[Crashem]

Any word on people with old cases and what options we have?

 

[Ross]

   
Everything can be reconstructed, provided your interpolation filter has enough taps to provide the necessary approximation of the sinc function (Nyquist-Shannon sampling theorem).
That's where the new Spartan 6 FPGA are like a breakthrough for audio processing: the advances in lithography made it possible to dramatically increase the number of gates while reducing the power consumption -> welcome Hugo!
 
Cost is of course a factor: the Spartan 6 tray costs < $100.
The Virtex 7 (flagship) features 13 times more gates but cost > $4,000...
 
We can just hope Xilinx and Altera will continue developing their process and offer better FPGA for cheaper in the future...

 
Thanks. I'm sure you're right but this still doesn't make sense to me, probably due to my miniscule understanding of sampling theory. If there is no information sampled at above half the sampling frequency, then the interpolation filter should be able to reconstruct the original waveform. However, assume sampling at 44.1, if a transient takes place between two samples, such as the initial thwack of a cymbol - ie within the 22uS gap - an interpolation filter will be able to interpolate from the samples on either side, but will miss the intial point of impact. It may correctly reproduce the frequency response, but not the timing of the orignal impulse, which has not been sampled. No doubt there is something I am failing to understand ...?

 

[barbes]

Pool:  [COLOR=B22222]BEST headphone for HUGO[/COLOR]
*********************************************

Now that I finally received my HUGO, and was told by Rob NOT tu use any external Amp  , I need to choose a "non leaking" headphone that has the best possible synergy with the HUGO  (I already have TH-900 & Roxane that pair very well with Hugo  -)

[COLOR=B22222]--> Anybody tested  AlphaDOG,  SignaturePRO,  or other TOL "isolating" headphone with the HUGO ? [/COLOR]



Any, on a wider scope, any other headphone recommended that offer even more Grunt & Musicality with the HUGO...

The Alpha Dogs sound fantastic.

 

[SoundFreaq]

SigPro sound fantastic.

+1 on the AP2

 

[WizardKnight]

Peter at Double Helix Cables just made my USB cable for the Chord Hugo and sent me some pics. Also a while back he recabled some of my headphones with Audeze connectors so my headphone cable can plug into any Audeze headphones. My recabled headphones sound superb. Will post some pics of my new USB cable & my recabled headphones below.







If anyone has any questions you are more than welcome to email Peter at Double Helix cables at the address provided below.

doublehelixcables@gmail.com

 

[SoundFreaq]

Peter does fantastic work. He's passionate about his work, and produces some of the best and most well thought out cable designs in the industry. Plus, his cables are just beautiful, and exude quality and elegance. A lot of times you'll get a great cable that doesn't look as great as it performs, but with Double Helix, you get fantastic looks and top performance. 

 

[Rob Watts]

 
Thanks Rob.  So a follow up:
* many of us who have reviewed/evaluated the Qute DACs (HD and EX) have found that we enjoy the sound from a USB/SPDIF converter better than direct through the USB receiver on the Qute (especially true on HD where a UK Hifi magazine measure the USB output has having significantly more measured jitter than the SPDIF..although still within reasonable spec).  Do these reports make any sense to you?  If so, would the same hold true for the Hugo (one dealer/user, Richard/Aumamp has already reported he thinks the Hugo goes to another level using an Audiophileo AP2 and a good linear ps for said converter)?  As another data point, I have recently been able to send the EX unpowered USB and subsequently found the differences between it and the SPDIF not a real issue anymore (which tells me maybe we were hearing rf noise, not jitter).  FYI.

Yes Qute and Hugo both share isochronous USB, so timing comes from the DAC, not the computer, so any sound quality differences is categorically not jitter. Your right, my explanation too is RF noise, as this is the only thing I can think of that can upset the sound, as it's not jitter and it is bit perfect (otherwise DoP simply would not work). Moreover, the SQ improvements sound exactly like reduced noise floor modulation due to lower RF noise - smoother sound with more body. These reports are very useful, as it helps me very much on new projects. 

 

[Rob Watts]

Two questions for Rob:

What's the best battery powered source you've used with the Hugo?
What's the best source you've ever used with the Hugo?

I'd expect the latter to be a Chord product of course, but I'm curious about the first.

 
I am not the right person to ask, as I have not specifically tried lots of things. When I am on the go, I use a regular lap-top. Now I have heard lots of sources, (not in an AB sense so please take this comment with a pinch of salt) but Hugo just sounds like Hugo whatever the source.

 

[Rob Watts]

   
Rob, a question on your answer above.
 
You say that redbook is capable of better timing, possibly exceeding DSD. Doesn't this depend on whether the information is in the initial recording? If a recording has been made at 16/44 for example, is the timing information still present in the original recording able to be reconstructed below 22uS? Surely the smallest interval recorded is 22us and therefore transients smaller than this will be lost and cannot be reconstructed ...?

Yes good point. The key is that the sampled data must be bandwidth limited to 22.05kHz, so there is no information of smaller than 22uS samples to start off with. Then if you use an infinite tap length FIR interpolation filter the original bandwidth limited signal is perfectly reconstructed. This also means if a transient starts half way through a sample, then the reconstructed signal will also start half way through. But to do this, you must have infinite taps for perfection. as soon as you start reducing the number of taps, then the precision of the timing of the transient in the middle falls off. I know these concepts are difficult, I have been thinking along these lines for 30 years, and I still find it hard!  

