NOS1 USB vs Berkeley Alpha + Lynx AES16

[praphan]

Hello everyone,

I have promised somewhere in this forum that I am going to spend some time conducting critical listening to my both NOS1 USB and Berkeley Alpha DAC and share my experience.  I have my NOS1 USB set up and sound very stable now and it is a good time to report the comparison.

Most of us should have heard more or less about the Alpha DAC which was designed by three Pacific Microsonics alumni – Micheal Pflaumer , Michael Ritter (both were co-founder of Pacific Microsonics and co-inventor of HDCD process) and Rene Jaeger. All of them involved in the development of the legendary fine Model One and Two.
 
I have no clue if Alpha would reproduce music the same way as Model Two which is also owned by Mani. But I found that Alpha is highly resolving and musical at all signal levels to submerge you into the sound of music. Alpha’s resolution to information is of high quality that the music does not tend to be too analytical with too much detail making the listening experience a fatigue one. It is rather persuasive, refined and silky smooth. The music in front of you becomes palpably real. All digital information is processed to deliver well balanced, natural and in an unforced manner accentuating a sense of ease, relaxation and musical involvement. Alpha is also capable of handling a fair amount of dynamic contrast with quick transient. All in all, I really like my Alpha played back from XXHE through my music server driving Lynx AES 16 sound card.
 
To compare Alpha with NOS1 USB, all my gears used for comparison are exactly identical ranging from PC, power cables, interconnects, amps and all the electronics in both signal chain. Although my electronic gears are not true reference quality with half a million dollar price tags, but they are good enough for analytical and critical listening. In fact, I wish to eliminate the use of Lynx AES 16 but my SPDIF output on my NOS1 USB is not active yet. I still have to wait for another new SPDIF driver to be coded by Peter for this comparison. Therefore at this stage, the comparison will be between NOS1 USB vs Alpha + Lynx. Five albums were chosen for my critical listening ie The Chieftains – tears of stone ; Gershwin by Grofe ; Friends of Carlotta (direct to disc); Diana Krall - Live in Paris and We Get Requests by The Oscar Peterson Trio.

To me the first noticeable difference between the two DACs is the sense of space and the naturalness of the staging. NOS1 USB has a more forward perspective along with greater sound stage depth. I can imagine myself sitting in an expensive front seat of a concert hall playing back through NOS1 USB. When you buy a more expensive ticket to enjoy live performance you will be able to perceive a better sense of depth of the band or orchestra. The instruments in the back of the hall sounded more distant while the instruments near you will be very presence but can be sometime fatigue with extended listening. This is particularly true for NOS1 USB playing back “You look good to me” by the Oscar Peterson Trio. NOS1 USB produced a much greater sense of ride cymbal and hi-hat’s presence before you than the Alpha + Lynx. I have to lower the volume by one notch to avoid tendency of fatigue. Having said that I do not feel too much overwhelmingly tired at all listening to this album from NOS1 USB. But I like it played back from Alpha more than from NOS1 USB. Please do not get me wrong, cymbal sound produced by NOS1 USB is of resolved with transparent texture without graininess and edgyness. But it is a bit bright for me. On the contrary, when playing back Friends of Carlotta, NOS1 USB revealed much better sense of three vocal singers in different positions in the recording studio. The image and sense of position are stunning in which Alpha is not much resolving in this aspect. Through Alpha + Lynx, I am not quite sure if there are 3 or 4 vocals. The NOS1 USB revealed three singers right away and you know where they are in relation to your listening position. Playing back Gershwin through NOS1 USB reveals a much better sense of perspective that makes listening the same track through Alpha relatively flat like you buy a mid price ticket to a live performance. I went to Celine Dion live at Caesar Palace in Vegas some years back and did not book the ticket well in advance. My seat is skew on the side and at least two third of the depth of hall from the stage. Although the sound quality is world class with excellent dynamic, tranparency and sense of fresh and real, what I heard is so flat due to the distance from the stage. For me, I like to sit closer to the stage and pay premium. This is what you will get from NOS1 USB.

