Significant steps? (DAC)


Thought I'd get some opinions on where the significant audible steps are in DAC/DAC amp/stack are.   Sure, ESS sounds different from AKM, Op amps are different. Good discrete class A can be different. Tiny tiny differences.  Where are the big differences. Big for a Dac of course as there is little real difference from a Apple Dongle to a Pontius to insanity? 

Step one. The dongle. $12. Good for portable. Most would agree, not quite hi-fi but actually listenable on the move. 

Step Two. Real DACs made for hi-fi. $200 stack, DAC/amp. e.g Schiit or JDS stack, Topping DX3pro+, etc. These can be pretty good. Similar to respectable internals in integrated amps of respect.  Popular priced, high volume,  chip based.  They all sound about the same to me.  Maybe with $1000 headphones you can pick out differences. The better of them "do-no-wrong"  which is significant in my book. The Sabaj A20 and the Gustard A26 have the same chip set, both mass produced, op amp etc.  Why a 3X price difference? Just the streamer feature? Is this group, sub 1K maybe, where the "all DACs sound the same comes from?  Or like me, having heard an expensive but not very good one biased them into a steadfast spec chasers?  ( Bad experience, but my mind/ears are still open) 

Step three.  It gets harder. I would expect not just not bad, but something actually better.  Different moving to the cheapest R2R or higher end D/S.  Is the  difference from the DX3 to a D70 really significant? RME in here?  Better or just different to justify 2 to 5 times the price?  Every reviewer at this price claims every new one is worlds apart and the new budget champ. Really?  Higher price may be justified as more expensive chips, R2R are of course difficult, better power supplies, better output buffers. Without direct comparison, is this a level you can live with happily and just listen to the music? Are there still actual flaws in the performance? Do you need to have a system appropriately balanced components with speakers above 10K to hear the difference?  Can you hear it on an Audiolab integrated and set of Sonas Faber Illuminas? 

Step four. Expecting serious entry to high end.  Is it at the maybe Mearson or Gustard R26/A26 level? Qutest?  Or are these really small increments from less?   Is the real step having to go Holo or mid-line Denafrips?   2K or is it 3? 4? Where is that "real" step? Does dealer vs direct sale move things between steps? What would a Aris cost if it was through a dealer chain? Price difference and the need for a preamp would put it almost 4K.  Is that good?  

Higher?  After several grand, does it really cost more and justify higher prices, or are you buying the case and prestige?  I have probably not heard the right ones, as the upper end I have heard, the difference to my bottom end was very slight. Does a Birkman or Berkley really change things? What can they do in a box at DCS that is worth $140K that actually sounds better than one for 3 or 5K, or $200 for that matter? Maybe technically close to perfect, but how far above the studio recording and mastering makes any difference. GIGO. 

So, where are the dividing lines that separate significant performance differences?  What product is a bulwark for that tier? 

tvrgeek

Showing 17 responses by tvrgeek

There are significant steps, giant steps, and steps into fancy cases and ego bragging rights.  Giant may be more than I am willing to do. There is a point somewhere in the giant range where there is not actually any improvement to be made regardless of price. I use a PC as my server so no streaming, no MQA, etc. Mostly redbook FLAC files. 

I guess I am looking upper significant, lower giant. It would have to something better as I have tuned the system to the do-no-wrong. I could stop probably. 

Chord ( both Mojo and Qutest) Pontus or Holo Spring would be in the list. Quite a stretch for me.  Bummer Holo wants $600 to add a pot and buffer and only the top end has the improved USB port. Brings it up close to 4K.   That should be universal as Schiit can do it in their $230 unit.  An AMD galvanic transformer costs $12 retail, so feeble excuses.  There should  be no need for a DDC any more for USB.  Band-aid for a bad USB implementation. Once upon a time and for SPDIF,  but not now on USB. At the multi-K price point, even SPDIF re-clocking should be built in. 

