CHORD ELECTRONICS DAVE

[GraveNoX]

I tested optical vs USB on 3 different days on Hugo 2, optical sound better, smoother, less grainier, cleaner and clearer than USB. Easily noticeable. My USB's power are way cleaner than before, but still not enough to catch the optical. Measurements may show same results, but listening is all that matters in audio world.

 

[yaccobo]

I am ready to purchase a DAVE, the only thing holding me back is the fact that the DAVE was lauched already 2.5 years ago, and that in the meantime there were other developments improving on the sound quality (like the M Scaler in the Blu2 CD-transport). I do not want to invest in Blu2 as i don't own any CDs anymore.

I know that this is a contentious question and that the company cannot divulge their 2018 launch plans, but can the users here who follow the discussions about Chord technologies more closely than I offer any thoughts whether purchasing a DAVE at this point would be a risk.

I am just afraid that perhaps a successor to DAVE, with better sound quality, will be announced later in 2018. Are there any hints that perhaps a dedicated M Scaler may be released that could be used in conjuction with DAVE to increase the sound quality?

I know that there is always an intrinsic risk in purchasing gadgets in a rapidly developing technology area since nobody has a crystal ball, but the investment in DAVE is quite substantial, therefore my reservation in investing into a 2.5 years old technology.

 

[Triode User]

I am ready to purchase a DAVE, the only thing holding me back is the fact that the DAVE was lauched already 2.5 years ago, and that in the meantime there were other developments improving on the sound quality (like the M Scaler in the Blu2 CD-transport). I do not want to invest in Blu2 as i don't own any CDs anymore.

I know that this is a contentious question and that the company cannot divulge their 2018 launch plans, but can the users here who follow the discussions about Chord technologies more closely than I offer any thoughts whether purchasing a DAVE at this point would be a risk.

I am just afraid that perhaps a successor to DAVE, with better sound quality, will be announced later in 2018. Are there any hints that perhaps a dedicated M Scaler may be released that could be used in conjuction with DAVE to increase the sound quality?

I know that there is always an intrinsic risk in purchasing gadgets in a rapidly developing technology area since nobody has a crystal ball, but the investment in DAVE is quite substantial, therefore my reservation in investing into a 2.5 years old technology.

Only Chord can answer and they will play their cards tight to their chests for obvious reasons.

My impression is that they have got other launches to occupy them for at the rest of 2018 and that whilst Dave is still at the top of it's game it is unlikely to be succeeded. All the signs are that any standalone MScaler is also way off.

Every now and then a second hand Dave pops up so that would minimise any financial risk.

 

[rayl]

FWIW I had similar qualms about Blu2 vs maybe a forthcoming Mscaler stand-alone but took the plunge... just ahead of the NAmerican price increases.

Main thinking is that the combo is quite advance... I haven’t heard of any breakthroughs in the works... the only potential flaw is the RF from Blu2 entering via S/PDIF and reports here are that ferrite chokes work to an extent. So I felt that the better is the enemy of the good and placed my order. Life is always short.

The wait here was 5 wks DAVE (actual, as it arrived, the estimate was 4), 14 estimated for Blu2. I am guessing until the Blu2 backlog numbers come down, nothing new will be offered... and the intros at CES are now done so you have more info than I did.

And if it goes past Brexit, is that a win (lower price) if there is a major GBP depreciation or a loss if no deal with whomever and WTO terms come into effect .... so many unknowns.

 

[ZappaMan]

I’m in a similar position.

 

[Rob Watts]

Thanks Rob!

Based on this, then if Dave were directly powering PASSIVE speakers, i.e. only Dave's power supply involved therefore there's no ground or power supply 'loop' , then could Dave's housing THEORETICALLY be made of plastic and still suffer zero effects from airborne RFI, even if a mobile USB source was used with all it's wireless radios active on full blast??

And the same for someone using Dave purely with headphones with a mobile USB source?

Before anyone falls off their chair reading this, I'm not suggesting Dave should be made of plastic . I'm just trying to use this example to close the loop (pun intended) on my understanding of when airborne RFI is of concern, and when it's not.

Cheers in advance.

No then external RFI would create currents on Dave's ground plane - it is the solid block of aluminum that screens the PCB ground plane from external RF.

 

[Rob Watts]

I am ready to purchase a DAVE, the only thing holding me back is the fact that the DAVE was lauched already 2.5 years ago, and that in the meantime there were other developments improving on the sound quality (like the M Scaler in the Blu2 CD-transport). I do not want to invest in Blu2 as i don't own any CDs anymore.