The other distinct aspect of NOS1 USB is its capability of delivering greater dynamic contracts, along with lightning-fast portrayal of transient information. The music swings effortlessly from quite to full steam with tremendous speed, but with no sense of etch on the transient leading edges. Playing Oscar Peterson and Diana Krall through NOS1 USB is tantamount to listening to their first live performance in their world tour. They are still full of energy and enthusiasm to portray their skill in music as much as they can. On the contrary, playing back the same tracks through Alpha + Lynx gives you an impression that they are a bit tired towards the end of their world tour. Diana Krall’s fingers rolling on her Steinway grand piano are not so punchy and powerful as she is still energetic. In sum, ability to resolve dynamic contrasts and transient is clearly to be more superior by NOS1 USB.

The last clear virtue of NOS1 USB over Alpha is the sense of transparency as a result of ability to reveal more details of musical information. Remember I played 24/768 through XX z-6 (see my signature). The presentation had a pristine clarity and vividness that seemed to strip away all the filters/windows placed between me and the musical notes. The results are so stunning that I could hear vividly a lively plug of Ray Brown’s double bass (Oscar Peterson Trio) which is not as resolving from Alpha + Lynx. While Alpha deliver timbres which are simultaneously immediate and palpable, yet gentle and relaxed, NOS1 USB wraps musical vividness with ease and grace making me so involved in music I hear.
 
On other aspects of the sound quality such as tonal balance, pace, rhythm, timing, musicality, both DACs are performing very well and have a little difference to compare. They do not sound exactly the same anyway but both sound excellent on these aspects.

Alpha Audio Design just recently released their USB interface which was reviewed and praised highly by Robert Harley. According to Robert, Alpha USB brings Alpha DAC to the next bar in all aspects of sound reproduction. This implies that Lynx card is no more a perfect match for Alpha DAC. It took Michael Pflaumer two years to design Alpha USB interface which is considered by Robert to be state-of-the-art. Does anyone here on this forum own Alpha DAC and its USB interface? If so please share your experience.

I definitely enjoy NOS1 USB over Alpha plus Lynx. NOS1 USB is now a permanent out board DAC in my listening room. Alpha is moved to other room. I am not quite sure if I should spend USD 1700 on Alpha USB. Would appreciate if anyone can share.

Praphan

[PeterSt]

Praphan - Time for me to say WOW.

WOW not only for such an offically looking review, but merely for your honesty. And you know what ? it can only help. You, and the others. And me btw.

So, all summarized in one little phrase : I 200% agree with all you said. I know the albums too, and indeed this is what happens.

But our fun is : You NOS1-USB does nothing ...
Ok, it transfers digital to analogue, but will show THD figures which are real, IOW really representative for what it is fed with.

And so we must be at the outside of it ...

Ok, what would you say if I add some 20K of value to it ? Still not that half a million of your undervalued poor system (I wonder haha), but at some stage you will get there. Just from the music it will produce ... Just watch it ...

I will send you an email about something to try out. It may work, it may not. For me it works because I was disturbed about the exact same one thing which disturbed you. Yea, blame that recording. But I have been skipping too many of them lately.
If you go through the roof (in the positive sense) because of it, just let it know in here, but don't tell what it is you applied. When you don't go through the roof, also let it know as honest as you were and are. But again, don't tell what it is you applied.

Thank you for such a great review.
Peter

[manisandher]

Hey Praphan, thanks for the write up - a really great read.

I've been banging on for longer than I can remember that the performance of the Model Two is ultimately determined by the interface used to connect it to the computer. I'm not 100% happy with either my Weiss AFI1 (even after upgrading its power supply to a linear Paul Hynes one) or my RME AES-32. But I reckon I might be able to improve the performance of the RME quite significantly if I adopted the 'PCI extender' method that Peter employed in the original NOS1. Perhaps you should give something like this a go with your Lynx before even thinking of buying the Berkeley USB interface? It might save you some money, especially considering someone we all know should have quite a few of these PCI extenders hanging around by now ;-)

Mani.

[praphan]

So, all summarized in one little phrase : I 200% agree with all you said. I know the albums too, and indeed this is what happens.

But our fun is : You NOS1-USB does nothing ...
Ok, it transfers digital to analogue, but will show THD figures which are real, IOW really representative for what it is fed with.

And so we must be at the outside of it ...

Hi Peter,

I am glad to be able to share some insight from my humble listening experience. Glad that you value my input too. My skill is still far away from many experts in this forum.

I got your email and will do some experiment and report back. I fully understand why I should keep low profile on what you suggest me to apply.

best,
Praphan

[earflappin]

Praphan, thanks also for taking the time to share your experience with the NOS1 and BADA.  I owned the BADA for about 18 months and did extensive experimentations with the AES interface. 

