Phasure NOS1 vs. Pacific Microsonics Model Two

[PeterSt]

Hi Mani - That's some nice sounding post ! Of course I knew half of it already, but that was only half. Well, thank you - and I feel honoured.

 

Maybe it's nice to say outloud that at the time you first heard the NOS1 (well over a year back now), you didn't own a PM Model Two yet; you bought that later and from that moment on it has been quite exciting to wait for what would happen when you'd be able to compare the both. I guess you must have been more confident than I myself.

 

I will make some pictures allright and add some graphs too.

 

Sasaki - Yes, standard ESI driver. And sorry I didn't respond to your email yet (same apologies to others).

 

Thanks,

Peter

 

[PeterSt]

I suppose trying this require me to get a PC desktop machine or can one use a mini PCI-E port in a laptop

 

No, I'm afraid that won't go.

But, PC-Express (card) slot : yes. But, only "officially" for now and it has never been tested yet.

 

On another note : laptops really don't cut it. They are really too slow (never mind the type and processor) and won't give you a means of convenient playback. Ok, the playback itself will be allright, but not the preprocessing needed. Please notice that I am talking XXHighEnd here, and think about a FLAC needing to be pre-converted which really can go well within half of a second for a 10 minute track. A laptop ? could be 30 seconds.

But it's easy to try out; it's not DAC related.

 

Regards,

Peter

 

[PeterSt]

Ted, Hi,

 

Of course you are correct, but it is not really a friendly means for playback. It's a bit similar to copying files to special media and play them from there (like SSD, RAMDisk) because it sounds better; who really likes doing that ?

 

So, a player like what's suggested best here, does that all for you and it goes "on the fly". This suggests real time and it will come so to you indeed once the PC is fast enough. I mean, something like half of a second is sufficiently that (IMO). Just play like always, but in the mean time it plays fully from memory and as in the Windows case, just PCM (WAV).

It doesn't matter what (sequence of) conversions were needed, and it's just done for you, including ending up in that media you asked for (set).

 

Of course, all this can be done in advance by yourself too, and the cpu etc. burden for it is not there. But you will be busy copying etc. those files more than listening to music.

 

About things like "up conversion" in general ... this too is not to be underestimated. I mean, my cons from above can be eliminated largely to convert your whole collection in advance, but :

 

a. It will be a kind of fixed (while still many variations exist you may want to try out tomorrow);

b. It still needs the original to be kept on disk;

c. Look at the size of a 32/352.8 album; 2-3GB !

 

I hope this enlightens a bit, but of course not to forget : is takes for starting point these things can go all automatically. If they don't anyway (or let's say when a player plays FLAC and converts real time during playback itself) - yes, your general option will be "do it yourself in advance". But suppose we're only talking about the ever hot subject of FLAC ... who really does this ?

 

All 'n all we could say that maybe the way I approach this all is "taking distance" from things we shouldn't be bothered with. And now this includes the filtering (which is the underlaying subject really) which is not fixed in-DAC, but which is flexible outside it. It really shouldn't imply you suddenly must start copying things. So, it's a concept and I hope to have made it consistent. But now it does need a PC which is sufficiently fast ...

 

Thanks Ted,

Peter

 

[PeterSt]

”My interest is in going right from the DAC to amp, so the output section matters a lot too.”

 

and

 

I’ll let Peter chime in here, as I don’t actually know much about the output section. But I know that Peter has specifically designed the NOS1 to drive power amps directly. Many of us do this and enjoy improved SQ as a result.

 

Since I'm asked I guess I have to say something here, but what can be added (apart from the output impedance being 33 Ohms, which Mani told in an earlier post) ?

But I guess I can tell a little story maybe ...

 

At knowing how much SQ will improve at skipping any means of "pre-amp" (but replace "preamp" by "any means in the signal path, up to a single $$ resistor"), it has been my explicit objective right from the start to let it work (regardless of amps and interlink lengths (~~)) "on its own" so to say;

If we skip some years of thinking about everything, and now knowing that the whole physical traject took some two years, I can tell you that it took around two years to do just this very subject well. It is related to nearly everything.

