World's first NOS 24/384 filterless DAC

[LydMekk]

Ok...just some thoughts on an sunday afternoon here. Read what Telstar wrote, and although not heard your DACs, I kinda symphatize with some remarks: Several releases of XX have had too much energy in the treble and upper mid. MB this is because Peter has tweaked it that way to suit his own PB chain and speakers, I don't know.

But these are characteristics which also has come forward in my system, which is very different fra Peters.

Also some releases are very 2D and you get the feeling of looking thorugh a window as Telstar mentioned. But then again, some versions are VERY 3D.

The problem is that one versions good SQ is often ruined in the next version due to implementing more formats, changes etc. that has nothing to do with SQ in themselves.

Looking forward to hearing some of these DACs before long.

[PeterSt]

Several releases of XX have had too much energy in the treble and upper mid. MB this is because Peter has tweaked it that way to suit his own PB chain and speakers, I don't know.

Pwew, difficult ...
Ok, some brief (objective I hope) remarks :

1. This should hardly be about my system, but then personal likings, if any. My high output is 16dB uplevelled already opposed to how it officially should be.
2. I don't recall to (wanting to) head into those directions by means of the software. By accident it does though, sometimes.
3. It may be useless to refer to my playback chain from what we heard those 15 hours, because that was SPDIF while the virtues of the NOS1 are with I2S.

The differences between SPDIF and I2S are rather large, but (or *thus*) SPDIF was used for both DACS to compare apples with apples better.

Important or not, I just copied the remarks of the visitors without any colouring of myself, and might the statement have been "the worst system" or "the best system" either wouldn't have said much because "the system" is so much created by the software already, let alone the DAC. But of course the rest of the chain does have its influence, and in the case of the visitors it may be good to say that both were explicit dipole lovers, which is quite some opposite from horns.

In the very end it is dangerous - or not a good thing anyway - to explain (XX) things through a system, no matter how truthful and logical that would be. We never did because we never felt the need to, and that is very very useful because we all could discuss about the same thing. However, I think that a. we all have become much more experienced on listening (I sure have in the past 2 years) and b. we may be reaching limits where the rest of the PB chain just can't be avoided anymore. Therefore, let me add to this "pwew" response my latest experience :

Referring to the post from today from SeVeReD (Re: Q experiment anyone ?) he talks (including earlier posts in that topic) about "a wall of sound". Apparently he can encourage for this by means of Invert On;
Then, this week I visited the designer of these horn speakers, and by means of a small tweak in the filters (about eliminating a dip in the 120 Hz region) he created a massive wall of sound, which I "stated" to be way too much a wall of sound. Voiced and instruments became to big of it (imagine a voice to be over a meter in diameter).
Lastly, by pure coincidence (that it happened right tonight) I was able to play the NOS1 how it should be and how I want(ed) it : completely NOS withpout filter over I2S. Together with Invert now *that* creates a wall of sound in my room and PB system which I could not with this DAC so far (and the software versions since ??).

Now if you watch above alinea closely you see that "wall of sound" (which may be very deterministic for "a system") was created 3 times in 3 very different ways, the 2nd one actually changing "the system" really. The first is just a software setting, the third a change of the DAC including a software setting.
I can't speak for SeVeRed really of course, but if I didn't know otherwise, the other two means of changes would make me swear I was listening to complete different systems. And as you can imagine, more laid back opposed to a wall of sound just *is* completely different.

Btw, the most characteristic of the NOS1 operating really NOS (and filterless) I2S connected, is the unrivaled dynamics going together with the sweetness of I2S. SPDIF connected the dynamics should be as high, but as how it comes to me, is "smeared" again because of the roughness (opposed to sweetness) of the SPDIF connection. So, these dynamics come to you as very very clean. The fun is, this would lead to sterile without further changes, but including the wall of sound andthe warmth added because of that, just makes it a (3D 3D 3D) mixture I LOVE.
With normal sized voices. Haha.

The problem is that one versions good SQ is often ruined in the next version due to implementing more formats, changes etc. that has nothing to do with SQ in themselves.