I know that this is a contentious question and that the company cannot divulge their 2018 launch plans, but can the users here who follow the discussions about Chord technologies more closely than I offer any thoughts whether purchasing a DAVE at this point would be a risk.

I am just afraid that perhaps a successor to DAVE, with better sound quality, will be announced later in 2018. Are there any hints that perhaps a dedicated M Scaler may be released that could be used in conjuction with DAVE to increase the sound quality?

I know that there is always an intrinsic risk in purchasing gadgets in a rapidly developing technology area since nobody has a crystal ball, but the investment in DAVE is quite substantial, therefore my reservation in investing into a 2.5 years old technology.

I have many very interesting projects that I am currently working on - none of them include Dave 2. So assuming that the FPGA is not suddenly discontinued (extremely unlikely, Xilinx has kept production running for many years for old families - and the FPGA in Dave is still listed as current family), you are safe in that a replacement is not on my current horizon.

 

[yaccobo]

I have many very interesting projects that I am currently working on - none of them include Dave 2. So assuming that the FPGA is not suddenly discontinued (extremely unlikely, Xilinx has kept production running for many years for old families - and the FPGA in Dave is still listed as current family), you are safe in that a replacement is not on my current horizon.

thank you Rob, this reassurance is music to my ears, but my wallet is already in a panic mode as my last reservation to investing into a DAVE is now obsolete

 

[Mojo ideas]

I have many very interesting projects that I am currently working on - none of them include Dave 2. So assuming that the FPGA is not suddenly discontinued (extremely unlikely, Xilinx has kept production running for many years for old families - and the FPGA in Dave is still listed as current family), you are safe in that a replacement is not on my current horizon.

Also I can catagorically state Dave, as is

I have many very interesting projects that I am currently working on - none of them include Dave 2. So assuming that the FPGA is not suddenly discontinued (extremely unlikely, Xilinx has kept production running for many years for old families - and the FPGA in Dave is still listed as current family), you are safe in that a replacement is not on my current horizon.

Dave is and will remain technically the best Dac on

Only Chord can answer and they will play their cards tight to their chests for obvious reasons.

My impression is that they have got other launches to occupy them for at the rest of 2018 and that whilst Dave is still at the top of it's game it is unlikely to be succeeded. All the signs are that any standalone MScaler is also way off.

Every now and then a second hand Dave pops up so that would minimise any financial risk.

Dave will remain technically the finest Dac on the planet for many more years to come and as long as we can keep up production we will. Though Rob is sometimes musing on an Ultra Ultra high end Dac with a stupid amount of digital stuff inside and of three things I’m sure, it’s likely to be done one day. Boy is going to be costly, it would have to be. Finally it would take many years to develop and code. Besides all that Rob is very busy with other exciting stuff just now that far more people will get enjoy. So Dave will remain preeminent for a very long time.

 

[onlychild]

Yes optical is the best input - its really a question of getting the other inputs to match optical.

Not sure if I'm understanding this correctly, but is the consensus now that optical is the best input for DAVE vs USB?

I remember reading many posts from Rob stating USB was a better input for DAVE than optical.

 

[Crgreen]

With certain hi-res files, optical isn’t an option, and where you have to use some kind of USB to optical conversion, is there any real benefit? Other than with a CD transport, is optical really practicable or worthwhile?

 

[Rob Watts]

Not sure if I'm understanding this correctly, but is the consensus now that optical is the best input for DAVE vs USB?

I remember reading many posts from Rob stating USB was a better input for DAVE than optical.

Actually we are talking tiny differences - a battery powered lap-top via USB sonically is as good as optical, plus USB handles everything. When I do critical listening I pull out the power on my lap-top - but the difference is so small it's something I do not bother with when actually listening to music.

And yes I used to think that USB had the edge on optical - it's slightly brighter - but I was wrong about that, the brightness is not better timing as the clock comes from Dave, but worse RF noise pick-up. This gets eliminated when the lap-top is not connected to the mains or not grounded.

I will be visiting @romaz soon, so I can test out whether something else is going on with the work he has put into sources. It's always possible that other things are happening....

 

[onlychild]

Actually we are talking tiny differences - a battery powered lap-top via USB sonically is as good as optical, plus USB handles everything. When I do critical listening I pull out the power on my lap-top - but the difference is so small it's something I do not bother with when actually listening to music.