 

People who followed the thread over at Phasure about the first physical steps and how all emerged to a ready product at the end, will have seen that during the time of waiting for new PCBs, parts, sorting out other things, there has been one large big red thread all over : the "gain stage" as I'd like to call it for a general means of getting enough juice into the power amps. And as a side subject (though major) there has always been the "nothing in the signal path" subject, which is a sheer impossible thing. And for those experimenting with the PCM1704 chips ... they are a pain to this regard.

 

While wanting the direct connection (and which really is the explicit advice right now), there's always been de desire for having in some means of volume control. The theoretical possibility for this depends largely on the means of "gain" itself, and each means could you make sit back for a week and think about the now possibilities for that. It never worked, and without exception always degraded sound. Not that new customers would notice that, but that would only be because of not having the reference.

Even when the first DACs were in production, I was still trying out some newly found means with some sort of transformers (but luckily that did not work either, haha).

 

Btw, the fact that right from the start there was a very well working digital volume per means of XXHighEnd, never scared me much for the possible impossibility of not succeeding on the volume matter. This does, however, not mean that any potential customer will like it that way (the breaking windows syndrome), but in the end I choose that "commercial risk" over degration of SQ.

 

In a nutshell this little story -very indirectly- tells about the direct connection to power amps, but what I actually wanted to say is that it was this which took 2 years. Or let's say it really needed that throughput time.

Right now there's a fairly nice 1.5VRMS output for SE-RCA and 2.5VRMS for balanced-XLR which is on the limits of the 2.5V rails (5V plus/minus), allowing for a nice headroom for overshooting as well. No switches anywhere, and the only degrading element in the signal path (serial) is one (high quality) resistor.

 

Of course the implied question (quoted above) assumes a better SQ from a direct connection to begin with. So yes, this is good thinking. But you don't want to know how difficult is to do that right, and to not cause more harm than good. This morning I just wrote a little story on my own forum about me by now having more bass output than an amplified live band can do well with dedicated amps and speakers for the bass player. And the most interesting is : it really needs the software to complete that little part of it. Bits = still bits, but you don't want to know ...

(meaning something like : without the drive of a pre-amp or explicit "buffer" this will be the first thing going wrong; in the very end though, it will be the wobbling bass frequencies which may produce low perceived frequencies allright, but nothing like a wall of bass; low jitter combined with software which "controls", really does that job. Well, as it turned out.)

 

Peter

 

[PeterSt]

I'm curious which ESI card and driver supports this sample rate? I checked and the only current pro cards that goe above 96 seem to be the Juli@ and the Maxio.

 

It is the Juli@ allright.

We can use Miska's guess on how to achieve the sample rate as a good one; It needs programming on the DAC side (that FPGA = Field Programmable Gate Array) to get things together into a 32/384 stream, the software at the other end doing similar - there taking it apart. A switch on the DAC lets it operate this way, that switch being Off making it all behave normally (32/192 for physical input).

 

I'm not enthusiastic about PCI express because its a serial interface with a serdes at each end and a not too synchronous communications unlike PCI, but its much faster so that may not affect things as much.

 

Apart from what Miska again explained very well, the length of the interface runs OK on 10 meters and with a modest cable. Nothing can go wrong either (it's all normally error-checked); Jitter can't play a role here. That's further down stream, the oscillators directly clocking the DAC chips via i2s.

 

Peter

 

[PeterSt]

1. The XXHigh End massively upsamples. This lets you get an almost perfect impulse response from the DAC- would pretty much eliminate pre and post-ringing. Some high end DACS like dCs and Weiss (medea at 352.8 khz) do this too, but with hardware.

 

I can imagine how vague or confusing this all is, but no;

 

There is no "massive upsampling" as such at all. It is only 8 times (for redbook) which is just enough to sufficiently eliminate harmonic distortion from the too large steps for the too low sample rate (44.1 redbook) - while on the other hand transients are not killed. I always try to illustrate this by asking to grab a piece of paper, draw a squared wave on it with only a little rounded corners (like analogue will behave) and now start to evenly cut off those corners by means of a diagonal line from one side (vertical oine of the square) to the other (horizontal line). Do this 256 times (which I'd call massive upsampling) and you will end up with a nice sine. Here is your flute which ever was a violin ...

