded at higher rates anyway?
Personally I  care little or not at all for genres that are stuck in the  rbcd timeloop.
And I still have not got an answer why  only the bandwith limited 16/44.1 can be recovereed in an ideal 1 million taps world?
Real live acoustic instruments like  percussion have  content up to 90-100khz don't they ?
I suspect that one of the reasons apart from timing and flow  that simply more information as such,with  both DSD and HD PCM makes it  sound better more realistic is because it contains more HF information which even if we cant define it as tones with our hearing  could still influence  things in a favourable way.
I did a similar test as the one you suggested and could hear a pressure on my ears up to 22khz where the test stopped and could define a tone on my right ear up to 19khz  in my right ear and 16  khz on the left. via HD800 and Hugo.
I think there are "a lot of unknown unknowns" to quote Rob .
Regarding the rest my ears tell me basically the same as yours.

 

[bmichels]

Those few last posts from Romaz and Rob are so interesting that I printed this page to archive it...

   Thanks to you all.

 

[JaZZ]

  Hello again romaz, your last two posts are  very interesting indeed.
Since I haven't heard 16/44.1 via DAVE I will wait with my opinion regarding DSD 64 versus 16&/44.1. and DAVE. But via every thing else I have heard including Hugo 16/44.1 sounds better than ever via Hugo but still not as  good as  well recorded DSD64 IMHO. The timing and flow of DSD wins clearly for me over even interpolated 44.1 as with Hugo.
Ok at times resolution is a bit lacking in fff with lots of instruments playing con tutta forza.
But I basically never experience the digititis  listening fatigue 16/44.1 tends to induce even via Hugo after a while.
And among  all classical recording  pros and conductors  I know, and they are quite a few, nobody thinks 16/44.1 sounds better or even as good as DSD 64.
What I can't  understand is why  there is so much attention to adress the problems and missing timing information with 16/44.1 when basically everything since many years is recorded at higher rates anyway?
Personally I  care little or not at all for genres that are stuck in the  rbcd timeloop.
And I still have not got an answer why  only the bandwith limited 16/44.1 can be recovereed in an ideal 1 million taps world?
Real live acoustic instruments like  percussion have  content up to 90-100khz don't they ?
I suspect that one of the reasons apart from timing and flow  that simply more information as such,with  both DSD and HD PCM makes it  sound better more realistic is because it contains more HF information which even if we cant define it as tones with our hearing  could still influence  things in a favourable way.
I did a similar test as the one you suggested and could hear a pressure on my ears up to 22khz where the test stopped and could define a tone on my right ear up to 19khz  in my right ear and 16  khz on the left. via HD800 and Hugo.
I think there are "a lot of unknown unknowns" to quote Rob .
Regarding the rest my ears tell me basically the same as yours.

 
My theory as to why higher sampling rate matters is interference (beside the filter resonance itself). Ultrasonics may not be audible themselves, but may leave audible traces nonetheless.
 
Here's an image with two tones of equal loudness, 15.2 and 23.9 kHz, sampled with 192 kHz:
 

 
The two tones mixed together create a signal with massive amplitude fluctuation. Let's consider it an analogue signal to be sampled with 44.1 kHz. Which implies the mandatory low-pass filter for preventing aliasing (actually before and after digitization). The filter with its high Q factor and correspondingly massive filter resonance will completely eliminate the amplitude modulation on the (audible) 15.2 kHz tone caused by the (inaudible) 23.9 kHz tone.
 
It's up to debate if the amplitude modulation be audible or not. Personally I see no reason why it shouldn't be. It was my strong impression since the first few beats I heard from my first CD that overtones were smeared by a coolish glare, everything was too clean and sleek.
 
Now with DAVE this is reduced to an easily tolerable rest – to my ears. As mentioned, equalizing my headphones is of higher importance to me, and since my FiiO X5 II allows equalizing only up to 48 kHz... Moreover, most of my music collection is 44.1 kHz anyway.