I ran an AES16 Lynx card from a fanless PC.  Bottom line, I was able to notably improve the BADA's performance by using either a Grimm or Antelope reclocker and a Legato USB/SPIF interface.  I think I was experiencing two different improvements.  First, a better clock and second, galvanic isolation. 

The AES/SPDIF interfaces are fundamentally flawed IMHO, but the Berkeley brain trust continue to espouse the position that USB/Firewire hardware do not belong inside the DAC and prefer instead to optimize the AES/SPDIF interface.  Despite the improvements I was able to achieve I still felt the BADA was adding a HF "zing" on all low rez material I fed it no matter what filter setting I used.  Nonetheless, we are talking fine strokes here....it is certainly a high performance DAC/Pre.

So I have little doubt their new USB/AES-SPDIF box would yield improvements over your current set-up.  BTW, they have also come out with a BADA2 which uses the same improved clocks they use in their new USB/AES-SPDIF box. 

Personally, I look forward to hearing just how much this new technology improves the BADA.  I have much respect for the Berkeley guys. 

Thanks again.

[juanpmar]

Hi praphan,

Great review even for the ones who don´t own a Berkeley dac but are in waiting list for the NOS1 USB.
This review gives us a very good idea about the NOS1 USB qualities.

Thank you,
Juan

[praphan]

Hey Praphan, thanks for the write up - a really great read.

Thanks for the kind words, Mani

Perhaps you should give something like this a go with your Lynx before even thinking of buying the Berkeley USB interface? It might save you some money, especially considering someone we all know should have quite a few of these PCI extenders hanging around by now ;-)

I still have 1095 PCIe extension card lying around on my test bench. However, my AES16 is a PCI version. I shield my Lynx card with EMI/RFI shielding paper to reduce the noise. The system has a very dark background. So I think Lynx need external powerful linear power supply to improve its sound quality.

thanks and regards,
Praphan

[praphan]

Great review even for the ones who don´t own a Berkeley dac but are in waiting list for the NOS1 USB.
This review gives us a very good idea about the NOS1 USB qualities.

Hello Juan,

Thanks. I am sure you will enjoy the upgrade which brings your NOS1 to the next bar.

Regards,
Praphan

[Bigear]

Hi Praphan,

nice review.

I also own a BADA with Lynx interface (as it was recommended by the folks from Berkeley) and even the non-USB NOS1 is already superior imho in the overall definition, especially in the low and high's and the dynamics. I had the opportunity to compare it at Peter's home when I purchased the NOS1 and the difference was pretty obvious. 

[praphan]

The instruments in the back of the hall sounded more distant while the instruments near you will be very presence but can be sometime fatigue with extended listening. This is particularly true for NOS1 USB playing back “You Look Good to Me” by the Oscar Peterson Trio. NOS1 USB produced a much greater sense of ride cymbal and hi-hat’s presence before you than the Alpha + Lynx. I have to lower the volume by one notch to avoid tendency of fatigue.

I have listened to this remastered track (You Look Good to Me- Oscar Peterson Trio) over and over and can only say that Winston Ma of FIM must have applied certain unique remastering techniques to improve dynamic range aiming at recreating excitement of this fine jazz music. I have an impression that Winston was trying to replicate "analog sound" to present transparency, detailed timbre and transience. During the last part of this track, the reproduction of cymbal sound can turn into distortion if the playback system is not resolving enough. I have high respect on Winston Ma and his FIM label and do not think that "You Look Good to Me" can be his remastering flaw. One needs a comparatively more resolving system to get this remastered track played back right. So it became my reference track. I feel that if I get this track right and move me, then I have the right system.

So, all summarized in one little phrase : I 200% agree with all you said. I know the albums too, and indeed this is what happens.
But our fun is : You NOS1-USB does nothing ...
Ok, it transfers digital to analogue, but will show THD figures which are real, IOW really representative for what it is fed with.
And so we must be at the outside of it ...
Ok, what would you say if I add some 20K of value to it ? Still not that half a million of your undervalued poor system (I wonder haha), but at some stage you will get there. Just from the music it will produce ... Just watch it ...
I will send you an email about something to try out. It may work, it may not. For me it works because I was disturbed about the exact same one thing which disturbed you. Yea, blame that recording. But I have been skipping too many of them lately.

