Ok, a small update.

I think I have everything under control for connections and the best sound, and I must say that besides all the virtues I mentioned earlier, the DAC now also sounds "sweet".

Sweet is in contrast to the crispyness and sharpness (sharp in the good sense !) an NOS DAC usually shows, and which can be more refined. Thus, that crispyness and all, also shows a kind of roughness, and where "roughness" is a negative, I managed to sweeten that.
Btw in due time I will start measuring the DAC, and show some figures.

What I didn't tell so far, is that nearly all preparations have been taken in order to produce this DAC and sell it. Yep, and in its own dedicated "Phasure NOS1" cabinet. Before production really can start though, one thing has to be worked out and this is a tough one. This is btw also the reason of telling about the production, because I know quite some people are waiting for answers on this, but in the mean time I now know this is going to take a while. So Yes, but Wait.

So what are we waiting for ?

With my ever planned 32/384 DAC I planned a special connection to the PC which has never been done before. That is, not that I know of, and there isn't any reason for it also with other manufacturers;
Then, in this topic I expressed about wanting a general I2S connection. This I2S connection should work for PC's just the same. And as you may know, it is (AFAIK) the most jitter free connection;
I tried this with Firewire as the first carrier, but this is a dead end. I just can't find any Firewire boards with manufacturers that want to sell something, or otherwise the drivers s*ck and access to the SDKs etc. is (too) difficult.
So now I combined my old plan with this new DAC ...

The good news is : I have it working. Thus, no SPDIF, no USB, no Firewire/ADAT or something, but native I2S with a "secret" connection to the PC.
The bad news is : the stupid DAC board only accepts I2S when in Oversampling mode ...

So, after all the hard work it now appears that I2S cannot be used in NOS mode. Yuk.
And "sadly" the nature of the I2S connection indeed sounds better than SPDIF, and it just gives the sweetness I talked about in the above. I know, comparing is a bit of apples and oranges, but since SPDIF can be used in OS mode just the same, this difference in the nature of the sound is very clear.

Again (like with the ESS) 1 million people most certainly will love the sound coming from OS mode. But this will last until they hear NOS mode. Again OS just doesn't cut it, and compared with NOS mode, OS to me sounds like a dead bird. Too sweet, and nothing seems to jump out, while I am sure each musician does his/her stinking best to let his/her instrument jump out. With NOS they do ...

All 'n all right now I am into deep trouble. Well, kind of;
The DAC board wasn't explicitly designed for NOS mode, and although everything looked good so far, this is where the tweaking stops.
Someone has to go back to the drawing board ...

Peter

What I indeed expect is that people experiment with e.g. a better sounding XXEngine3 from an older version with a newer XXHighEnd to control it. Generally this shouldn't be done, and anomalies will come from it most of the time. But in this case the volume mapping has changed, with the possibility of a sudden loud volume at the old setting. So this time I thought to warn people.

Hi Paul,

It's all in here : Vista Users ... prepare yourselves ...

Anymore questions ? please call again.
Peter

You most probably did everything fine, but the procedure just doesn't work (anymore). Or at least not for everyone.
I just sent you a PM though (My Messages above). Please see there.

Peter

Do not combine the XXHighEnd.exe from the zip (right at the bottom) with any older XXEngine3.exe !!
(the volume settings are not concistent with other versions)


The following changes apply to this version :

• When the whole lot of playback means complies to some rules, the Digital Volume now has four times more steps above -48dB (towards -0dB). These rules are :
  
  1. The DAC has to support 24 bits for the sample rate used;
  2. The source file has to be 16 bits.
  
  In this case each of the Digital Volume settings allows to play without any losses as with the rougher steps from before.
  By now we could say this is worth a small patent but never mind; The whole thing requires a. a lot of imagination and creativity (can it be done really ?), b. simulation (how to prove it really is good), c. real practice application (man, is this difficult). So let's wait until this shows up anywhere else.
  
  You could wonder why this out of all works with 16 bit files, and not with 24 bit (or more) files. Well, this is because of 16 bit files don't use 24 bits, and the additional 8 bits (when all is uprated to 24 (hence 32 which it is in practice) bits) gives "empty space" headroom allowing for some very dedicated manipulation. With 24 bit files this space is not empty, and the bits in that space would be molested.
  That's all.
  
  The granularity now is 1.5dB, or better, is "near" 1.5dB. This means that from one step to the other it may go like -1.2dB, -1.6dB, -1.55dB etc., while all is visualized as 1.5dB. Only at the 6dB boundaries from before the, 6dB is exactly that.
  