 

A perfect impulse response is not related to this at all (thinking this is about an impulse which does not "ring" (pre- or post echoes). The means of filtering *is* related to this though, and Arc Prediction Filtering from XXHighEnd (not coincidentally) does not ring one sample.

Let the "not coincidentally" be related to "NOS" which should be just like that, where part of that (from origine) is related to "no filtering". So, the "no filtering" *is* there, because it's really not filtering as such what happens. Instead it is interpolated upsampling (upscaling) which nicely takes care of getting the stepping distortion out of the way. So it filters as should, but it's no filter ...

 

For additional confusement : the DAC really "physically" is NOS, but including the software it needs (previous alinea) it's not. Of course it can be used without it (to your decision), but it would be wrong. Not more wrong than current NOS lovers use NOS/Filterless, but wrong.

 

No other DACs (high-end or not) do this, guaranteed.

The most close will be the Pasific Microsonics Model Two (emphasis on "Two") because of its 24/192 design which allows for a similar "NOS" principle, but with the difference that it contains filtering - we all hoping to have them eliminated by the process, because the process shouldn't be doing really much when fed with "already filtered data".

I'll stop here before it's confusing only.

 

2. XXHigh end uses some proprietary interpolation: "Arc prediction" for a more analogue processing

 

Well, true. Maybe not so much "for a more analogue processing", but for a more 1:1 analogue result;

In the other thread I talked about "head nor tail" when the analogue result would be held against the original file; this (A.P.)) just can. And you don't want to see how much off it all still is ... (this was clearly no commercial statement :-)

 

3. the NOS design avoids filtering done by DAC chips, which Peter dislikes. Again, some high end DAC makers do their own DSP which may result in avoiding the "bad" filtering of DAC chips.

 

I don't think that something like "bad filtering in DAC chips" exists, although quite many can, and for almost as many it can be shut off. But the point is, next comes that DSP chip (or separate filtering chip), and there you again have what I indeed don't like. Mind you, just for theories.

 

4. By pre-processing the data in XXHighEnd, the computer isn't working hard at the time of playback.

 

Assuming you refer to the upsampling process ... it depends on how you look at this;

Mani's outlay on laptops not being fast enough is true; however, it is true only for when music starts, or for when the digital volume changes; you'd have to wait for it "to come through". More to the real merits this description isn't literally correct because of parallel processing and settings dictating that. So, without what you say being untrue, during playback this process happens just the same, because when a next track comes along you don't want silence. So, in a parallel process this happens - during playback itself. You won't notice it. Still for the first notes to play, you'd have to wait for it.

 

4. The Phasure Dac has extremely low jitter.

 

I guess it has. Partly this is because of the best low jitter oscillators available (mind you, for audio and which is all about phase noise), but the largest deal here is the direct connection; as I said in the earlier post, there's just nothing in between those oscillators (each separately shunt regulated) and the chips that can add jitter. So the figures should be net. I must honestly add that so far I had no good means to really measure "figures" as such, and all I can tell that there isn't any way I can find a single spur of jitter by means of graphs. Only when something is wrong (poor oscillator) I can show it, and what shows, shows 120dB down. Says nothing, and let's stick to lacking equipment for it.

 

So I wonder if Peter hasn't (entirely to his credit) recreated much of the hard work done by certain DAC manufacturers, but through software? Is that fair to say?

 

Please allow me to speak for myself, but I really don't think so. Hopefully this shows from my responses above ?

It will be true that The Phasure NOS1 does things through sofware which others do in hardware, but what it does through software is different from anything else. Besides that, it allows you to choose yourself what that "through software" exactly is. You are given a hint by the software provided via XX, but nothing witholds you from improving on it if you think it can. Or more easily thought : just use and compare anything you can find (online (real time) or offline (pre-processed into a new file) and pick your choice. One thing you will have a guarantee about : nothing in the DAC will change whatever it is you do.

 

Let me thank you for putting up questions which really matter; You didn't get the answers you maybe expected, but just for that reasons the questions (or statements) were good.

 

Regards,

Peter

 

PS: I bought a new PC recently; it costed 1000 euro exactly including one 90GB SSD. Not 100% silent, but silent enough. But notice that any more modern PC will do. If there's only no energy saving stuff in there (laptop processors etc.).

 

[PeterSt]

Wait ... please ...