I hear you very, very well;
As far as I can tell, those reasons for changes are behind us, because it just all has been done. At this moment I can only hope that the current version is good for SQ for everybody, and if not, the Q2-Q5 can do something to it. I am sure (!). The problem of course is how.

Right, having said this all, I think I only now see how to interpret this one :

But these are characteristics which also has come forward in my system, which is very different fra Peters.

... which makes all I just said unnecessary.  After proper interpretation (I think) I said the same as you wanted to express. I never like to scratch carefully typed words (can I ? hehe), but what it comes down to is that earlier versions sounded better, right ?
Hmm ...

[LydMekk]

Earlier versions was not uniformly better than todays, no, IMO. The right description would be more like a rollercoaster ride with incredible ups and downs...

There has been gold nuggets all around. But there's been a lot of cr*p too...

(Beeing truthfully here, hope you don't mind, P!)

What I really would like to see in XX is support for either a simple EQ with some 16-20 bands, or possibilities for linking to e.g. Audiolense etc.

Then each of our (very different) choice of TASTE in SQ reproduction will be taken care of in a good way. I for one would be very happy with an standard EQ.

Sry about the distraction somewhat from the disc. about the NOS and other DACs but the issue kind of cropped up earlier in this thread...

[SeVeReD]

Earlier versions was not uniformly better than todays, no, IMO. The right description would be more like a rollercoaster ride with incredible ups and downs...

There has been gold nuggets all around. But there's been a lot of cr*p too...

(Beeing truthfully here, hope you don't mind, P!)

What I really would like to see in XX is support for either a simple EQ with some 16-20 bands, or possibilities for linking to e.g. Audiolense etc.

Then each of our (very different) choice of TASTE in SQ reproduction will be taken care of in a good way. I for one would be very happy with an standard EQ.

Sry about the distraction somewhat from the disc. about the NOS and other DACs but the issue kind of cropped up earlier in this thread...

But don't typical EQs mess with the phase in a way that just destroys... you'll be chasing your tail forever? Not saying "don't do it Peter", but ... maybe there's a reason he hasn't done it.  EQs and xovers settings are harder to implement (and maybe impossible somewhat) than we all may think.  I agree it's been an up and down ride, but I think we've been heading in the right direction...(says me with being very happy right now and too scared to move to X version hehe).  What worries me is when Peter ups several changes quickly and I dl them without evaluating them against a good old player... I think that happened with later V versions and early W versions and I got lost ... There was some good things and some bad things that happened.... made it confusing.

[Telstar]

What I really would like to see in XX is support for either a simple EQ with some 16-20 bands, or possibilities for linking to e.g. Audiolense etc.

Then each of our (very different) choice of TASTE in SQ reproduction will be taken care of in a good way. I for one would be very happy with an standard EQ.

The "simple EQ" cant do a good job in most cases (including mine).
What would be really good, is to interface (i.e. use a filter) with DRC:
http://drc-fir.sourceforge.net/

It's open source and on total different level than paremetric eq, or even more sophisticate piece of hardware such as Lyngdorf and DEQX.

I know that Peter is working on some filters, so this may be the right moment to look over room correction.

[pedal]

If you want EQ and/or roomcorrection there is one amazing program to buy: AUDIOLENSE XO. It even includes digital XO functions if you want to, for as many channels you like at whatever samplingfrequenzy. At €360.- it's a bargin.

I use it as a EQ/DRC in 2-chanel mode with final Group Delay Correction to minimize any phase errors occuring after the FIR filters. After measurements the program creats a correction file, which I enter into the playback chain, through Convolver plug-in running in a digital loop inside the RME FFMixer software. Life will never be the same.

SOMETHING ELSE:
Peter, -does your new DAC prototype have a capacitor on the output, or is it DC-coupled?

[Telstar]

If you want EQ and/or roomcorrection there is one amazing program to buy: AUDIOLENSE XO. It even includes digital XO functions if you want to, for as many channels you like at whatever samplingfrequenzy. At €360.- it's a bargin.

I use it as a EQ/DRC in 2-chanel mode with final Group Delay Correction to minimize any phase errors occuring after the FIR filters. After measurements the program creats a correction file, which I enter into the playback chain, through Convolver plug-in running in a digital loop inside the RME FFMixer software. Life will never be the same.

