Oh, and I forgot to ask...is this buffer a dual mono configuration or do the two channels share the power supply?

Thanks Peter.

Don't worry about the vinyl.  I am more concerned with other analogue inputs such as an FM tuner et cetera.  It would be nice to (still) not have a preamplifier but to be able to use your active buffer for more than one input.  So if I was listening to some music through my DAC and then wanted to tune into a live FM concert it could be done without changing cables and all that associated messing about.

I could sort out a remote volume control whether that be a high end passive attenuator or something with the Muse chip and then I would have the flexibility I need in that department.

Reading between the lines above it seems as though you don't think more than one input could work optimally for SQ...is this true?

Regards,

Anthony

Pictures please Peter.  It is a pretty big box.

A couple of questions...

Can it be built into the NOS1a?  I would prefer to have fewer boxes rather than more.

Do you think a switch and phono-in and perhaps a volume control could be incorporated?  After all it seems as though this new device is acting as a preamplifier.

Do you have some specs for the new device?  Gain, power draw etc.

Regards,

Anthony

I just spent half an hour reading that post Peter and I am no closer to knowing what the New Device is.  At first I thought a new I/V stage for the NOS1a and then I thought you have developed a preamplifier and later still your writing made me think of a passive buffer to go between DAC and amplifier (eliminates DC offset, breaks ground connection) but your descriptions are really of something between amplifier and speaker...at least I think they are...unless you have developed an active buffer.

Scratching my head...

Ok, so many of us are using RAM-OS now and our audio playback is run completely from ram.  Is it time to take a closer look at the ram we are using?  I ask because there is a group of people in my part of the world that claim single-rank ram sounds better in their playbacks (not XXHE or NOS1a playbacks mind you) than dual or quad-rank ram.

From the little reading I have done, single-rank ram is about 7% faster and the whole of the ram can be accessed at any one time, whereas the other configurations mean only a portion of the ram is accessible at any one time.  What does this mean to XXHE?  I don't know.  But with RAM-OS we are in a position where this sort of thing might just make a difference.

Single-rank might allow us to use even lower SFS.  It might have other benefits or it might do nothing at all.

Is there anybody out there using single rank ram and RAM-OS?  I am considering getting some just to try out.

Cheers,

Anthony

After making the post above, I decided to update my profile so that Peter could get a better idea of what my settings are and the gear that is in use.  Well,in the Xeon RAM-OS I found two settings that were not the same..."Balanced Load" was 64 (not 63) and "Provide Stable Power" was 0 (not 1).  I changed these in W10586 RAM but they did not hold, so I booted back into W10586 Base and changed them there.  When I went back into W10586 RAM the "Provide Stable Power" had changed from 0 to 1, but the "Balanced Load" had not changed as was still on 64.  Not sure what happened there.

So I retested the Xeon and this is what I came up with:

Item	Xeon	i7-3930K
Start XXHE	18s	18s
Load mixed 41 song directory from embedded Explorer	7s	8s
Play same mixed 41 song directory	1m32s	1m31s
Load Music Root in embedded Explorer	5s	4s
Load Chieftans Album WAV format	3s	3s
Play Chieftans Album WAV format	18s	18s

So the Xeon and i7 are level pegging now that the Xtweaks settings are similar.  Peter, do you think the times similar to those you experience at your end?

And call me officially "amazed" at the performance change of the PC with one Xtweak setting changed one digit from 0 to 1.  Wow!  Not sure if I should try to get that 64 for "Balanced Load" down to 63.

Hi Peter,

I have here my XXHE PC which I made myself some time ago (with your direction) except I used a second hand Xeon processor rather than the i7-3930K that was recommended at that time.  Here is a link to a specification comparison between the i7 and Xeon.  Note the only real differences are in L2/L3 cache, number of cores, bus architecture and base clock speed.

