Chord Hugo

rob watts said:

  It won't derail the thread, as the experience gave me a direction for digital design.
 
In the '80s, there was a buzz going around that mains cables made a big difference to the sound, so I tried experimenting. Yes you could hear a difference, even under blind conditions. And this was with one of my pre-amps that had 3 stages of regulation - 210dB of rejection on the PSU. This clearly did not make sense. So I experimented, and found that the RF characteristics of the mains cable made a big difference. This then led to the design of an RF noise mains filter, and this ended up being a 12 stage filter, going from 100 kHz to 1 GHz. Then when I used the filter, the connection from the wall socket to the filter now had no effect on the sound quality. So this proved it was RF noise, and because of the observation that small changes had a big affect, it also indicated that the ear/brain was highly sensitive to it.
 
So what was going on? The RF noise intermodulates at RF frequencies with the audio component active stages, producing extra noise in the audio bandwidth -  thus creating noise floor modulation. It turns out that noise floor modulation is highly audible, as it messes up with the brains ability to separate sounds up into individual instruments. Reduce RF noise, reduces noise floor modulation, and it sounds smoother with much better instrument separation and focus.
 
Now with digital design there are countless ways for noise floor modulation to creep in, and I go to extreme lengths to reduce these effects, as well as reduce RF noise coming out of the DAC - hence the 2048 times digital filtering and pulse array noise shapers running at 2048FS. There are lots of ways for noise floor modulation to creep in.
 
So what levels are significant? I have had noise floor changing at -200dB, and it was audible. That's 32 bit accuracy. Once you start saying that the digital domain performance must be better than 32 bit performance, then that opens up a whole can of worms for ways for things to make a difference.
 
So messing around with mains cables ended up having a major influence in my analogue and digital design.
 
Rob 

Click to expand...

 
 
Thank you very much for sharing your insight on mains cable Rob! Always been curious as to the why!

perhaps the info you talk about is how the power cables help. al

alrainbow said:

rob this is a little off topic but it is about the hugo . i own an offramp 5 and there psu all the options . i have tried it with the hugo a two other fpga type dacs . it not does not improve it my question is why . it not only does not help it makes it worse.

Click to expand...

USB input has two variables that can change the sound - jitter and the amount of RF noise injected into Hugo. Jitter problem is solved within Hugo, so a separate USB unit will have no affect on jitter. As too RF noise, adding extra components will always add more RF noise, but there may be some benefit in isolating Hugo from the computer. It's pros and cons.
 
Anyhow, my job is to make Hugo sound fantastic whatever the quality of the USB, so I am pleased that USB add-ons are not helping!
 
Rob

clemmaster said:

  Would you mind sharing your thoughts on the OR5+Master-7 vs. Hugo as a DAC-only?

Click to expand...

I will try it. Hopefully next week or later.

alrainbow said:

I have the same combo. I can do it tonight . But from memory the hugo is much better. The M 7 is really nice DAC but lacks the clarity , separation , and deatils of the hugo. If you go from the hugo to the M7 it's not pretty . If you go from the M7 to the hugo it seems light but super clear .
Al

Click to expand...

Thanks Al.
 
Are the Lampi (4 and up), AMR and such THAT much better than a Master-7 + OR5?
What's the point of owning the combo if a standalone Lampi is better for roughly the same price?

LOL. As a 35 year audiophille with $250,000 system once said, " Power will make up for 60% of the sound of your system".

That same guy is broke and living under a bridge right now

clemmaster said:

  Thanks Al.
 
Are the Lampi (4 and up), AMR and such THAT much better than a Master-7 + OR5?
What's the point of owning the combo if a standalone Lampi is better for roughly the same price?

Click to expand...

 
 
 
I heard the Lampi 7 and man it was music magic. I said to myself I can listen to this **** all day long... and that rarely happens.

agb100 said:

 
A digital to analog converter's main job is to convert a digital signal to an analog signal. If it failed to do that it has failed, period. No amount of power supply capacitance, transformer, regulation, fancy output stage, tubes to tame the nasties, can correct for what is not there. It is, accordingly, NOT about how well the analog output stage is executed. In fact, that is about the last thing to worry about. Other than being body-snatched by extraterrestrials who may do sexual experiments on you,

Click to expand...

vert said:

   
I've actually had both the Chord QBD76 and a Naim DAC. The Chord uses FPGA, which you would consider superior than the Naim, correct? Since the Naim uses off the shelf chips? The Naim is more analog sounding than the Chord.

Click to expand...

 
Quote:

agb100 said:

 
I'm not the only one who considers FPGA (properly implemented) superior. It is audibly and measurably vastly superior to common dac chips.  The operative words are "properly implemented."
 