 

If this means that people may be offended, or this is appreciated as inapproriate otherwise, I am happy to answer questions - but not if it really shouldn't be done here.

 

Yes, maybe I like to share because I'm proud a little too, but it really would be ok for me to answer questions on my own forum.

 

On the other hand, apologizes if I read this as a message while it is not one at all. English you know ...

 

Regards,

Peter

 

[PeterSt]

Ok, thank you guys. That makes me smile again.

(but if Chris interrupts this ...)

 

Barrows,

 

Actually this is the most interesting question.

But before I will give some sort of an answer, I was thinking ...

 

You wouldn't mean DC Offsets which may be eliminated by digital filtering, or do you ?

 

If not, for fun and honesty ... what would you think will be acceptable for a(n electrical) DC offset ?

 

Btw, by accident Mani already knows the answer, so I'm afraid I can't lie (later).

 

Peter

 

 

[PeterSt]

Yes, easy answers obtain priority :

 

17.32" x 17.32" x 7.5cm.

 

The height excludes any feet (3 or 4) of 2" more or less because they are not provided.

 

Oh, for people not common to inches and stuff, I calculated 2.5cm per inch here.

 

And by these larger dimensions (feet, inches or otherwise) all the connections are on the inside. So, sticking out nowhere (see earlier picture).

 

Yes, I promised pictures, and promised myself each day. Will happen tomorrow. Maybe someone should remind me and keep me to it.

 

 

 

 

 

[PeterSt]

Hi,

 

Earlier in the thread you asked a one line question. I think I thoroughly answered that. For your rememberance, this was your question :

 

Peter, but what about doing the upconversion offline?

 

Now, what about thinking about the time the answer took me ?

And how easy it would have been to say something like :

 

"No Ted, please think further. I don't think that's a real option."

 

Maybe I got carried away, and too extensive answers are not appreciated.

I will try to keep some senses.

 

Peter

 

PS: We posted along eachother (and never saw your last one); otherwise I would have said this first.

 

 

[PeterSt]

(the self-congrat comment is too over the top and not warranted IMo).

 

Man, I am sorry. Please understand that this only takes some more seconds to read this correctly (but blame my englisch not being able to do it within the second). I will take the blame fully on me, but this is how I read (again, too fast), and allow me, it's just for your fun (hopefully) :

 

a. Gkoone's questions should be more private ? eh, huh, who said that ?

 

b. No, nobody did. Hmm. Ah, wait, DISagree. Ok. Still nobody said that.

 

c. The quote above (confused because of the before) ... self-congrat. Man, see, it *was* a message, and here it is repeated. Not warranted. Ok, so not. Well, I can agree.

 

d. Heck, why does Ted say this. He asked a question himself. Ok, answer was not appreciated I guess. Well, why answer longer than the question ever was in the first place.

 

Ad c.

Yes, I left out the "however" (never read that), and I guess all it needs now, is to see that "comment is" should really be "comments are" to make my failure here not happen.

All it further needs is your comment about USD 3900 which sounds a lot more to me than EUR 2900 (but which sadly is so at this moment ... just wait :-), and some comment about things not fitting in any rack. Disaster is about to happen (my side). Of course I continue to read that this time you won't agree with Barrows (who shows interest in, say, "me"), and at this moment I am biased enough to think you have an investment in the Metric Halo or something. All it needs further is to briefly browse throuhg this thread and meet your post titled "Ignorance is showing here".

 

Again, this was for your fun, but this is what happened. Even waited an hour before clicking "Save" to the post, because I really couldn't imagine you "behaved" like this. Still I did against all odds.

 

I was wrong.

Peter

 

 

 

[PeterSt]

I'll try to be brief ^-)

 

The subject has been passed numerous times, and I think I never have interfered in any of those threads about it. But IMO :

 

A (connected) Word Clock as a master can't do much in a system (DAC) that is really bit clock driven (the Word Clock formally derived from that, or from the Master Clock operating on an even higher (or the same) frequency).

For SPDIF I am not 100% sure about this all, because SPDIF is (a clock signal) "derived" to begin with. Still, in-DAC a bit clock would be some master of it all first, and I don't think there's a good way to derive the bit clock (as a devider ?) from the Word Clock. Of course, the Word Clock determines jitter (letting the samples go and such), but still the bits must be readily available first.