And yes I used to think that USB had the edge on optical - it's slightly brighter - but I was wrong about that, the brightness is not better timing as the clock comes from Dave, but worse RF noise pick-up. This gets eliminated when the lap-top is not connected to the mains or not grounded.

I will be visiting @romaz soon, so I can test out whether something else is going on with the work he has put into sources. It's always possible that other things are happening....

Thanks for the quick response Rob!!

I've got an Innuos Zenith SE arriving next week and its USB only, but I'm not concerned with tiny differences if it means a cleaner and simpler setup like the Innuos direct to Dave provides.

 

[Jawed]

Very interesting results, it's good you have spent the time gathering data.

Jawed, where do you think the bottleneck is?

You got me curious to find out exactly how expensive 1MM taps is. It certainly isn't the bottleneck.

The bottleneck is that it's not commercially realisable as a stand-alone DAC. The next step from Blu 2 is a million taps in a device of the size, portability and power consumption of Hugo or Mojo. Nothing else is worth doing. At least not for Chord. Some other competitor? Sure, they can try. If it sounds better than DAVE or Hugo 2, then yay.

The power consumption of a general purpose processor to run FFT based convolution is the plainest bottleneck. There's also the matter of latency. Latency can be reduced by running a partitioned convolution algorithm, but then power consumption goes up. This page is very interesting (I got there via your link to Convolver):

https://www.ludd.ltu.se/~torger/brutefir.html#whatis

on the subject of performance, latency and practicality. As you can see from that page, long FIR has been practical in real time for 15+ years on an ordinary consumer PC.

When running a million taps you need to use 64-bit (double precision) arithmetic. If you want to introduce GPUs as a solution, you have to bear in mind there's about 5 chips out there that aren't absurdly slow when running double-precision. Most of the GPUs are 1/16 or 1/32 rate on DP, entirely nullifying any theoretical benefits they might have over an ordinary CPU. Also, the fast double-precision GPUs are silly expensive, etc.

If your test processor was a phone - which could represent a device that replaces Hugo or Mojo with built in 1 million taps processing - then whatever numbers you produced would be more interesting! See if it could run for 5 hours, say, off battery...

As it happens, back in 2016 we had a discussion about FFT-based algorithms:

https://www.head-fi.org/threads/chord-electronics-dave.766517/page-273#post-12778587
https://www.head-fi.org/threads/chord-electronics-dave.766517/page-275#post-12782076

That discussion obviously pre-dated Blu 2, which moved the goalposts rather dramatically.

When you have a specific algorithm to run, a general purpose processor (CPU or GPU) is going to be worse than an FPGA and utterly comical in the face of an ASIC. That's why so much AI research is focussing on FPGA and ASIC. It's why investment banks use FPGAs not GPUs for their high performance, low latency, workloads.

GPUs are great for proof of concept, because of their programmability and dense compute capability. Programmability could be attractive to Rob (it's likely easier to make C work than an FPGA when building a million taps) but that doesn't get you a device with a million taps in a portable replacment for Hugo.

It's worth noting that GPUs have hardly progressed in compute capability in the last 9 years. Factor of approximately 5x is not very impressive, compared to the previous 9 years:

https://techreport.com/review/17618/amd-radeon-hd-5870-graphics-processor/5
https://techreport.com/review/17618/amd-radeon-hd-5870-graphics-processor/5
And that chip has very high double-precision capability (1/4 rate) and to get 11x today requires Volta V100

https://www.anandtech.com/show/1136...v100-gpu-and-tesla-v100-accelerator-announced

which took 8 years with a chip that's 2.4x bigger. So in effect it took 8 years to get 4.5x higher capability in the same chip area (cost) though it is probably half the power. Though this GPU (in a consumer card that's just been released) is about 4x the cost of the card from 2009, adjusted for inflation.

So if you have a DSP algorithm you want to deploy to 10s of thousands of customers, an FGPA (or a grid of them, they're usually small and low power) is going to be preferable to a GPU. If you have a one-off application installation or you're changing the application all the time, then sure, use a GPU. Or a supercomputer full of them.

---

You can find some of my GPU related coding here:

https://github.com/JawedAshraf

though it's for noise reduction in video, not audio processing.

 

[Jawed]

Yes optical is the best input - its really a question of getting the other inputs to match optical.

Glad we agree on this!

Now playing: Leif Ove Andsnes - Sibelius - 6 Bagatelles for Piano, Op. 97: Lied, No. 2