It is very easy to design a common converter, the literature not only exists in the public domain, but today a manufacturer can purchase ready-made solutions (almost complete DAC assemblies for pennies from the Orient), meaning that most DACs you buy and unknown to you, may contain non-OEM manufactured assemblies made to order. The NAIM DAC is simply not competitive in this league we are discussing at this forum, as good as it may be, and in my experience no DAC using off the shelf IC DAC chips is. They cannot do the necessary computations for accurate conversion, being limited in both design and chip real estate.
 
You should be clear that only a handful of designers are capable of, have the knowledge and resources for, working with FPGAs. And FPGA's are not perfect, they are perfectible. But at the bottom line, they can be dramatically superior.
 
As for: you say yours is more analog sounding, that may be true, but then I ask, do you own the HUGO or Direct Stream to come to that conclusion, or are your words just wishful thinking?
 
Lastly, let's say that the NAIM is more analog sounding. I would take that to be a serious problem because analog is not the standard for musical reproduction, unless you are into low resolution formats, as it is at best 12 bits for 1/4" tape and about 10 bits for LP. To prove to you how limited analog tape is in both dynamic and frequency range, just spin an original direct-to-disc on your platter and compare it with the commercial LP variant of the same. The improvement of D2D, as pure as analog can be, is rather dramatic. I can't make the conversion to bits or numbers off my head, but I'd assume D2D capable of 14 bits of resolution, about -80db s/n under the very best conditions on the very best equipment. LP is about 60-70. The Hugo may perform at 24 bits at -110-140dB; at 32 bits - the next level, maybe 180?
 
Moreover, a number of the best known proponents of the D2D medium have admitted over the years that digital recording (not playback) surpassed analog recording a long time ago. Bob Katz for example, very well known, has stated that for the first time audiophiles have, with digital, direct access to the master. The reason is because the digital transfer can be exactly the same as the master.
 
Analog accordingly is not the standard for modern sound reproduction. Digital is, because digital alone duplicates the data on the master.
 
If your DAC sounds "more analog" than a Direct Stream or Hugo, I'd sell if off pronto if I were you.

Click to expand...

 
I'm going to have to respectfully disagree with some of AGB100's statements or at least his conclusions (this kind of got long from trying to explain the difference in conversion methods, but #5 gives the dumbed down analogy and read the parts about FPGA.  I'm probably inaccurate in some of the details but I believe my overall thoughts are correct).
 
1) First off, the statement of DAC's job of reproducing the original analog signal being the primary job.  AGB100 concludes therefore that all other factors don't really matter if it fails at its primary job.  Therefore it is implied that these other things don't matter as much. He is correct in the statement but not necessarily the conclusion.  The reality is that ALL DAC's fail to truly reproduce the original music as they are all using some type of lossy method to record and convert back to analog sound.  However it can be argued that all modern DACs at this level succeed at reproducing this music.  Modern DACs, for the most part, have pushed various errors in their technique (Delta Sigmas for the most part) to outside the commonly understood audible level to one degree of another.  You could therefore conclude how you implement is actually the only thing that matters.  For example, Rob Watts argues (which I don't necessarily disagree) that the brain can actually process errors in noise shaping at a much higher resolution than what most scientists believe we hear as these scientists are simply measuring the capabilities of our ear vs brain and that the Hugo does a more accurate job than the modern DAC he uses as an example.  But please remember pretty much all audiophile level DACs are fighting over stats/figures that most audio related scientists would argue humans can't hear.  Audiophiles would argue that we can hear some differences and the various DAC manufacturers are asserting that one factor or another that is beyond what audio related scientists consider audible, is what is the difference.
 
2) Additionally, most audiophile level DACs designers would vehemently argue with AGB100 conclusion that the last thing you worry about is the various parts outside of the DAC chip (although the common definition of DAC chip is fraying) are "the last thing you worry about."  My guess is when Chord releases their desktop Hugo Dac, it will concentrate on these other factors like power supply etc.
 
3) vert makes a statement that he believes Naim DAC sounds more analog and implies that it sounds better to him.  AGB100 argues back that vinyl is much less resolving than what the Chord Hugo can do.  AGB100 is again right, but again not necessarily right about his conclusions.  The reality is that the 10 bit resolution that vinyl is capable of and the 32 bit that Hugo is capable are well beyond commonly understood abilities for humans to hear/process.  In fact, if AGB100 is correct about this resolution being the key factor for a while now in how music sounds, then all DACs for a while now (since they have been above 10bit resolution for a long time) should have been sounding better than vinyl.  I believe many audiophiles would argue that is not true.
 