 

[Christer]

   
My theory as to why higher sampling rate matters is interference (beside the filter resonance itself). Ultrasonics may not be audible themselves, but may leave audible traces nonetheless.
 
Here's an image with two tones of equal loudness, 15.2 and 23.9 kHz, sampled with 192 kHz:
 

 
The two tones mixed together create a signal with massive amplitude fluctuation. Let's consider it an analogue signal to be sampled with 44.1 kHz. Which implies the mandatory low-pass filter for preventing aliasing (actually before and after digitization). The filter with its high Q factor and correspondingly massive filter resonance will completely eliminate the amplitude modulation on the (audible) 15.2 kHz tone caused by the (inaudible) 23.9 kHz tone.
 
It's up to debate if the amplitude modulation be audible or not. Personally I see no reason why it shouldn't be. It was my strong impression since the first few beats I heard from my first CD that overtones were smeared by a coolish glare, everything was too clean and sleek.
 
Now with DAVE this is reduced to an easily tolerable rest – to my ears. As mentioned, equalizing my headphones is of higher importance to me, and since my FiiO X5 II allows equalizing only up to 48 kHz... Moreover, most of my music collection is 44.1 kHz anyway.

Sounds plausible to me.
My first CD audition of classical music still remains the most disappointing HI FI or rather lack of the same I have experienced.
Even a direct cut 78rpm from the 30s  sounded more like real music to me, than what I heard from Philips and Sony's first cd players at the first demos.
Luckily it sounds a lot better these days.
And the difference between 16/44.1 and higher res is not day and night to me either.
But one of the things that has always troubled me with  rbcd, 16/44.1,and still does, has been the smearing  you mention which to me  has been especially bad with strings and percussion instruments.
Prof Johnsson's HDCD  was the first real  step towards better rbcd sound.
I wonder how close his carrier signal system is to what Rob Watts is doing  know?
Or did it add more bits? I can't remember. Anyway it sounded a bit better than plain rbcd.
Another  personal problem with rbcd which may sound silly to many here, is that I have found it difficult to conduct along  to rbcds compared to the same work on LPs .
With DSD and SACD  and now hi res pcm downloads I can conduct along to the music more easily again.Ok I am a lousy conductor and miss a lot of beats. But anyway.
I can only describe it as the music is flowing  again.

 

[STR-1]

I can share with you my own experience.

If you've read my previous posts, you know that I have not found any significant difference from one source to another.  If you recall, as a worse case scenario, I used a dirty 12-core Mac Pro with its switching PSU (which is a much dirtier source than any MacBook Pro as laptops contains lighter hardware with less RF and can run off batteries) connected to mains without line conditioning and with its standard 18g mains cable and to DAVE with a generic $10 USB cable.  This was compared against a variety of purpose-built music servers ranging from an $8,000 Aurender N10, $3,500 Aurender X100L + $1,200 custom Kenneth Lau linear PSU, $3,000 custom built PC with Windows Server 2012 R2 core + Audiophile Optimizer + linear PSU, and a $300 Sonore Sonicorbiter SE (Roon endpoint) + linear PSU.  While blinded, a group of 3 listeners could not state a preference for one source against another even though the optimized sources had the further advantage of being connected to a Son of Q balanced power supply (isolation transformer + RF filter) and benefited from expensive mains cables (Audience AU24SE or Digital Dynamics Challenger AE15 with active RF shield).  