If you go through the roof (in the positive sense) because of it, just let it know in here, but don't tell what it is you applied. When you don't go through the roof, also let it know as honest as you were and are. But again, don't tell what it is you applied.

I have applied what Peter suggested on a PM.  After a serious listening, I can reaffirm that the application works perfectly fine for me. It should work for other as well but only Peter can make the official announcement.  As promised, I am not in the position to reveal what I did. So stay tuned.

The only hint that I can give is that it has nothing to do with NOS1 USB. After the application to fix the “high” frequency syndrome that disturbed me on “You Look Good to Me”, I can feel the music with less analytical presentation.  Music  in front of me comes with more body with round texture and liquidity.  There is some trade off.  The image and staging are slightly diminished.  However, this can be subjective and only be my personal view.  To take sound stage out of the equation, I and Peter coincidentally conducted the same listening technique – “around the corner listening”. I stepped out of the listening room,  left the door wide open and stood by the door around the corner as if I was about to approach the entrance but stop right there. So what you will hear from that position is the music around the corner where sound stage, image and other dual effects are out of the picture. What counted are tonal balance, rhythm, dynamic range, transience and sound texture. I wanted to hear how interesting and involving the music sounded. Did my body move and gradually pull myself into the room and swing with the music ? Did I feel like the live performance is happening in the room ?

The result is a rewarding one.  I have a very beautiful playback for this unique track of Winston Ma and I am very happy with it. As I said, if I get this track right, all the rest in my HDDs should be fine.

I can say that I am a happy man now.

Kind regards,
Praphan

[easternlethal]

Peter - CA just reviewed Alpha's USB converter here:

http://www.computeraudiophile.com/content/Berkeley-Audio-Design-Alpha-USB-Review
(not very well, might I add)

Again, they say the USB controller should not be inside the DAC because the conversion to I2S affects the DAC circuitry (although I do not understand why converting from SPDIF to I2S would not run similar risks).

so i have to ask - do you agree, and does the NOS dac 'prefer' a particular connection?

[PeterSt]

Yea, interesting review. I mean "This is the best converter - period." ... what to do with that while there is no single expression about how it sounds.

I know, I express "sound" from audio maybe a little bit different from others, but to say nothing at all - No.

About the converter itself - I don't understand much of the setup / strategy at all. Well, that it looks like commercial talk, not knowing how to do it better;

This latter by itself is a bit nonsense, but let's say this is about money, and that the $1895 or whatever works better than a new Alpha which is $1895 higher priced. Furthermore, "of course" everybody wants to have asynch USB these days, so Berkely may expect to sell quite a few of those converters.
What about a rip off ?

Maybe I'm not talking blahbah all the way, if you only know that the original NOS1 was not supplied with (asynch) USB, and that people could have it upgraded by the huge amount of 335 euros (say $400). And this included the disassembly of the old stuff, and the assembly of the new. Ok, maybe that's way underpriced because of the work involved, but do notice that it includes the hardware as well, the licenses involved and some more stuff.

When we are at it anyway ...
The Alpha is the most dull sounding DAC I had in the house, so far. I didn't say "wrong sounding", but just completely dull. Filtered to death. Now, what I would have done - undoubtedly knowing the difference with e.g. a Weiss, was expressing what changed to that with the USB converter in front of it. But no. It is just the best converter - period.


Thanks,
Peter

[easternlethal]

Peter - would it be possible to describe a little more about the Nos DAC's approach with respect to ultimately getting signals nto the DAC's circuitry (which I believe requires ultimate conversion to I2S)? Do you recommend I2S direct via PCIe to the DAC (as I think it was originally supplied) or is the USB conversion within the DAC just as good now (with or without external USB SPDIF converter)? Is there any isolation involved? Sorry I have not been able to find any thread discussing this much. I read somewhere that usb/firewire controllers used to be added as an afterthought (Weiss actually subcontracts its firewire driver programming) so I was wondering how much attention was paid into the process for the NOS DAC..

[PeterSt]

Ok, let me type a few things ... 
(shouldn't you be sleepy or something ?)

[PeterSt]

The original NOS1 took the approach of having the clocks (oscillators) inside of the DAC while no outboard (in-PC) clocks were operative. It took the "soundcard" into the DAC while still being connected to - and being under the control of the PC. From there it was i2s to the DAC. No conversions in between (so, no SPIF in order or anything).