  Important : Partly because of the complexity, partly for your convenience, in all cases the higher granularity is allowed to be applied. Thus, also when your playback means does not comply to the two rules mentioned above, it is allowed to use the 1.5dB steps in between the 6dB steps from before. However, in the cases you use the 1.5dB steps while it is actually not allowed, this will be shown to you;
  
  Although not so easy to explain, in short it comes to this :
  We have "legit" stept and we have "not legit" steps. Now look at the pictures below;
  
  The first picture (Volume01) shows the situation when your playback means comply to the rules. Below and above the volume control two buttons show, and both buttons have a green color. The volume is set at -45dB here, and this is a legit volume. Legit means : no losses.
  The button below the volume slider allows to notch down the volume by one step, and the button above the slider notches up one step.
  
  The second picture (Volume02) shows the situation when the payback means does not comply to the rules. Note that the only thing changed to above situation, was loading a 24 bit album.
  Now two sets of notch up/down buttons show, and the left set shows in a red color. Red means : wrong ! In other words, -45dB is not good for lossless volume.
  The right set of buttons shows a green color. This green color only tells you this : when clicked on a green button, you will receive a legit volume.
  
  Now look at the third picture (Volume03). Here indeed a green button was used to notch up the volume, and the -45dB changed to -42dB which *is* a legit volume in this case. That this indeed is so, can be seen by the now black color of the left set of buttons. Remember, red = wrong.
  
  Summarized :
  
  - When the playback means does not comply to the two rules, two sets of notch up/down buttons show, and the rightmost set -showing the green color- will notch towards a legit volume. The leftmost set can still be used, but it will notch to any available 1.5dB step, and when you end up at a not legit volume, the left set of buttons will show the red color. Otherwise they will show black.
  
  - When the playback means *does* comply to the two rules, only one set of buttons will show, and they will always show green. This means that any click on a notch button will end up in a legit volume.
  
  Not to forget : Whether your playback means comply to the two rules, depends for rule #1 on the settings you set yourself. This obviously won't help you tweak things for the better. Thus, having a 16 bit DAC but tell the system it is 24 bits really, really won't make your volume better (if sound would come from this at all).
  
  Btw, clicking on the "V" label (showing the volume level) as introduced per 0.9w-6, has been disabled again.
  
  
• Theoretical support for Windows 7. This has not been tested, but Engine#3 should run as under Vista compatible OSes.
  When you are running Windows 7 and this is properly recognized by XXHighEnd, it should show "Windows 7" in the Playlist Area when no tracks are loaded (hence press clear should show it).
  
  
• It is now possible to play the files from the "XX drive" (only useful at UnAttended Playback !!);
  
  See Settings Area and the new checkbox "Copy to XX-drive by standard".
  
  The XX drive is supposed to be the drive where XXHighEnd resides which should be the same as the OS disk.
  Note : this latter is not mandatory, but if not this feature won't have an objective.
  The objective is to load the selected files from the Playlist onto the "OS disk" first, and play them from there. The result will be that the disk(s) where the files originally are stored is/are left alone during the time of playback so they will spin down (at your OS settings concerned). Furthermore it is the virtuual intention that the OS disk is a Solid State Disk which doesn't spin at all, and the whole lot implying better sound quality.
  
  Kind of important : In order to really let this work, the coverart needs to be loaded from the OS disk as well. This concerns the main coverart picture as shows in the left pane of XXHighEnd, or on the Wallpaper for that matter;
  This all only works when playback is initiated from a Gallery. Only then the coverart is available from the OS disk, assuming that XXHighEnd is installed on the OS disk.
  And ... note that before the implied coverart was loaded from the disk containing the original source; not the XX-drive ! Also, many lines of code had to change in order to let this work, and explicit testing for all the dozens of places this has impact, has not been done. Please let know when something produces errors now, or when wrong coverart shows.
  
  Note that with Cue Files (or very large files otherwise) this takes a noticeable time before playback commences, because the whole large WAV file needs to be copied to the OS disk in advance of playback, and although the setting "Start Engine3 during conversion" (Settings Area) may be applied by you, playback onviously cannot start before the whole WAV file has been copied.
  Also, when one track is played from a Cue File, again the whole large WAV file has to be copied first.
  Besides this, the copying is a kind of slowish, hence it looks like a faster means can be created. Later ...
  
  The copied files to your XX folder will disappear at any one moment XXHighEnd is quit (Off) while no playback is going on. Keep this in mind for your possibly not too large SSD containing the OS already, because you may run out of disk space there.
  
  Important : The individual files subject to copying to the XX-drive will have unique names equal to ... their names. This means that when two (or more) tracks with the exact same name are subject to this copying, a second (etc.) track will overwrite a first, and although these tracks will all play, they will all really be the same as the last one. Don't get fooled by the uniqueness of track numbers, because it may happen that the tracks subject to this don't have track numbers really, and they are only shown (added) by XXHighEnd.
  