I wonder if the R26 is any better than a Bifrost or just full of features I have no use for.  I imagine a sub 2K R2R would sound at least different. My focus is speakers only, vocals and acoustic instruments as natural as possible.  Price jumps quite a bit if one needs to add a preamp. Kara is about the entry point so now $1500. 

I have see way too many horror stories about the lack of support from Musician. It is a shame when someone puts out what may be a great product, but ignores the owners.  Or like Topping, support seems to be one gentleman who answerers questions on the forums.  I wonder about all these Chinese brands that don't even have a corporate WEB site.  A $200 DAC is throw away. $2000 is not.  Geselli, JDS, Schiit, Chord, and RME you can get on the phone with.  

No, I have not posted my system. It is resolving enough I can tell the differences between the Muse, 4 SMSLs, 2 toppings and my JDS.  The Sabaj and the DX3pro+ sounded the same to me. The difference between my own amp and the Vidar were quite clear. Which is better I am not sure but I am sticking with the Schiit for now.  Detail vs dynamics. I have not heard a pair of speakers under 5K I would prefer, and a lot way over it I sure would not. Come to think about it, I have not heard a speaker that I "gotta-have" and speakers are way more important than the DAC. TuneTot wasn't bad. They have made a lot of progress since the Watt. I would like to hear the mid-line Sonas Faber. I could probably match it myself but it takes about 6 months to do a speaker. 

I stick with my beer.  With beer, it is the recipe. Wine the right mouse has to piss on the right vine on the right day....   Great, but too hard to pick.  I'll stick with my Amber ales and porters. Great for $12/6-pack. 

I am going to get a loner Mojo. What I don't know is when the power adapter is got AC applied, does the dac power up, or do you have to hit the ball?  Same for a Qutest.   Dave is irrelevant for a retired Civil Servant on SS. 

Same question on the Pontus and Holo. Do they power up for use when AC is applied?

Everything is in a cabinet. When I power on the PC, via USB, it triggers a relay for my power sequencer and brings AC up for everything. I am not opening the door to fumble for a bunch of stupid keep-alive buttons. At least the Vidar goes into standby, but when sensing an input, it comes on. So does my sub. 

For this usability feature, I am being pushed to a Bifrost or E70 Velvet.  An outsider may be the Mearson. 

Of the Topping, JDS, Schiit and a Sabaj I tried, there were differences. Tiny and not very significant. Without A/B, I would say none. Only in direct swapping. I consider these to be well executed in their class. Agree, not all are. Several others went back to Amazon. 

Absolutely, execution matters. Chips not that much. But a billet aluminum case does not actually shield 60 Hz better than stamped steel. I can't hear a color display. 

So, is there a step up that is significant?   Are the improved circuits, code, the analog and reconstruction filters really better? Does Chord know something?  Does the R2R sidestep the worst of the filtering problems?  Are the interfaces well designed so as not to have compatibility issues?  Or is this all a joke playing off our egos.  Why does HOLO make you pay $1000 more for a decent USB. Schiit does it for $250 and it includes the DAC.  An Analog Devices transformer chip costs about a buck in quantity.   What's the excuse?

From my manufacturing experience, there is not much you can actually do above a few K. Which products take that money and actually do a better job?  As an example in an R2R, one can build a single ladder with SOTA .01% resistors, or select sets in parallel and get .001%. Better except thermal drift tosses it out the window. Better caps?  More processing power allows more complex filtering.  There are things you can't do for $100 in a DAC. Toss in useless interfaces, data rates exceeding anything available, licensing for MQA and you are up to $400. Add low volume production and a nice case it is  $800. Drop it into a dealer network and up to $1600  A few You-Tube reviews and audio shows, it is now a $2000 unit.  Same circuit.  

So what units are actually an audible step up. Focus on midrange smoothness. Then maybe dynamics. Don't want to go into adjectives that describe the source, speakers and room being applied to a DAC. Just not there.  Oh, the "midrange sounds forward"  Translation: small hump in the eq.  Would I actually hear a difference between a Pontus and Biftost? Three times the price, is it twice as good? Only slight?   A Qutest over my JDS? 16 times the price? How much better? How much more would I enjoy a few hundred more CDs over the electronics?  Goodwill for 69 cents each!  With discounts at B&N, around $7. I am willing to pay more if it matters. What step to what matters?