USB could have been done right from the start, but I was reluctant to go that route because USB isn't all that Walhalla people might think. About beautiful USB cables which improve sound and such ... which should not be (in my view).

The original solution appeared to be a tedious one, and we spent more time in getting everything going (per DAC !) than we could concentrate on putting DACs out, and moreover the solution appeared not to be compatible with all MoBos (all GigaByte) and in the end no solutions came - whatever we tried. I'd say, look at Praphans efforts (and his piles of MoBos on the floor) to see what I mean. It was no guaranteed solution.

And so USB came in afterall, but with the aim of being able to do DSD (this is what nobody knows ... that it was actually *this* which was the real motivation).

Having the hardware and driver software (around a year back now) immediately showed I was right; It didn't sound at all, was noisy as hell, but, nothing anywone would worry about. One little problem, it underperformed the original NOS1 which was the reference of course. Not only for me, but also for customers. So, with again Praphan as the example, how could I send him a now by guarantee working solution, while SQ was clearly less.

Along the lines there was heavy interaction with Nick in here, and we both worked on better noise resistant solutions; him doing that in the midst of the original NOS1 - me doing the same within the USB solution. And, although we both worked on actually different things, the aim was the same and what actually happened is that we motivated eachother to do it beter again.

Along the lines I had battery packs for everything, so the whole of the NOS1 ran on batteries (for 48 hours or so in one go). All was worked out for the recharging - a nasty thing for "users" anyway. From there I learned what inherent noise could be (on the low side) and I tweaked as long as it took to mimic that with the nomal PSUs. In the end that worked.

And so now there is the asynch USB solution which is resistant to any outside noise, while not having isolators or anything that would imply more jitter. This latter is one of the most important things. In the end the combination is.

Inside it goes directly from USB data to i2s, with the DAC right besides the USB connection. Nothing special here, but the low noise is, and the jitter specs are.

There's actually nothing much in the NOS1, like conversions or any other ciruitry that may debet the sound. Just the D/A chips (8 of them) which directly connect to i2s. Ok, there's an FPGA in the middle which controls the sample rate settings, and together with some FPGA on the USB board it all brings upgradeable firmware. So, the firmware (which comes along with the driver without notice) takes care of the leanest hardware operation - again thinking noise. Think of this like all sorts of connections being possible, but those not used are shut off in the firmware. Or shut on, when needed.

It is in this area where 768 emerged, which actually is a coorporation between playback software and the firmware in the DAC. May I find out how to let 1536 go today, tomorrow it will be operative at the customer's side. But okay, I won't, because the chips run at their limits.


Alright, this was maybe more history than answering the question directly, but there isn't much more to say I think. So now the story has some body. Prrrt.

Peter

[easternlethal]

ah i see. a few more questions if that's okay (please don't feel like you need to disclose anything you don't need to):

1) what usb chip does the dac use and is the driver programmed from scratch or based on some conventional standard like the vhpd1394? how do you avoid the noise in the usb signal from the beginning?

2) how do you convert the usb to i2s?

[PeterSt]

Let me try to be as vague as possible ...

There's a CS receiver and a Xilinx FPGA.

Driver was not programmed from scratch (who the h*ll do you think I am ? ), BUT, the source code was licensed to me. So, quite a few things are now proprietary to Phasure.

The noise is "just" a matter of wanting to avoid it, and next tweak as long as it takes to achieve that. Of course this is a way too simple explanation (as a matter of fact, it isn't at all), but what's the most important is that one (me) needs a reference to hunt down. To this regard I can tell you that in attempt to coorporate with other engineers in this (USB) field, I found no luck because "I shouldn't be doing difficult because any noise under XYZ would not be audible anyway".
That this relates to jitter specs is just that tad of knowledge where things can get better than in other random situations. Like I said in the before post : it is the combination of things which make it work.

Couldn't be more vague !
Peter

[TonyB]

Not true, you could be more vague 

It is interesting that you could bring noise down without isolation. If both the computer and DAC ground are earthed, there is a gound loop and some garbage signal will flow on this wire.

TonyB

[PeterSt]

While true, groundloops can be broken ...
Hint : this is all part of the DC Offset Control in the NOS1 ...
(which is without any circuitry for/against it)