Well, I have compared one more source because I couldn't leave well enough alone.  I was convinced I could find some source that the DAVE couldn't work its magic on.  I currently have on hand an Oppo BDP-105D which is an excellent CD/SACD transport.  Unlike typical CD/SACD transports, this unit is capable of Tidal streaming although those of you who have an Oppo know that Tidal streaming with the Oppo is utterly horrible.  While CD/SACD playback is reference quality, Tidal streaming sounds hollow and distant, worse than highly compressed MP3 playback from an iPhone. It is probably the worst implementation of Tidal streaming I have heard and I'm not sure why because the SQ I get from Netflix streaming from the Oppo is very good.  Well, I connected this Oppo unit to DAVE via my optical cable and as expected, CD and SACD playback is excellent, as good as any of the other sources I previously mentioned.  To my surprise (but maybe by now, I shouldn't be surprised), Tidal playback from the Oppo through the DAVE is also utterly spectacular and as good as any of the sources listed above.  If there was maybe a small iota of doubt I had before, that doubt is now gone.  Just feed the DAVE a bit-perfect file and find yourself rewarded.  I suspect you could stream a bit-perfect file wirelessly to the DAVE through Bluetooth and it would sound just as good.  Perhaps a feature to consider for the future?

Now what about DAVE itself?  As I shared with you, mechanical isolation matters.  What about the quality of mains power to the DAVE?  Here is the standard 18g US mains cable that came with my DAVE.  I had not used it on the DAVE until now:




Here is the $3,500 Dynamic Design Challenger AE15 mains cable I have been using with the DAVE.  




This mains cable is specifically designed for low power digital front ends like a DAC and has an active battery-operated RF shield and I can verify this shield works.  As you turn the shield from on to off with a switch, the music becomes less illuminated and probably has a similar impact as Shunyata's or Isotek's digital mains cables with RF shields.  Could the cheap 18g mains cable that comes with the DAVE sound as good as this expensive specialty cable?  Furthermore, with the DAVE, does a dedicated line conditioner/isolation transformer make a difference?

Well, here is my testing methodology:

1.  DAVE with cheap 18g mains cable > direct to wall VS expensive mains cable > direct to wall.
2.  DAVE with cheap 18g mains cable > direct to wall VS expensive mains cable > Son of Q (isolation transformer + RF conditioner) > wall
3.  DAVE with cheap 18g mains cable > Son of Q > wall VS expensive mains cable > Son of Q > wall

My listening room / home office has a dedicated AC line with good earth ground although this line also powers LED and incandescent lights in the room, a large LED television, several computers, computer monitor, scanner and printer, speaker amp and powered subwoofer.  My HE-1000 was used and connected to DAVE via DHC cable.  Two-channel was also evaluated.  Three tracks were chosen for their wide dynamic range, complexity and because I know them well.  I used (1) a DXD recording by 2L of Mozart's Requiem, a complex passage that contains amazing depth and a large chamber choir, (2) an organ improvisation of Deilig er Jorden (Fairest Lord Jesus), another DXD recording by 2L that I chose for its complex and rich bass dimensionality and (3) Beethoven's Piano Sonata No 29, a 16/44 recording performed by Alessio Bax, a young up-and-coming pianist I saw live a few months ago.

This is what I found.  While unfiltered power to components pre-DAVE (ie music server, CD transport) don't seem to impact the DAVE, good power to the DAVE does make a noticeable difference and while this difference is not as pronounced as I have heard with my other DACs, it is noticeable even with blind testing.

In each scenario, when DAVE was connected to the expensive mains cable, the volume seemed elevated and this is how I could so easily pick it out even when blinded.  I have noticed this phenomenon in the past with my other DACs and it isn't because the SPL is actually higher but more likely because the noise floor is lower and the dynamic contrasts are greater.  The highs are more extended, easily noticeable at the 45 second mark on Mozart's Requiem.  There is also more definition to the bass with the bass layers more finely delineated on the organ track.  On the Alessio Bax track, when he is hammering on the upper register keys of the piano, with the expensive mains cable, there is clearer delineation of the keys whereas with the cheap mains cable, there is more noticeable smearing.  On 5/5 tries, when blinded, I was able to accurately pick out the more expensive mains cable. 