I wake up the PC, drag several covers to the queue and walk away. Playing music need not be like a Japanese tea ceremony so not searching for on buttons inside a cabinet. What it looks like I don't care as my electronics are for listening to, not on display as industrial art. Electronics are not part of a craftsman/Asian decor. 

Problem is getting them to listen to.  I can borrow a Qutest, but that is about it. 

Having a background in engineering and manufacturing, I know there is a limit where you exceed marginal improvements and transition into magic.  There is just nothing left to do. Sorry if you have one, but $140K for a DAC is not perfection, it is pure bragging rights. Hopefully as good as one for 1/10the price.  3K? 5K? Within the real engineering realm audible or not. Maybe even a little more. My biggest gripe is at that level, they have left out important audible features. No excuse.  $100 or even 1K depending on technology is cost constrained. Not 5K. 

I was hoping for some units I have not heard of to search out. Not everything makes it into You-Tube shills glowing reviews. 

Good point on separates. For my last round of upgrades, I considered going to a better integrated amp. Looked at SIM, Hegel, Atoll, Audiolab, Rega, Exposure,  etc.  I thought the internal DACs were no improvement to my JDS. ( took it with me to A/B a couple of them)  In isolation, only the Moon seemed to be not as good so  one could live happily with a ton of 50 to 100W A/B integrated amps. But I don't need all the extra switching, headphones, or as it seems, the internal DAC.  None had the level of analog tone control I wanted. and most had the stupid power on to stand-by issue. ( Andiolab and Exposure did not)  So I bought a Vidar and Lokius for less than half the price. I was almost going to buy some Sonas Fabers, but the  upgrade to mine pushed them to their level at least.  That's the only reason I can get past the wife the idea of another DAC. 

And very true, a big part of boutique products is sales in the hundreds ( they hope) does not amortize as well as in the hundreds of thousands and engineering spread over many products.  As an engineer I may understand that, but as a consumer with limited or maybe just rational means, what do I get for my buck? What do I get from the brick and mortar distribution?  No one is going to show up and install a $900  DAC like a set of Wilsons. 

So I have an approach.  It seems once above the mass produced bottom tier, DACs that garner some interest seem to be stand alone requiring a preamp.  More money. So I ordered an Alps servo pot remote board from China-Inc. I hope it is a real ALPS. I am going to stuff this into my Asgard along with a little mixer to give a buffered mono sub out. So a very good remote control preamp and phones for under $100 outlay.  Single ended but my cables are all very short. 

That puts into the cost range the Mojo, Bifrost, Ares entry level non D/S chip wonders. Better or not, probably different. It still leaves open the better of the all-in-one that may be worth a try like the SU-9pro. Returnable if I hear no difference.  

I just can't get excited enough to do the big jump to the Holo Spring full version, Pontus, Hugo range.  Maybe that is the real step up as most have suggested. I hear you. I just can't go there.  There does just not seem to be anything of real improvement value between sup 1K and close to 3. I look at something like the R26 and can't help but to think it is a $500 DAC with $800 in features I don't need. A good one, but no better. 

I think a point I made was miss-understood. At some cost, there is nothing else you can do from a strait engineering improvement.  Above that, there may be value to some for exclusive, visuals, etc. but there just isn't anything known to make it sound better. Hand built does not make better sound. Mono-block billet machining does not sound better. Actually a stamped steel case has better shielding. ( OK, who is advertising Mu-Metal shielding? You want high end, use better materials! ) If it is a value to you, fine.  There is also the diminishing returns way below that level where the home system electronics is far superior to the studio.  To put it bluntly, at some point you are transporting a load of manure in a Rolls Royce hoping it will smell better. My quest is only sonic performance.  A Bugatti won't get me to the store any faster than my GTI and considering real roads ( like the music source) the GTI is a better match.  It is a big step up from a Yogo but above it, there is not really much more you can do in real performance. 