So, for SPDIF maybe, because I just don't know it really, but for I2S, no, I can't see how (btw, the length of the wire with the clock signal creating jitter already).

 

Peter

 

[PeterSt]

Let's start with this :

 

My opinion is that the lower the offset the better, below 50 mV is good

 

Haha, you just must have asked Mani.

Ok, kidding, but this is exactly what I said to Mani of what would be within specs if I had to determine them.

 

You didn't ask, but a well behaving DAC would be under 20mV on all output pins (which are 4, thinking balanced). A very good one may be all under 12. My own does 15. Usually two will be under 5, and one will be 0. It depends on how things "balance out" if I'm allowed to say it like this.

 

to deal with DC offsets there are basically three ways: design a circuit which can be carefully trimmed adjusted to have no/minimal offset, this usually requires a somewhat complex circuit. Use a servo feedback loop to eliminate offsets, or AC couple the circuit using caps. My feeling is that each of these approaches has advantages and disadvantages.

 

Maybe you didn't want to say it exactly like this, but I don't see any advantage. Well, except for keeping the DC low of course.

 

Earlier I told you that the subject is the most interesting, because, well, it just is. It is when you don't like to apply any circuitry (up to one cap) like I do. And hey, isn't this why you asked the question in the first place. And so I don't. Well, that's what I told not even in between the lines.

 

Now, to your surprise perhaps, I don't like much to tell what happened to let this happen, but what I can tell at this moment is that it depends a bit on the environment so to say. People may receive 50mV or even more, and although totally harmless it clearly is not the best for the best sound. In theory that is. Right now I just have another little project behind me, which guarantees the lowest possible DC a DAC can inherently have. With "inherently" I mean the internal current flowing which is unavoidable. But, it makes it all independent of the outside world to the DAC, so people who now may have a DC offset of 50mV because of environmental influences, *will* end up with that under 20mV of which I know can be the inherent DC (max that I have seen so far).

I am not 100% finished with this yet, but when I am people will be able to "upgrade" themselves.

 

No real answer, but I hope it will do, that indeed (still) nothing is in the signal path except this one resistor. Anybody can correct me, but I think it is still much better to have a little DC (which is about stuff like absolute phase only -> 50mV against 1.5VRMS which is about 42 times more and that against some wave shape and phase "offset" (per pin)).

 

Lastly before I forget (and hoping nobody made it till here because I'm almost too embarrassed to say it afer being 18 months or so being plain against it) : gain is done by an opamp. It really turned out to be the only way, and when I at last gave in, it turned out to even improve SQ which I never imagined. People who followed all a bit will know how many times it failed before I suddenly found a way which just worked. Is that in the signal path then ? let's say it is. But it depends on how it's used, or better : what it is used for. That is why I count it out.

 

Peter

 

 

[PeterSt]

Trust me that I didn't see Mani's post before hitting Save.

But somehow I had rather seen the two posts the other way around.

 

[PeterSt]

how about some of the more down to the earth details.

 

10.5Kgs

 

How long is the warranty?

 

Wasn't that stated two years ? maybe it was three. I really don't care much.

Once I had to pay for a car repair because warranty was over due 3 days. I will never forget that.

 

Is it sold directly or through retailers or distributors?

 

At this moment directly only. When that changes, price will go up. Turnover will be bigger, profit probably too.

But I guess this all didn't start out like that, and preferrably it will never happen.

 

What’s the financial stability of the manufacturer to back up that warranty?

 

Zero to none.

 

Despite my funny (meant) answer, I can tell you that your focus is on the wrong subject. Well, that is my personal opinion; It would be better to investigate how many PCM1704U-K chips I could obtain today, may it be for your DAC for tomorrow. The answer I just gave above.

Some people will sell their boat when things get tight. I will sell the 1,700 PCM1704's I already have, might that happen. At double the normal price. Next I will sell my boat.

 

I hope this tells you something about the real answer. Probably not (I wouldn't dig it :-).