When the cheap mains cable was kept plugged into the wall and the expensive cable was plugged into the Son of Q conditioner, this delta widened further although the benefit was less pronounced compared to the expensive cable by itself.  When the cheap mains cable itself was plugged into the Son of Q, the gap became narrower.  Both with and without the Son of Q, there was no noticeable compromise in dynamics, a claim made by some against line conditioners.  It became clear to me that both the isolation transformer/RF conditioner and the more expensive mains cable resulted in additive improvements essentially leading to a darker, more dynamic and refined presentation.  The effect was more noticeable on headphones than it was on 2-channel and I surmise that this is because "direct to DAC" is more capable of revealing improvements.

These are my personal observations about the DAVE so far:

1.  The digital file matters.  There is a clear improvement with my 24/192 and DXD recordings over 16/44 (I'm not talking about upsampled music but native recordings).  While PCM is technically superior to DSD, DSD and especially DSD128 can sound very good.  Unfortunately, since Roon cannot currently play DSD256 via DoP, I am unable to play my native DSD256 recordings and I am left wondering how good this sounds through the DAVE.  Regardless, I am pleased the DAVE is so versatile.

2.  While the digital file matters, the digital file server does not (and I never thought I would ever be saying that).  Simple bit-perfect delivery even with a cheap USB or optical cable is all that is required.

3.  The DAVE benefits from mechanic isolation, especially with 2 channel.

4.  The DAVE benefits from clean mains power.  While Rob has done a splendid job of isolating the DAVE against its own switching PSU, a mains cable with good RF shielding can make a noticeable difference.  While some believe connecting a DAC to a dedicated line conditioner can potentially rob it of some of its dynamic potential, with either the Audience aR6-TSSOX passive conditioner that I have on hand as well as the Son of Q transformer-based conditioner that I have decided to replace the Audience unit with, I have not found this to be the case.  With these types of devices, YMMV and will depend on how much noise is in your mains line and so try before you buy.  In my system, the DAVE with the cheap mains cable by no means sounds harsh or lifeless but my better mains cable + isolation transformer / RF filter clearly adds a desirable sense of refinement.

6.  While the quality of digital interconnects don't seem to matter, the quality of analog interconnects, headphone cables and speaker cables absolutely matter as the DAVE has the potential to reveal the best qualities but also the limitations and flaws of such cables.

7.  With the DAVE, headphones potentially will sound better than 2 channel or any scenario where you have to add an outboard preamp or amplifier.  At the present time, because most DACs do not have the ability to connect directly to a headphone without an intermediary amplifier, it is my belief that the DAVE is the best headphone DAC in the world for most headphones. The immediacy and clarity that this direct connection provides cannot be overstated and while I cannot speak for the preference of others, personally, I would try to find another headphone before I considered adding a headphone amplifier.  As for 2-channel, for me it is a forgone conclusion that I will be adding Chord's new amp to my system once it becomes available if it provides the same level of transparency I am getting now with my HE-1000 direct to DAVE.

Roman, thanks for that - great post! As soon as I get my Dave, and after I have run it in, I'll try some mains cable options. Cheers

 

[lovethatsound]

Romaz, another great series of posts by you,offering us Dave users valuable information and insight into the Dave.Are you sure you and Rob are not twins?lol,ANYWAY thanks alot.

 

[romaz]

romaz, I'd like to rock your world slightly:

http://people.xiph.org/~xiphmont/demo/1kHz-105-dither.wav

This is a -105dB 1Khz sine wave encoded in 16-bit WAV format at 48KHz sampling rate. On my Hugo TT with HD 800 S I have to go to almost maximum volume to be able to hear the tone. But there it is

Thanks for this but with my DAVE and HE-1000 (my HD800 S has not yet arrived) and at maximum volume level, I am unable to hear this tone.  Can anyone else with a DAVE hear it?

 

[JaZZ]

Yes, I can (just) hear it with a HD 800 and DAVE at 0 dB. At +19 dB it's clearly audible and even relatively clean. The same goes for the HE1000 (even though a bit lower in volume).