  Lastly, keep in mind that using this setting only has the implied effect opposed to not using the setting, when the tracks concerned don't need to be converted in the first place. Or in other words : each file which needs some kind of conversion, will have its converted result on the XX-drive anyway.
  
  
• Support for FLAC Cue Files.
  
  Note : Currently this will not work for all means tracks can be loaded into the Playlist Area, but selecting albums in the Library Area following by clicking "Load" will work, and all the tracks from the Cue File should show.
  Since pretty much nothing else has been tested regarding FLAC Cue Files, even errors may occur at the attempts comcerned.
  
  
• Most probably since the first 0.9w version, rips from dbPowerAmp showed static at the end of each track.
  Note that dbPowerAmp (stupidly) adds tag data at the end of tracks.
  This has been solved again, but because of the totally different way of working since 0.9w(-5) all theoretically is subject to the ending of playback at positions where it shouldn't. With "all", all means of sources and combinations with 16/24 bit DACs, upsampling and the whole lot of XX settings (like Ldn (Less Dynamics) etc.) is meant. Much has been tested though, and thinks look OK.
  
  
• When Crack Detect (Settings Area) is disabled, a message now shows that this really should not be done, unless decoded material (like DTS) is played.
  Never shut off Crack Detect for normal material in order to skip any Crack Detect message, unless you anticipate on static. However, *if* you do so and don't notice any anomalies during playback, please let this know instead of just doing it. Remember, the necessity to do it for sure implies something to be wrong somewhere.
  
  
• It was found that Cue Files really did not make use of the "Split File at Size" setting (Settings Area). Now they do.
  
  
• It appeared that clicking Next when a Cue File was playing Unattendedly (Engine#3), the time position would not reset and playback of that next track started somewhere in the middle. Solved now.
  
  

Edit 1a : Do not download this version if you wish to play Engine#3 Attendedly only (Unattended checkbox at the bottom). Half of the tracks will end in a crash of XXEngine3.
No problem with Unattended Playback.
Edit 1b : It looks like this can be solved by downloading the zip here : Re: No Track Given Error. Unzip it over the folder you used for the main zip file below.

Edit 2 : It was found that the digital volume is wrong at -19.5dB ! (way too low). So skip the -19.5 volume setting please.

Ok, thanks ! It's in there for 0.9w-7.
Peter

This requires explicit recognition of the OS from within XXHighEnd, and it is not there yet.

Unzip below 0.9w-6a version into your current XX folder (and save the old XXHighEnd.exe first !), and start it.
Set switch Log Activities to On (Settings Area), quit XXHighEnd and restart it.

Go to the TemporaryData folder (subfolder from your current XX folder), and find the newest XX- logfile. In there the second line shows how the OS is reported. Please copy-paste (not write it yourself) this line (which looks like : OS Name is ...) in here.

After doing this, set back your saved XXHighEnd.exe version, because 0.9w-6a is half baked !! and it should not be used for normal operation.

If you can do this ASAP (hence within a few hours from now) you can use Windows7 in the upcoming 0.9w-7 version.
... If it works for Engine#3 ... which we can't tell yet.

Thanks,
Peter

Hi Paul,

As I just said in the other post ... the means how USB is used by Empirical is not conform standards.


That would only be the case when Engine#2 is used instead of Engine#1, and this should happen under XP only (files can be too large for Engine#1 under XP, not under Vista).

Peter

Hey ... that is very nice to hear. Thank you !
But I sure hope you can create something for the OffRamp to work. I now recall that Emirical doesn't use the normal protocol for USB, which really doesn't allow for 24/96 (16/96 it does), and which will be your problem. In the end it will be WASAPI not understanding this protocol.
But maybe I'm wrong.

Peter

Yes Chris, you are right.
Gerard, that is a good idea (sometimes things are so obvious !).

Thanks,
Peter

Got it. Thanks Glynn.

And guys, *never* switch off Crack Detect. It is there for a good reason. It is allowed to be switched off only when you know in advance things are just right, like with DTS and a DTS receiver;
If you have to switch it off for normal playback because otherwise things won't play, please warn me instead ! Even if you don't hear anything wrong with it !!

Peter

I can't solve it without an example Glynn. Please don't forget this !
Note that it happens at the end of a track (and not at the beginning).
Thanks ! Peter

Glynn, I have a rather important question in order to find *where* I went off track :
Was this the same with 0.9w-5 ?


Btw, here's the reference to an earlier thread about this (which reference is merely for myself and checking dates and things in the program) : 0.9u-6 -- Loud pop when playing next track

That is not difficult at all. However, it depends on the granularity I can achieve. And it looks like this is less as I wanted.
Man, this is difficult in general.

bgt, just sent you a PM.

Peter