 

 

 

[PeterSt]

Hi Demian,

 

It the PM Model 2 there is a dual crystal oscillator located very close to the DAC's. It runs at the "master clock" frequency. It is divided down to the bit clock and to the word clock frequency, a process that reduces the phase noise, the jitter remains constant. The external data source when locked to the word clock should have essentially no effect on the clock if the isolation is good. The key issue becomes eye pattern and data accuracy which is essentially perfect in this scenario. The data stream is driven by the clock next to the DAC, which is again ideal. (The PM uses AES dual wire for 192, not SPDIF.

 

You will know more about the PM M2 than I do, but I can not understand all of what you are saying here.

 

It the PM Model 2 there is a dual crystal oscillator located very close to the DAC's.

 

It is allright to have oscillators in there, but they (IMO !) will be there for reclocking purposes only. Or some anynchronous SRC.

 

The data stream is driven by the clock next to the DAC, which is again ideal. (The PM uses AES dual wire for 192, not SPDIF.

 

Functionally speaking, AES will be SPDIF allright. Now, I really don't see how the data stream can be driven by oscillators in the DAC, while at the other end some oscillator(s) will be running, and *them* obtaining the data really;

If the connection would be async FireWire, yes, then I could understand.

but

Although I don't read it as such (and it will need the context of earlier posts I guess), when looked at this some other way around I don't see much what is wrong with your outlay. So, when the clock of the PM2 is connected to the original source, and *in there* the clock virtually is replaced by the one(s) in the PM2, yes. But only for theory and up to some degree, because it again (see my earlier post) depends completely on the architecture of this other source, and obviously this architecture is not under control of the PM2. So YMMV - it is nothing to depend on - users can not even check the real merits except for listening.

 

If we additionally take into account (and of course you said that yourself I think) that no word clock connection has ever been made for an application like this (better jitter specs etc.) - certainly not 10 or more years back - then I guess all we are doing is finding good reasons for an existing word clock connection of which we actually don't understand why it is there in the first place.

That all the Big Ben's jumped onto this market and that people happily buy these devices is good for them, but we may really wonder what perceived differences are really, while I'm not saying that they are not there. But as tests showed (sorry I don't have a link), all of these devices could only equal the jitter performance of the original source (IIRC but one, but now think of my "YMMV").

 

FYI (and where I stand in this), the specs (not up anywhere yet) of the NOS1 state explicitly that no word clock connection is provided because there won't be a single way it can be improved upon. It also states that the connection *is* there though, but internally, and with some DIY it can be exploited. Instead the specs also state that the oscillators can be replaced within 5 seconds (not counting the removal of the cover), might better ones become available. They could have been soldered and be more or less fixed in there, but we did not, for the better reason of being able to replace them in a whimp.

And of course this is all so because I think this is "the" way to deal with these things, or otherwise the word clock connection would have been there for commercial reasons at least.

 

 

It is divided down to the bit clock and to the word clock frequency, a process that reduces the phase noise, the jitter remains constant.

 

This may come across as nit picking, but if it were for me, then each division of the clock will result in additional jitter. Something like 6dB.

About the reduction of Phase Noise I can agree I think. But this will be merely because of the specs of the oscillator which are unavoidable for its frequency, while not using that frequency in practice (but a division of it).

I can easily be proven wrong here though.

 

Different DAC architectures are more sensitive to jitter at different clock inputs. Its not obvious from the outside which is more important without a good understanding of the guts.

 

Obviously (to me) this is completely true, and just what I said myself. So, I'm sure we don't disagree much.

I also don't want to know it better for someone who never even has seen a PM2 from the outside.

So let's wrap this up by me saying that I have my own ways, but also that I certainly don't know everything and all.

 

Thanks,

Peter

 

[PeterSt]

Mani,

I was hoping you could post picture(s)

 

Hi Tim,

 

Please don't blame Mani for not doing it. I said I would do it, so he needn't to. But I guess I'm not good at doing things some half way, and at promising "pictures" I want good pictures. This is not easy, and it may take me half of a day. It needs the good light too, and blahblahblah. But I'll make a picture of that interface in a few hours time now, after sunrise.

 

Mr Commercial. Hmm ...

 

 

[PeterSt]

At last ...

 

This is a not completely ready DAC. But it allows for some site-seeing.

 

 

 

The PCIe (1x) interface in the PC is in the middle of the picture here.