4) The reality is that there are various techniques to reproduce music with their various positives and negatives.  Vinyl proponents would argue that vinyl is analog, as it the music itself, and while it is less resolving that most modern DACS, it is resolving enough and also captures something about the music that is lost in the digital translation.  The two main digital DAC techniques are R2R and Delta Sigma.  You will have to do your own reading on the two techniques and advantages/disadvantages of each, but I will make some gross arguments for each.  R2R (or Ladder) DACs use what seems to be a simpler technique, and R2R proponents would argue more likely accurate, to interpret PCM data into analog, but R2R Dacs become harder and harder to implement at resolution increases and more and more expensive.  Think of it as an assembly line in that the longer it gets, the more likely a mistake is made somewhere along the assembly line and the more expensive it is to make an accurate one.  R2R proponents would argue that R2R DACs have lost popularity because they have become too expensive to manufacture compared to Delta Sigma DACs.  The NAIM DAC uses the PCM1704 DAC which many consider the top R2R commercial DAC.  MSB actually makes their own customized R2R DAC Chip.  On the other hand, the Hugo is using Delta Sigma modulation.  Delta Sigma is even harder to understand.  Basically the PCM data of 1/0 representing the music is converted via Delta Sigma modulation, a complex math formula, into a series of pulses similarly to how the original ADC converted the analog music into PCM data.  However, Delta Sigma modulation is not accurate in reproducing the PCM data into the indidicated pulses.  This inaccuracy is reduced as there are more "taps," as Chord likes to call them, and theoretically is 100% accurate at infinite "tap" length.  If R2R Dacs are like an assembly line, the Delta Sigma Dacs are like a Savant able to rattle off Pi (3.14 etc.).  The "taps" are how many decimal places this savant can give you.  Since Pi is infinite, it would take a Savant able to give you the infinite answer to be 100% accurate.  Delta Sigma caught on as computing power, to perform the Delta Sigma conversion, became cheaper and cheaper and represent the vast bulk of Dacs made today.  Delta Sigma proponents would argue that the errors introduced by the Delta Sigma conversion are well beyond the level that humans can hear and that Delta Sigma modulations look more "true" to the analog signal as the Delta Sigma pulses resemble the original analog signal (btw, heart of the SACD vs PCM debate).  Btw, all these conversions with Delta Sigma as well as R2R introduce other possible issues like jitter, timing, etc. that I haven't even begun to cover.
 
5) So to simply, Vinyl, R2R DACs, and Delta Sigma (Hugo being one of them) use significantly different techniques to regenerate the original music. It would be like comparing different fruits (apple, orange, pear) and arguing one being better than another.  You can try to measure sugar levels etc.and say one is better than another, but personal tastes matter just as much.  Now, this isn't to say that at times, most people wouldn't say one particular fruit is better than another.  Currently, I believe the differences of the qualities of the various fruits is close enough that personal taste plays a big part and that there is no clear winner.
 
6) AGB100 also asserts that the Chord Hugo measures and audibly is better than off the shelf DACs.  This is just plain untrue.  I have only seen one chart from Chord itself that claims better results around noise shaping against an unnamed off the shelf DAC.  I haven't seen common measurements against top end off the shelf Delta Sigma Dac chips (like the ES9018) by neutral 3rd parties.  If manufacturers are to be believed, the ES9018 claims better gross measurements than the Hugo (ES9018 135db Dynamic Range with THD of .0001% vs Hugo claims of 120db Dynamic Range and THD of .0005%).  As for the Audibly better, I think that is a matter of personal preference.
 
7) AGB100 implies something magical about FPGA based DACs.  To be clear, a FPGA is simply a programmable chip.  The Hugo was a leap in performance in large part to new FPGA chips that were able to do more real time calculations while using low power which allowed Hugo to employ more "taps," than ever before while using significantly lower power (hence the possibility of mobile form).  However, the state of the art FPGA is still slower than state of the art non-FPGA chip.  The new FPGA's just narrowed the gap in performance.  I can't however speak specifically about whether or not the Hugo FPGA is more capable than the hardwire chip the ES9018 uses (as I believe the ES9018 is proprietary and details unreleased) for example in terms of what Chord calls "taps."  I would suspect that the Hugo still isn't as capable as ES9018 "taps" wise given the gross claimed measurements but, if you look at qute and hugo stats, you can see that it has narrowed significantly.
 
8) Chord's overall approach is custom in which it has taken control of every step of the Delta Sigma process vs. the black box approach, for much of the process, of using off the shelf DAC.  Since Delta Sigma relies on complex math equation, the new capable FPGA chips have narrowed the gap in computing power that I believe that a custom approach to Delta Sigma could result in a tangibly different sound (and hopefully better).  Since Chord has lots of practice producing custom Delta Sigma designs, it is no surprise the Hugo has been as successful as it has been.  But let's not forget it is likely that the ES9018 has a lot more research dollars/man hours involved it in and likely measures better on gross measurements.
 
