[mikey8811]

Hi Torq
 
I just tried it myself.
 
I just compared the Direct Stream with Torreys upgrade to a Cary DAC 200ts. I  found the Direct Stream a little sibilant and  lean . Don’t get me wrong, I liked it in its own way – there was more openness, inner detail and microdynamics were better, especially with strings and the way instrumental sounds decayed. There was also sweetness in the upper mids that gave some vocals a pleasing lilting quality. Soundstaging and imaging were pretty good, as was the layering of instruments. Vocals were clear and reached out into the room. With the Cary, the vocalist had more body and presence but vocals were not as clear, open and had less inner detail.
 
However, I missed the full bodied sound of the Cary and the slam it had.
 
Quote:

   
No, I didn't.
 
I can't see anything interesting about it, let along interesting enough to warrant it's price tag.
 
Most of the "clever" features I'd never use, so that doesn't help much there.
 
Beyond which, while when I started out on this path I was quite happy the AKM 4490 as a converter, at least lower down the price spectrum, the longer I've lived with one the more I'm thinking I actually prefer the 4396 in several areas - despite the newer chip measuring better.  We'll see whether that impression remains when I do my side-by-side all-up Bifrost variant comparisons.
 
Also, I'm just not adding any more DACs to the list this go around - if it's not listed in the first post at this point, it's not going to be (and the only exception I'd make is for a device that's doing something very different to everything else - on the order of the big Schiit, PS Audio and Chord units).  CotS-chip based DACs really aren't fairing that well in this line up.

 

[Torq]

 

Hi Torq
 
I just tried it myself.
 
I just compared the Direct Stream with Torreys upgrade to a Cary DAC 200ts. I  found the Direct Stream a little sibilant and  lean . Don’t get me wrong, I liked it in its own way – there was more openness, inner detail and microdynamics were better, especially with strings and the way instrumental sounds decayed. There was also sweetness in the upper mids that gave some vocals a pleasing lilting quality.
 
However, I missed the full bodied sound of the Cary and the slam it had.

 
Interesting.
 
I'd expect a "full bodied" presentation from AK4490 based units as it seems they were deliberately tuned that way as part of their "Velvet Sound" tuning philosophy.  I'm finding, in general, a bit of bloom with units running that chip, which can be quite pleasing in some situations, but isn't really my thing overall.
 
Normally I'd put that sort of response (mine, not yours) down to expectation bias ... but when you've observed it yourself, across more than one DAC, and prior to knowing there was any technical difference, and then had the manufacturer publicly state that they engaging in that sort of tuning exercise, it's a bit harder to dismiss.
 
The PS Audio stuff is interesting to me as much due to the potential that exists with their field-upgradability as anything to do with the sound they produce currently.  Which is something none of the CotS chip converters offer (unless they're implementing their filter stages via DSP or FPGA).  Of course, there's no guarantee that the next OS update will improve the sound, or, I guess, that there'd ever be another release ... but the possibility is there.
 
While the Cary will stay off the list, I do appreciate the update!

 

[Baldr]

An interesting CotS point: I recently, for the sake of completion, evolved a reference design DAC from a chipmaker I had never considered. When complete, I was shocked at how the sound differed from all published/3rd party impressions of this chipmaker’s family of DACs. Puzzled, I then obtained and read the CotS’ evaluation board schematics, only to find a suggested post DAC analog filter which was significantly different than the one I heuristically developed. Further reading proved that identical eval board filter was suggested/implemented across their line of DACs. From this, I speculate the converters using this chip IMHO could be made to sound/measure much more to my liking.

 

[artur9]

  An interesting CotS point: I recently, for the sake of completion, evolved a reference design DAC from a chipmaker I had never considered. 

I think it would be kind of fun to listen to your mad scientist experiments.  

 

[mikey8811]

Originally Posted by Torq /img/forum/go_quote.gif
 
I'd expect a "full bodied" presentation from AK4490 based units as it seems they were deliberately tuned that way as part of their "Velvet Sound" tuning philosophy.  I'm finding, in general, a bit of bloom with units running that chip, which can be quite pleasing in some situations, but isn't really my thing overall.
 
Normally I'd put that sort of response (mine, not yours) down to expectation bias ... but when you've observed it yourself, across more than one DAC, and prior to knowing there was any technical difference, and then had the manufacturer publicly state that they engaging in that sort of tuning exercise, it's a bit harder to dismiss.
 

 
Bloom is my sort of thing. I didn't think of DAC selection by the chip it uses as I am not as informed as you are. I attributed the bloom to it being a tubed unit but when I switched it to solid state output (it does both) the bloom was there all the same.
 