Well Porsche USED to make useable sports cars. You could actually make it to work and back every day. They do make a re-badged VW SUV.  Drove a couple a few years ago. Bought an MG. Better car. More fun. 

I was going to compare the Lambo to a $140K DAC, but the DAC will actually sit on the shelf and do a perfectly good job while you brag how rich you are, where the Lambo is just bragging rights for rich guys  that if it does not catch on fire every time you rev it, won't clear a driveway apron let alone a speed bump and is darn right dangerous in even a light rain.  So it is not really a car and no comparison is valid. Not even a decent technology demonstrator.   But as I said, there is a point when more makes no useable difference and in case of super cars, more is actually less. 

Oh for the days of the Ferrari 400 or  the 330   Useable cars on real streets. Chord Dave?   Then the did the F40 and they have been useless ever since. The last useful Lambo was a farm tractor. Come to think of it Porsche made a series of very useful tractors too. Bill Scott had a complete collection. 

 

As the consensus is the big step up is more than I am willing to pay, and my entry level chip based do not have an obvious defect, I decided to see what a higher end current chip based does within a "play" budget.    So Amazon, sale price SMSL D400es.   $800 list, sale $550

Well, ONE parameter that improved when going from entry level  (DX3pro+ I am using or the Sabaj that I sent back with the SMSL  is the remote works across the room!  Usability counts.  Thumb's down on the front spinner. If I keep it, it will get a knob glued to it. 

 It will take a lot of music to see if I can hear a difference. Main system speakers, desk system speakers, Desk headphones.  Amanda McBroom sounds great, but she always sounds great!  

One of the tests will be to see if this unit requires the decrease in player gain, or if by stacking the DACs they have sufficient bits to capture overshoot in the filter calculations and then scale before the conversion.  Going to have to dig into the architecture deeper.  For sure the Schiit and JDS did. It would be great if this was another problem recognized and fixed.  

Trying to think how to test with and without a galvanic isolator without opening the case.  Probably not possible. 

.Jumping to Spring or Pontus is more than I can do. I decided against the Chord at least for now. 

 

Yes. 

I guess some people do one thing then they die. Not me.   I started my career with Storage Technology and ended retiring from the DoD. 

 

I agree, the most important part of the DAC is the reconstruction and output analog, though from my listening I would have to give a lot of props to the digital implementation ( not the chip) as the Atom+ still holds it's own with a 3-generation old chip and generic OP-Amp output.  John seems to know what he is doing.  The same has been said about Geselli but I have not heard one.   One of the reasons I decided to try the D300es was their higher end output from their entry level. ( multi parallel OPA OP-Amps) and claimed better power supplies.  I also have a Schiit Asgard that has a very good discrete output.  Granted, Benchmark is , well the Benchmark.  If I had the bucks, ( $2200?) a Benchmark  might have been on the table.  Their literature suggests they have addressed all the issues I know about. At least with Benchmark, you are paying for the circuit, not the advertising. I was looking at an AHB2, but too much money.  I think of Bryston kind of the same way but don;t know if it is still true. 

The differences are slight, but maybe the SMSL does have a little sharper details compared to the JDS, but I am not sure it is as pleasant. The older AKM Schiit DAC was edgier until I reduced the JRiver output. They all got closer then. 

Unfortunately, we are dealing with new and re-mastered tracks post "loudness wars" and we are lucky if that 24 bit download is more than 12 bits.   Now as far as internal DSP, having many more bits to deal with for overhead allows more filter choices and then the ability to adjust the level without adding additional clipping. Another advantage to newer chips or stacking them. 

I agree an integrated DAC and Pre, or for that matter full integrated amp is a better engineering solution as it eliminates any cabling issues or other engineering tradeoffs by not knowing the following stage. It can reduce money spent on things like fancy cabinets and sharing things like the digital power supply separate from analog supplies.   After looking at several integrated amps, I chose to remain separates and picked up the Vidar. Of the very few that actually power up with AC supplied, I was not impressed with the DAC in any I could afford.  I am putting a servo remote in my Asgard, the Topping and SMSL both have preamp modes so the integrated preamp section is redundant. 