Some dust cleaning is needed allright.

 

 

 

And the outside of the PC ...

 

 

 

Towards the DAC ...

 

 

 

Connected now ...

 

 

 

... in some perspective ...

 

 

 

Seen from the inside of the DAC.

 

 

 

From some other angle ...

 

 

 

... in close up.

 

 

 

But back to this angle.

 

 

 

From the other side ...

 

 

 

... and from this close ...

 

 

 

... and again from another angle ...

 

 

 

... and from the outside of that.

 

 

 

From far, far away.

 

 

 

But from closer again here with a walk around.

 

 

 

And step a little to the right ...

 

 

 

... or a little back.

 

 

 

Seen from the interface itself ...

 

 

 

... in closeup.

 

 

 

Seen from that closeup.

 

 

 

And from a helicopter ...

 

 

 

... just on top if it.

(due 8 channel X-Over version, 16 chip)

 

I hope this suffices.

And isn't too much of it.

 

Peter

 

[PeterSt]

Thanks Ted. But, strange. Here they're just visible. Also from the "outside" (a VPN).

 

But I have heard similar before ...

 

If you briefly ask for e.g. this one : http://www.phasure.com/index.php?topic=833.0

(all old stuff)

Do you see that ? (pics)

And if so, if you now refresh this thread, does it help ?

 

I'm off and dead anyway, so I hope to read your response later.

 

Thank you,

Peter

 

[PeterSt]

Mani - Maybe it's a stupid idea, but why not ask Keith Johnson what he used in the past for best performance ? Of course he shouldn't come up with equipment we don't have in our homes (hence can't judge that). Anything "modern" wouldn't be helpful for an answer either I think, but since the thing will be 10 years old ... what did he use back then ?

 

Could be some Lynx ...

 

Maybe to keep in mind : It is my personal idea that all Firewire is a kind of "out of control" so to speak, be that BridgeCo based or be it DICE (Weiss afaik) based.

RME is left out of this equation, because it is proprietary.

 

You may not know about the subject, but I have been working for a year (throughput time) on a Firewire interface, but when I at last had everything together (my !) latency specs couldn't be met, and I gave up for that reason alone (apart from it's a mafia world). This was before we could use the ultra low latency in software, but I guess since we now can, the data prooves that this is just so. You were one of the first to find out ... (Jealous Guy).

 

Just speculating ...

Peter

 

[PeterSt]

It is a public secret that the cost of parts is around half of the sales price. I ever back put that on my forum somewhere, so it is not a secret at all.

 

For calculations, if needed, we have PCM1704 chips worth of more than your car (this would be gkoones car btw). This not only tells about some stupid value of things, but also that it all includes high quantities in the first place, or otherwise the whole thing wouldn't be affordable to anyone.

 

So, half instead of 10 fold or more.

 

Next, Miska is quite exactly right about two days of assembling. But FYI, "assembling" is about putting ready parts together. Putting them in the case, so to speak.

But to be honest, this excludes quality problems, like things not working, bad THD figures or anything, caused by unknown causes (like a faulty capacitor, whatever).

Lastly, nearly the whole thing is (SMD) soldered by ourselves. And no, this is not included in these two days, nor is it obtained in parts costs of course. But I like to see you do that within another two days, whether you are a prisoned Chinese or not (this hardly makes sense, and will only tell how stupid it is to do this ourselves).

 

The software ? think about maybe one year for this alone, but add a few years to make it "workable". All the trials of obtaining hardware ? think a small car for Firewire attempts I talked about, which really are all in the trashcan by now. So already that.

A DAC like this really should cost 10K at least to cover for this stupid stuff alone. And then I calculate 100 for sales.

Now multiply. Just for fun. That would be 7000 euros more for a 100. The figures match what has been spent. It can't be gotten back by selling even a 10000. Why ? because of those 2 days assembly which are already excluding the soldering of the PCBs.

 

I don't tell this because I'm pittyful.

But it does proove that I am quite crazy, and maybe that it doesn't feel nice to read that someone thinks that this -of course- is all about making crazy amounts of money. So that disturbs a little, and that's why I tell you this. You just don't have a clue (as it seems).