Can you suggest another DAC using the AK4490 or better, with that sound trait which retains the detail, open and natural sounding qualities of the PS DS?

 

[Torq]

   
Bloom is my sort of thing. I didn't think of DAC selection by the chip it uses as I am not as informed as you are. I attributed the bloom to it being a tubed unit but when I switched it to solid state output (it does both) the bloom was there all the same.
 
Can you suggest another DAC using the AK4490 or better, with that sound trait which retains the detail, open and natural sounding qualities of the PS DS?

I think a lot of people find it (bloom) very pleasing ... both listeners and manufacturers, which is probably how it got tuned that way.  It's a mood thing for me ... sometimes I want it, sometimes not.  Nothing wrong with that at all!  Buy stuff that you like the sound of and sod everything else ... because it's the only way to properly enjoy the end-result!
 
I quite enjoyed my 4490-issue Bifrost and it was one of those that I actually liked over the, 4399-based, predecessor.  So far, I'm liking the multi-bit version even more, though.
 
To answer your question, or try to, I'm afraid I don't know of any pure CotS based DACs (i.e. ones that use the up-sampling, DAC, and filtering all off the CotS chip/chipset), be they AK4490 or something else, currently, that play in the same space as the PS DS.  Yggdrasil does, and a bit above it sonically I think still, partly (or maybe wholly) by way of an very high-spec DAC chip (four of them, in fact) and completely custom filtering.  Chord's products are there too.  All three of those products pay very special attention to filtering.  And, per Mike's post above, handling filtering in ways not baked into the chip itself* seems to be able to pay big dividends.  Linn, for example, use the aging Wolfson 8741 ... and get much better results than I'd ever expect ... largely because they also rolled their own, FPGA-based, re-sampling and filtering code.
 
I only look at the design of DAC, or the chips they use, as a point of engineering interest.  Often, if I pull the thing apart, I just find a "reference" or "data sheet" design, with a gussied up PSU (not a bad thing per-se, but the levels it gets taken to can be a bit daft - particularly when much simpler implementations yield better measured and audible results) and some "high spec" components (often in places that shouldn't matter).  This sort of thing doesn't want to make me plonk down larges sums of money ... I'd rather spend it, right now, in places where there's more innovation and experimentation going on.
 
Which is a big part of why this whole process was initially biased towards discrete, clever, non-CotS/non-D/S based converters in the first place.
 
(*I'd have to pull data sheets to be sure, but from memory a lot of the CotS DAC chips have filter implementations that measure in the hundreds of taps, compared to the definite high-end trend towards more powerful filter implementations with tens of thousands, or more, taps, again Schiit, Chord and so on are pushing on this pretty heavily - and it seems to be paying off).

 

[Torq]

  An interesting CotS point: I recently, for the sake of completion, evolved a reference design DAC from a chipmaker I had never considered. When complete, I was shocked at how the sound differed from all published/3rd party impressions of this chipmaker’s family of DACs. Puzzled, I then obtained and read the CotS’ evaluation board schematics, only to find a suggested post DAC analog filter which was significantly different than the one I heuristically developed. Further reading proved that identical eval board filter was suggested/implemented across their line of DACs. From this, I speculate the converters using this chip IMHO could be made to sound/measure much more to my liking.

It would definitely be interesting to see where things wound up if more designers approached their implementations in that manner.
 
All too often the core of these things is just right off the reference design/data-sheet.  Which works, I suppose, for various not-very-interesting values of "works".
 
Makes me wonder if, at some point, we're going to see a separation of the implementation of DACs (referring to the entire box/product, not just the IC) start to separate out into two or more boxes.  At a minimum the raw converter, and then a separate filter box, so they could be swapped in and out.  Obviously there are limits there, unless the filter implementation is field-programmable/software driven, as the nature and tuning of the filter is going to vary based on the type of converter, and there are various other problems that come from breaking things up into different boxes (maybe just different boards which, I suppose, you already do!).
 
Anyway, just idle speculation ... 

 

[artur9]

  All too often the core of these things is just right off the reference design/data-sheet.  Which works, I suppose, for various not-very-interesting values of "works".

In the software world, this is the rule rather than the exception.  So much so that when I design APIs I no longer provide examples of their use because I know that others will just copy the example rather than code the thing up properly.  Tends to piss people off but I am beyond caring.

 

[artur9]

The AQUA Formula seems to be hitting the streets now.  Hope you manage to secure one for a demo.

 

[Torq]

  The AQUA Formula seems to be hitting the streets now.  Hope you manage to secure one for a demo.