 

[jamato8]

Playing some Dire Straights and the sound is wonderful (another dimension to music I thought I knew very well). Surprising how good red book can sound after all these years. 

 

[Ross]

Quote:

  Yes good point. The key is that the sampled data must be bandwidth limited to 22.05kHz, so there is no information of smaller than 22uS samples to start off with. Then if you use an infinite tap length FIR interpolation filter the original bandwidth limited signal is perfectly reconstructed. This also means if a transient starts half way through a sample, then the reconstructed signal will also start half way through. But to do this, you must have infinite taps for perfection. as soon as you start reducing the number of taps, then the precision of the timing of the transient in the middle falls off. I know these concepts are difficult, I have been thinking along these lines for 30 years, and I still find it hard! 

 
Thanks, I think I understand that. But isn't this where theory departs from practice? In reality, there will always be information above 22.05Khz (or half whatever sampling frequency is used), and artificially limiting bandwidth at the recording stage means that you will always exclude transients which start between samples. The only way to avoid this is increasingly high sampling frequencies. Transients lost during early digital recordings at 44.1khz will never be able to be reconstructed even with infinite taps, which is one reason why early digital sounds like un-lifelike.

 

[Rob Watts]

   
Everything can be reconstructed, provided your interpolation filter has enough taps to provide the necessary approximation of the sinc function (Nyquist-Shannon sampling theorem).
That's where the new Spartan 6 FPGA are like a breakthrough for audio processing: the advances in lithography made it possible to dramatically increase the number of gates while reducing the power consumption -> welcome Hugo!
 
Cost is of course a factor: the Spartan 6 tray costs < $100.
The Virtex 7 (flagship) features 13 times more gates but cost > $4,000...
 
We can just hope Xilinx and Altera will continue developing their process and offer better FPGA for cheaper in the future...

Yes performance goes up, costs come down, unlike everything else in the high-end! Maybe we will be talking about 10M taps in 10 years time. At some point though it will cease to make a difference, but I don't know where this point is.
 
When reading this post I was struck by two contradictory emotions - very pleased that you appreciate the number of taps argument, but a tiny bit bothered by the implication that it is all down to FPGA's, which of course in one sense is true. But it is also about appreciating that tap length and timing is important. I have been banging on about the importance of timing and tap length since late 1990's, when the DAC64 came out, and I have been a lonely voice, with a lot of opposition from qualified engineers saying I am talking nonsense. Nobody else talks about filter timing, nobody else talks about the importance of tap length. Indeed, the opposite is true - the high-end keeps banging on about the importance of no pre-ringing filters, completely the opposite to my thinking. 
 
The good thing about Spartan 6 and Hugo is that I can now do very serious tap length WTA filters and make it available at a reasonable price.
 
And Hugo talks for itself.    

 

[Rob Watts]

  Quote:
 
Thanks, I think I understand that. But isn't this where theory departs from practice? In reality, there will always be information above 22.05Khz (or half whatever sampling frequency is used), and artificially limiting bandwidth at the recording stage means that you will always exclude transients which start between samples. The only way to avoid this is increasingly high sampling frequencies. Transients lost during early digital recordings at 44.1khz will never be able to be reconstructed even with infinite taps, which is one reason why early digital sounds like un-lifelike.

Yes there will be info above 22.05 kHz - but I don't think you can hear it! So when I talk about transients, its about the timing of the transients, not the speed of the rising edge of the transient. I can't hear above 16 kHz, so info above this frequency will be ignored. But the timing of what I can hear is vital as we can perceive 4uS.
 
But like you I am bothered about the recording side of things, so I have been working on a pulse array ADC. This will have proper aliasing filters, and multiple OP's at different sample rates, so from the same recording and equipment one will be able to see what the true affects of sample rate is. This won't be available for some time though.

 

[Neil08]

  Love my Hugo but why is Bluetooth input so damm noisy? Does not matter what Bluetooth device used I still get a lot of  noise.
 
Anyone else getting  the same problem?

 
My Hugo has finally arrived and I'm having the same problem. The bluetooth input is unlistenable at times. Especially on quiet classical passages. 
 
Unfortunately for me, it is the input I will be using the most until I sort out a source and cabling. 
 
I've asked my dealer to chase Chord on this. Perhaps Rob Watts can comment?

 

[Ross]

  ...when I talk about transients, its about the timing of the transients, not the speed of the rising edge of the transient...
 

That is exactly what I mean. Music is not a series of sine waves, but a series of transient impulses followed by decaying harmonics. Capturing the transient impulses seems to be the one thing that digital recordings have not been able to do well, and is one reason why vinyl remains popular, because it does seem to reproduce transients with more fidelity. Someone insightfully defined music as "the organisation of time", which may be why digital recordings sometimes sound less like real music than analog recordings, because time has become slightly "dis-ordered". Correcting this at both at the recording stage with the ADC as well as the playback stage will hopefully mean that we have digital sound that is unquestionably better than analog in all respects.
 
PS The more I listen to the Hugo, the more I appreciate how fundamentally different - and better - it is than any digital I have heard. It seems so much easier to forget the equipment and enjoy the music!