 

Lastly, if not clear already, the cost of the parts is "only half" just because of the quantities bought. So it's also about investments (which phenomenon should tell that it is about earning back whatever factor). Listen to someone like Barrows who also builds a fine DAC. But for himself. What did he say ? 2000 euros or dollars, I'm not sure. But the best parts to be found, and they cost just that. No case yet, no hours charged, and the development was just done by other fools who don't charge these hours. Too much maybe still because of not buying in quantities, but also just 2000 euros. Who cares when you have 2000 euros spare anyway. But now what if that is times 100 just to keep the price down (yeah yeah, the 2000 will drop).

 

Hey, no problem. Let's say it is just business. Anyway I decided for it myself. No complaints anywhere.

But now imagine that it's not much nice to hear what you are telling about this all. Design your own case (instead of getting yourself any cheap looking thing like I would do if it were for myself), go with it somewhere, ask what it costs, and be sure you have it 100% right before giving it into production. If not, you've spent 100x that amount and ditched it. Aha, so you make a prototype. Good. Now ask what that costs. Can you do it in one time ? good again. If not, spend that money another time, or otherwise you will loose a 100 times that money. Of course, you can make 20 only ! But now ask for the difference in price.

Wait, you wanted aluminum. Also good. Much nicer. Start with 6000 for that mall needed and hope you do it allright in one go. If you didn't, spend another 6. But keep your sales target in mind, and see the revenue drop only because you had to spent that other 6k. So rise the price.

And see the sales drop. And ...

 

It's not that easy.

Peter

 

 

[PeterSt]

The last thing I want is interfering with this. So, let's say I don't. But but but ...

 

As you know the SPDIF out on the NOS1 currently is there as a "gadget". But, also it is there for people to try out my "theories" on a very low jitter SPDIF-out. And, we (you too, Mani) know of at least one customer who tried it for its merits on his Audio Note. He received a totally different DAC (AN) by it ...

So let's say that just works.

 

Now, with your second NOS1 being at the "painter" right today, what would happen (at least for you) if we'd try to make an AES output from it ? It would be 24/192 only (so, no 384). Would that technically work for the Model Two, or does it need 2x 96 ?

 

Don't ask me what it would take and cost to make it (well), and it will be for some further offline communication anyway. But ?

 

Peter

 

[PeterSt]

I hope to have a demo board in a few days that will give the clues to this solution. If it looks like it will work I'll pass along how to do it. If Peter is up to tweaking one and hooking it up to your box (in place of his DAC) that would be the best way to see how they compare.

 

But of course I want to try !

And honestly -as I told Mani privatly- I currently wouldn't know where to start at making *this* 2 x 96 into one 192 at the other end. And I didn't suggest a "if Demian would ..." - but I was sure thinking of it. So if you now come up with this yourself ...

Seems great.

 

I'd say, take your time with this (so I have some time left for now :-), and maybe it's best to continue this via email ? Well, if you want, contact sales at phasure.com

 

Thanks a lot,

Peter

 

[PeterSt]

The Phasure site seems to contain a typo. Peter?

 

Wow, thank you Mani, and Demian of course !

 

Well, I guess that happens when you copy-paste a Word document into an HTML editor.

 

I just tried it again :

 

17mVRMS

 

So yep. Mani, I'm sure you noticed this too ! But you changed the m into an u.

 

Well, that's nice ...

Dangerous stuff !

Btw, 17mVRMS (which for noise really can be 50mV easily for peaks) whould just be "sound" ...

 

In the attachment below a picture of what I just copy-pasted.

 

[PeterSt]

This DAC seems like a shoe-in for a Thunderbolt interface, as Thunderbolt is 100% PCIe compliant...

 

Of course this is correct or a good remark at least. It is also why I said (in this forum) that "Thunderbolt" existed for some longer time than when Apple came about with it, when it was anounced. It existed for ages on PC's (but read the anouncements from back then to get what I mean).

 

In the end we didn't go that route (of other means of PCIe), and choose for async USB instead. The reason ? not all mobo's are compliant to what the NOS1 uses from PCIe and it was too much of a DIY look or feel. I think that matters. So, that there's now the Phasure NOS1 async USB shipping could be called a commercial act, but it is merely about the unexpected support it needed occasionally. Not so with USB.

 

Regards,

Peter