Other than my being able to settle on a date, their distributor here has been very pro-active about this, so it should not be a problem at all!

 

[maxh22]

Hey Torq, when are you getting the DCS and MSB stuff? Have you auditioned any of it yet?

 

[Torq]

  Hey Torq, when are you getting the DCS and MSB stuff? Have you auditioned any of it yet?

 
No, sadly, I haven't auditioned either line yet.
 
Timing will be least a couple of weeks out at this point, probably closer to a month, and that's just on my side - will be down to demo availability once I'm free enough, schedule wise, to tackle them.

 

[ToroFiestaSol]

  In the case of the Adagio, for me, the significant additional expense of the unit with the super volume-control isn't warranted unless it is somehow improving the sound of the cheaper, but supposedly identical from a DAC perspective, Pavane.  If the Pavane wasn't available then the presence of said volume control on the Adagio wouldn't deter me - just for my use case it seems to be cost for it's own sake.
 
With the Chord stuff I think it comes down to people's long-term conditioning, especially with portable gear - where the general mantra has been "pure line out", which probably started when people began strapping external amplifiers to their iPods and wanted to bypass the built-in amp stage.  That sort of thinking stuck - even in cases where the implementation makes it moot.

 
Differences between Adagio and Pavane, from Metrum's Facebook:
 
"Since we have our new Adagio sometimes people do not understand the difference with our Pavane. From the outside it seems just a volume potmeter and a different type of remote but the inside tells more.
• As volume is controlled by changing the reference voltage of the dacs, max ref voltage is increased 3 times. This approach created a noise floor of - 155 dB ,extreme linearity and very low distortion!
• Instead of using our DAC ONE module as used in our Pavane, for the Adagio we used our DAC TWO. DAC TWO is using two R2R ladders on the inside which means that the Adagio is using 16 ladder dacs instead of 8 ladder dacs as used in the Pavane.
• The Adagio is running on double voltages compared to the Pavane.
• The Pavane is using 15 VA transformers for each channel, the Adagio is using 30VA for each channel.
• By using DAC TWO modules ( in fact every module contains a FPGA + two R2R ladders) there is no need to use the external FPGA anymore as used in the Pavane. This location on the digital input board can be used for future options
• As the Adagio is a digital preamp there is an option to control two power amps.
• As the Adagio can be connected to every type of power amp it has two possible output voltages
• And of course the Adagio will drive our future stereo and mono blocks, hopefully ready end of 2016"
 
Totally different DACs.
 
PS: Link
https://www.facebook.com/digitaltoanalog/?fref=nf

 

[Torq]

   
Differences between Adagio and Pavane, from Metrum's Facebook:
 
"Since we have our new Adagio sometimes people do not understand the difference with our Pavane. From the outside it seems just a volume potmeter and a different type of remote but the inside tells more.
• As volume is controlled by changing the reference voltage of the dacs, max ref voltage is increased 3 times. This approach created a noise floor of - 155 dB ,extreme linearity and very low distortion!
• Instead of using our DAC ONE module as used in our Pavane, for the Adagio we used our DAC TWO. DAC TWO is using two R2R ladders on the inside which means that the Adagio is using 16 ladder dacs instead of 8 ladder dacs as used in the Pavane.
• The Adagio is running on double voltages compared to the Pavane.
• The Pavane is using 15 VA transformers for each channel, the Adagio is using 30VA for each channel.
• By using DAC TWO modules ( in fact every module contains a FPGA + two R2R ladders) there is no need to use the external FPGA anymore as used in the Pavane. This location on the digital input board can be used for future options
• As the Adagio is a digital preamp there is an option to control two power amps.
• As the Adagio can be connected to every type of power amp it has two possible output voltages
• And of course the Adagio will drive our future stereo and mono blocks, hopefully ready end of 2016"
 
Totally different DACs.

It's a shame that that information was not apparent when I looked at the Adagio on Metrum's web-site!
 
I don't do Facebook, nor am I ever likely too, so had you not posted this I'd probably have remained unaware of the most important differences, so thank you for that!
 
Given the performance of Pavane, and the changes are this big, I'll add it to the list (assuming I can get an audition).

 

[ToroFiestaSol]

 
It's a shame that that information was not apparent when I looked at the Adagio on Metrum's web-site!
 
I don't do Facebook, nor am I ever likely too, so had you not posted this I'd probably have remained unaware of the most important differences, so thank you for that!
 
Given the performance of Pavane, and the changes in this bit, I'll add it to the list (assuming I can get an audition).

 
You're welcome
 
I wasn't aware of that until today too, they should include that info at the same time they lunch the product, not two months later.