Why should I buy Audiolense (which is good btw), when DRC is free?
BTW can audiolense create filters for each channel separately?

PS: I dont have the Fireface and if things go as they should, I'll never have it.

[leifchristensen]

what´s wrong with flutes?
in my system which pedal know quite well the , the sabre dac fed through I2S bus is the best so far
it equals the 5 tower dddac in all over AND bass slam
it surpasses it in resolution(not difficult)
with resolution and low noise-floor comes 3d and the rest by itself imo
so there´s no problem following complex passages of massed instruments and voices e.g. G.F.Haendel Messiah( my desert island favourit)
and this is with the stock IVY buffer version and the std psu´s
(will at a later stage try to build a tube based buffer ala  lampizator, but my tube psu´s tend to be big.......my riaa psu is 48kg for 4 x E810F!))

when it comes to XX I agree that close to perfect sounding versions are sometimes ruined in the persuit of  tweaking of features
agree that som prior versions sounds as if they were tuned to a specific need/purpose(soundwise)
( maybe the treshold for releasing new (improved) versions should go through a beta panel before "final" release)
so e.g let 0.9w-9b stay as the official latest working version till the next level is quality-proven.

maybe remove versions older than ????      there is something called too many options   (also on the different settings)
maybe there should be a working reference with fixed settings(at least some)
e.g 9b with 4/17/17/0/0   
easier to move on fw. when the new (intermediate) reference is set.......no more turning back except for P to locate weird faults
then the beta users can communicate on a select and"closed" part of the forum in the pursuit of REAL PROGRESS
this would also save a lot of ranting in the forum time for Peter

ALL this is of course IMO
ANYWAY THXS 4 the effort Peter,but make it easier on yourself
best
Leif

[pedal]

Why should I buy Audiolense (which is good btw), when DRC is free?

Why did you pay for XXHighEnd when Foobar is free?  
-The point is that digital EQ with lots of FIR filters generate phase errors. They sing like a dead bird, to quote Mr P. So you need a program which can perform a Group Delay Correction in the final stage, before D/A conversion, in order to get (most of) the transient sharpness and dynamics back into the music. But of course, if you wanna live a dull life for free, it's your choice. He-he. (No offense!)

BTW can audiolense create filters for each channel separately?

Yes. For as many channels you like.

PS: I dont have the Fireface and if things go as they should, I'll never have it.

Some players, like Foobar and WMP, accept plug-ins. Then you don't need aditional soundcard software with digital loop facilities. Hopefully XXhighend will have a possibility of plug-ins too in the near future?

[Telstar]

Why should I buy Audiolense (which is good btw), when DRC is free?

Why did you pay for XXHighEnd when Foobar is free?  

Because XXHE sounds better?
AFAIK XXHE is the only WASAPI player available. WASAPI is a much more solid api than ASIO, it's less dependent on the drivers produced by the manifacturers.

-The point is that digital EQ with lots of FIR filters generate phase errors. They sing like a dead bird, to quote Mr P. So you need a program which can perform a Group Delay Correction in the final stage, before D/A conversion, in order to get (most of) the transient sharpness and dynamics back into the music. But of course, if you wanna live a dull life for free, it's your choice. He-he. (No offense!)

First of all, did you try DRC? AFAIK it creates only one filter (per channel), and the filters can be 24/192.
I assume that you must create at least ONE filter to apply any kind of digital correction.

AFAIK, the features of DRC are superior to any other software, but I could be wrong, so if I am, correct me.
It's not like I dont want to spend the 300€, but I believe(d) that besides graphic representation and ease of use, the functionalities of DRC surpass the ones of Audiolense.

Hopefully XXhighend will have a possibility of plug-ins too in the near future?

I hope so, but at first I hope the possibility to use digital filters.

[PeterSt]

Peter, -does your new DAC prototype have a capacitor on the output, or is it DC-coupled?

No capacitors on any of the output possibilities, and no capacitors anywhere in the signal path.
Who'd want those ? hehe

[PeterSt]

Hopefully XXhighend will have a possibility of plug-ins too in the near future?