For me, it isn't that the Hugo is so much obviously better than ES9018 based DACs.  In fact, I feel that good ES9018 implementation might be a bit more resolving than the Hugo.  It is more to do with that all the ES9018 based DACs that I have heard have similar voicing/sound of which some are better than others and of which I have or had several.  I have been listening to ES9018 DACs for a while now so I have very little interest in another ES9018 based DAC.  On the other Hugo, the Hugo sounds different to me.  ES9018 sound more "digital," for lack of a better word, while the Hugo sounds more natural to me.  Hugo actually sounds more like a R2R Dac in its musicality.  R2R Dacs are significantly different architecture and sound.  It is getting increasingly hard to find R2R Dacs as they are expensive and complicated to manufacture and implement.  Choices like the Linn mentioned or the MSB or the Lampziator are either using NOS R2R dacs or designing their own R2R Dacs.  Often times the ES9018 will measure better than these R2R dacs on gross measurements like dynamic range and THD, but somehow R2R dacs sound more natural to me in the same way Hugo's do.  Other DACs like the PS Audio DirectStream use completely custom and different approach using a DSD focused conversion techniques and essentially converting PCM to DSD internally.  In any case, I think one of these custom approaches or R2R approaches can result in the very least, a different sound than off the shelf Delta Sigma Dacs, and hopefully a more likely better sound.

rob watts said:

USB input has two variables that can change the sound - jitter and the amount of RF noise injected into Hugo. Jitter problem is solved within Hugo, so a separate USB unit will have no affect on jitter. As too RF noise, adding extra components will always add more RF noise, but there may be some benefit in isolating Hugo from the computer. It's pros and cons.

Anyhow, my job is to make Hugo sound fantastic whatever the quality of the USB, so I am pleased that USB add-ons are not helping!

Rob

Click to expand...

Hahaha yes I am happy too. There was a time that these devices did not perform so well and Needed the help. I am just wondering why it actually make sit loose its mojo .
Even the direct stream doe the same thing. Anyway thanks and enjoy your profound of improvements and lower price point .
I am happy .
Al

rob watts said:

  Sure, lots of things happening! Been doing huge number of listening tests as we speak... 
 
Rob

Click to expand...

It'd be great if we could buy a Hugo without a battery, the same dac, same size but for use at home with a computer. Something to upgrade from my Meridian Explorer. Maybe Nogo?
 
I know Hugo can already do that but I don't need a dac with a battery really, and I'm afraid the battery might die in a few years.

I think the battery inside the DAC would be good for a few years at least . Relax and buy one you'll be happy .

I fly rc helis. And the batteries last a while no worries. .
Al

The batteries can be changed by the dealer when they died.

While it's nice that it seems that Chord has produced an apparently stunningly sounding DAC, I'm sure the engineering needed to achieve this extreme miniaturization here adds to the production cost.
(Maybe even special engineering for heat dissipation or cool-operation is required...)
 
And it's also a problem because it may prevent the typical audiophile purchaser of this unit from using the expensive cables they probably already own.
And since the DAC is the crown jewel here and the cost is high, if ever a lower-cost DAC-only offering was called for, this is it.
 
So I hope Chord sees fit to produce a lower cost version of the Hugo (a "Huge Hugo" maybe) with the option of just a DAC.
 
(Afterthought: maybe a larger, battery/power-supply could also be used in a slightly larger unit, increasing battery life or even improving performance -- just speculating here.)

Hi to all Hugo fans.
Long time reader of this thread and others around the net about this "Hugo" that is said to be so awesome.
 
I just turned 40 and thought about what kind of 40 year crisis i would have and be spending money on, a motorcycle, new car...
I decided it to be the Hugo and a pair of Ed8 Romea ultrasones (yes i know there might be others that are better, but i look for portability, zero leakage and isolation), also when turning 40 you can spend money on things that you wouldn't do otherwise so... Here i am.
 


Just go the Hugo today and currently playing songs in sqeeezeslave via USB from my macbook pro through my EX1000 IEMs. So far plain old 16 bit FLACs.
 
I'm not a person that analyzes things into tiny details or can talk about the many different technical terms, however what i can say is this. Music has never sounded better to me.
It's not that i hear details that i have not heard before, the main thing for me is that the sound, be it a guitar or someone singing has never sounded more RIGHT! And that is what it's all about i think!
 
So thank Chord for making this little amazing thing (smaller than imagined but heavier), and thanks to everyone here that has been posting impressions which made me take the plunge.