One trap I think we get into is "hearing new things" not because they are there, but because we are listening for them.  How many times has someone pointed out something odd in a track you ever noticed, but once noticed, you don't un-hear it even when you chill out and are just listening to music.  

So , one day only I am not impressed with the SMSL.  I may want a more artificially smoother DAC like an R2R.   Unfortunately R2R is an expensive technology to implement, so the "Equivalent" quality level is likely several times a chip DAC.  Spring and Pontius seem to be the suggested actual improvement.   Neither is available for me to borrow.  A Bifrost is about the only option. 

Alas, I am forced with my living room as it is as it is our living room, not a sound demonstration room.  Not bad, but a little bright.  I have tracked down all the little things that can buzz.  

I repetitive comment is criticizing  non-mega-buck systems.  None of those folks know mine.   At the listening levels I use, mine speakers are very resolving. Granted, in a bigger room and played much louder, there are better choices.  I don;t need a 1000W super amp as I never use mote than about 10. 

Listening on my desk system is concluded.  Headphones and speakers. There might be a slight difference in upper treble detail, but insignificant to me. In about the same range as switching filters from minimum phase to Anodizing. I hear no differences from probably 6K down.   Not looking good for the D400es as an upgrade. 

More on the main system were first impressions were stronger. Detail, but at the expense of smoothness. Fine is a cymbals crash is cleaner but not if a vocal is edgier. 

 

I think the conclusions that all "tier one" chip DACs sound almost the same are pretty close.  I do hear a tiny difference it the very top end "air" but it seems to be at the expense of vocal smoothness. Filter 3 smoother, 4 sharper .  Overall, I prefer my old JDS Atom+ for just enjoying music. It could be the differences are above the 17K or so range where my hearing is falling off. Or the "massive differences" from the You-Tube shills are all made up. Is it a shame that good hearing is wasted on the young who can't afford top tier equipment? 

The front knob is USELESS. One could glue a knob on it. But a desktop device you have to use the remote for is just plain stupid. On the plus side, Remote range is far better than Topping, it has XLR outputs, and does not really do anything wrong, so in a vacuum, most would be quite happy with it.  I think they did spend the money on the power and output, but it does not seem to be the determining factor. 

So the SMSL is going back.  Keeping the DX3pro+ in the main stereo for now and the JDS Atom+ stack on the desk.  I was considering a Denafrips/Ladder/Musician entry level R2R, but the horror stories about defective units or returning because they did not like them is preventing me. Not risking that kind of bucks for no support.  Jumping up to mid tier is even riskier! There may be great positive stories but I did not find any. As we know, do it right and you tell a couple people, do it wrong and you tell the world. That leaves the Bifrost or Mojo as the only two sub 1K  DACs that may sound different with minimal risk. I think I'll wait a while. 

I was thinking about the above article I referenced.  Do consider the test signal that stepped from 0 to max in one bit is an artificial situation that would never happen unless a corrupted file and it exceeds the rise time of a sine wave within the 20K spectrum.  It implies a single sample wide pulse.  So his comments on looking at the ripple seems reasonable.  Not sure I agree with his filter selection, but of course that is a "hearing" choice.   Looking at the response of the Schiit and JDS, it looks like both are close to "linear fast" but as these are older chips, the filters are custom and not the hardcoded  ESS numbered ones.  I also noticed the documentation in Topping and SMSL do not match the numbers as published in the ESS data sheet. They scramble the numbers a little.  You can figure it out looking at their published documentation.   

This is a complicated subject and I would like to track down more just for understanding. Bits are bits, but what you do with them matters.   Someone asked Mike Moffit which filter he used in the Schiit DACs. His answer was "The right one".  I suspect the filter is a large pert of why John Seaber's DACs are so good using older cheap parts.  It's the execution, not the parts. 

Well then Jacob,  I an sure everyone here appreciates your in-depth analysis and contributions with great insight.