I hope so, but at first I hope the possibility to use digital filters.

There will be filters in XX. The first will be for applying proper filtering for NOS DACs that work without - or can switch off filtering.
The mechanism of applying the filter (of any kind) will be general and applied onto the music file. Thus, not real time, compared to FLAC decoding.
Any filter available for such an application can be applied then. In a later stage more output channels will be supported, so XOvers will work as well.

The first one is the most important, because it will give NOS DAC users the opportunity to use a filter afterall, that filter not compromising anything to music (output) data ... assuming that I know what I am doing, can mesasure all I want/need and ... it will be dynamic. It will be dynamic/adaptive  to the various input rates, knowing that 44.1 needs a different filter than 96 etc.
When this works out as I intend, it will be a most good reason why XXHighEnd will sound better in NOS circumstances. But of course this is a kind of DSPing, although one of a special kind, and with the best purpose imagineable *and* it will just replace something a DAC normaly does - or should do.
Of course my NOS1 is the "perfect" example to test with, because it just has the options in hardware. If you'd only see that
a. filterless sounds best
b. what a complete mess measuring shows in that case
you'd know how much better it can be theoretically. Of course, unless this mess is what creates the sound kind of explicitly ... then the mess should just stay.

Please note the difference with something like equalizing. I am not in favour of that, and will "allow" it by the grace of the means to apply it (because : any filter can be applied, as explained before). So the only thing which I like about it is that it can be applied without loosing on SQ just because of applying it (mind you, like decoding FLAC is audible when done in real time).

Peter

[PeterSt]

Hey Leif, thanks for your input. A few remarks which I think deserve some attention;

so there´s no problem following complex passages of massed instruments and voices e.g. G.F.Haendel Messiah( my desert island favourit)
and this is with the stock IVY buffer version and the std psu´s

Yes, this is just equal to my first "review" of my Buffalo, and in the end is about my remark how very much more difficult it is to let sound an NOS good in such circumstances. With the Buffalo one can immediately feel with how much ease such passages are interpreted.

when it comes to XX I agree that close to perfect sounding versions are sometimes ruined in the persuit of  tweaking of features
agree that som prior versions sounds as if they were tuned to a specific need/purpose(soundwise)

In itself this is true of course. The latter though is not true. Each and every explicit SQ change emerged from
a. some theory of mine *not* knowing the result in advance
or
b. undoing a degration and hoping that some counteracting means would help.

then the beta users can communicate on a select and"closed" part of the forum in the pursuit of REAL PROGRESS
this would also save a lot of ranting in the forum time for Peter

Haha, but maybe you forgot that XX just *is* in beta ... all the time so far.
Ranting (if any ... I don't think so) is for me about remarks steering me in the right direction again. If remarks/complaints were not there, it can only be me myself discovering anomalies, and who knows I'd never hear them.
Please keep in mind that I hardly have time for listening myself, and that relative to that already one of you may have 10 times more listening opportunities. IOW, don't underestimate what a huge amount of time it takes to find something like "No track given", while the whole stupid thing emerges at, say, 16 minutes of testing. It could take 100 times of testing these stupid 16 minutes, which I rather do without sound after the 25th time.
In the end you could say that I abuse you all, but the more there are, the earlier someone finds this anomaly (which as we know often is dedicated to a few only). In the end it is for the good cause of course. That it takes a few years more than I expected in the beginning, is only because at some stage I thought of the Unattened avoiding the GUI interference, with the example of right now : remote control. It is a kind of stupid of course to remotely control a player which is not there. But okay, with some additional months it will work (the best is yest to come).


And Leif, might you be sailing the North Sea some time ... you're welcome.
Peter

[leifchristensen]

where do you live? I have driven to Nunspeet to Bert a few years ago. after the Kiel ferry it isn´t that far.
reg. all the beta versions:
I´ve just learned slowly to resist trying all the new versions all the time;
out of experience I´ve learned the hard way that I often get into all sorts of hopeless "unsolveable" trouble when installing a more up to date version
I simply don´t have time to test them all
and now I´m very happy with 9b
best
Leif

[PeterSt]

20 Km closer from the Kiel direction ...