I have had another members XXHE PC here that is identical to my own with exception of the cpu...the PSU is the same, ram is the same exact model and amount, same coolers, wiring, RAM-OS and so on and so forth.  My Xeon based AudioPC takes more time to do pretty much everything than the i7, but I think that the Xeon sounds marginally better (need to do more testing of that finding).  I have compared the time it takes the two computers to perform identical XXHE related tasks both controlled from the same Music Server and both running identical XXHE settings (see below)

 

Item	Xeon	i7-3930K
Start XXHE	31s	18s
Load mixed 41 song directory from embedded Explorer	13s	8s
Play same mixed 41 song directory	2m23s	1m31s
Load Music Root in embedded Explorer	22s	4s
Load Chieftans Album WAV format	3s	3s
Play Chieftans Album WAV format	23s	18s

So you can see the difference between the two is significant.

As far as I can tell the only difference between the two PC's is the processor, but on paper the Xeon should not be so far different in performance to the i7 after underclocking, so I wonder if you can see anything else that may be causing the issue?  Or perhaps the Xeon is underclocking more than the i7? 

Regards,

Anthony

Peter, the other question is if they are WAV files...they take considerably longer to load than FLAC because only one core is used at a time to upsample/convert them.  I just loaded a random 30 tracks of mixed WAV and FLAC into a playlist and the time from when I pressed play to sound coming out was 1m25s.

I am not 100% sure that the crash will not happen with just one album downloading at a time.  It certainly seems safer to do one album at a time, but I do have a memory of one album being downloaded three times...first as part of a larger group of say half a dozen, then by itself (which was still no good), then by itself again (which plays ok).

Thanks for the detailed response Peter.  Excellent service as usual.    I will give galleries a go.

Cheers,

Anthony

Peter, are Galleries quicker to load than the embedded Explorer?  I ask because Explorer can seem quite slow at times but I have never looked at Galleries.  I should though, rather than ask you questions.

Cheers,

Anthony

More observations...

So I tried downloading one album at a time.  Works.

Two at a time.  Works.

Three at a time.  Works.

Four at a time.  Works.

Five at a time.  Works.

Nine at a time. Did not work.  Had an internet outage and the dialogue boxes as described above came up.  Some of the albums had already downloaded but some had not finished.  Checking the track lists for each album against the Tidal data and it seems as thought the tracks that were downloading at the time of the internet outage (and the dialogue boxes that I clicked "Yes" to) are not there, but subsequent tracks in the album are there.

So this morning I decided to gives parallel downloads a try.  I searched Abdullah Ibrahim and started to prepare the last 5 albums in the search results.  Things seemed to be going well until I lost internet and the following dialogue box appears:

Header = "Prepare Streams"

Message = "5: The remote name could not be resolved: '53.audio.tidal.com'  J:\Streaming Temp\Tidal\AbdullahIbrahim - [2014] - The Song Is My Story\Streams\03- Open Door Within.flac  Continue?"

Options = Yes/No


I clicked Yes and then I got four more similar dialogue boxes for other albums say "Timed Out" and "Continue" with Yes/No buttons.  I clicked Yes to each of these and things seem to continue as normal (am watching the files download in their appropriate directories in Explorer - so all is still working).


EDIT:  I lost internet a second time and had a second round of dialogue boxes appear similar to those above which I clicked away "Yes" to continue when the internet came back and shortly after XXHE crashed.

On the 5 albums started the following tracks were downloaded:
  - 1/2/ 4/5/6/7/8/9/10/11 13 (missing 3/12/14/15/16)
  - 1 & 4
  - 3
  - 3
  - 3

I will do some more experiments now that I know for sure that I should be able to prepare multiple albums in parallel and I will take particular notice of what happens and report back.

When I have had only a partial album download I have verified it by using Explorer to check the number and size of the downloaded files in their appropriate directory. 

At one stage I set about half a dozen Bruckner symphonies to download in parallel and none of those managed to download fully.  The first track or two of most of them managed to download but eventually XXHE crashed and I had to start again, one album at a time.

A few times now I have had a track just end in stuttering part way through and I thought it was to do with the other stuttering problems I have that seem to be related to too low SFS but in fact when I looked at the downloaded file it was too small for its duration and it would end in silence (not stuttering) when played in other software.  To solve the problem I just re-prepare (download) the album in XXHE and all is good.

Hi Peter,

I do my Tidal search and right-click the "Prepare Stream" for the first album I want to listen to, then I do the same for a second or third or more albums while the first are downloading. 

Is this what you are asking?


Regards,

Anthony