Well the footy coverage on TV has finished for the afternoon so it's back to testing

As Peter suggested, I replicated my 0.9y-2(a) folder and replaced the Engine#3 from the one out of 0.9y-3, and did some more A/Bing.

My conclusion .... it's the engine itself which is making the discernable difference in SQ. The GUI is simply a means to an end, and if it is having any influence, then it is minor at best and I can't pick any differences between them.

So for now, I will continue with my initial hybrid which consisted of the bug reduced GUI from 0.9y-3 combined with the Engine#3 from 0.9y-2(a).

Not sure where to progress from here. It's probably got Peter scratching his head though .. LOL

Cheers all,

Russ

OK ... further thoughts here ... I was trying to get the best of both worlds ... updated interface combined with what I believe to be the best engine so far, which is in fact different to what Peter suggested, which was going back to the 'broken' interface ( B.B. King issues etc ), and using the latest engine.

I'll try that shortly. Sorry for confusion!

Cheers,

Russ

The contents of zips for 0.9y-2 ( with 2a exe added shortly after ) and 0.9y-3 appear to be identical apart from the actual GUI ( XXHighEnd.exe ) and the engine#3, so it should not matter which way the files are flipped .. you will end up with the same result. I double checked that before I created the hybrid.

The more I listen to the earlier engine, the more I enjoy it and would be happy to call it a day at that. It really is fabulous here. Of course it still has the glitches of not starting always and occassionally stopping mid stream, but when it gets going ... damn .. it's just great

I'm experimenting with some changes to my scheme's and priorities to see if I can get it a little more reliable is starting and continued running.

Cheers,

Russ

It's just occurred to me that Engine#3 changes between 0.9y-2(a) and 0.9y-3 would likely incorporate some memory manipulation recoding to cure some bugs we've noticed.

Peter, I assume you keep notation within your source code to keep track of alterations, are you able to cast light on what you might have changed to cause the effect we ( I ) am hearing between engines? I'm convinced I'm not imagining things now that i've been playing my 'hybrid' for a couple of hours now.

Ta,

Russ

Hi Gerard(A),

good to read you had some success.

My Q settings were arrived at by simply playing a couple of test tracks that are very plain in composition, but contain instruments that are readily discerned. I think it helps to keep things simple initially and tweak until you think the instruments are as close to real life as possible. My two main test pieces contain triangles, cowbells, castinets and a guitar mainly, which are articles that should be remembered reasonably well.

I also think you need to give a set of parameters time to 'invade' the subconscious over a period of time. After a number of hours you should know whether it's for the better or not, and I don't think Mr Placebo has any influence here as there is no right or wrong settings , but simply those that give you the closest feeling to reality in your environment and using your own equipment.

As a general rule I've noticed, as have some others, that the lower the Q1 parameter is set, the more precise and distinct an instrument appears to become but may be perceived as too clinical for some ears. Your DAC also needs to run comfortably at these low settings and the lower the buffer size the better ( usually ).

Happy tweaking

Russ

Another dawn ... another delightful day of testing

With a clear and fresh mind, I have just finished another session of comparisons between 0.9y-2a and 0.9y-3 and my conclusions are still the same as last night .. that 0.9y-3 lacked that little sparkle and clarity I have been enjoying of late.

Picking up on the recent post by Peter that it's the Engine that does the SQ work ( something that we perhaps tend to overlook or forget with all these changes of versions ), I created another test folder and plonked 0.9y-3 into it, then COPIED Engine#3 from 0.9y-2(a) into it, then queued up my two standard test pieces of music ... and pressed PLAY.

Well wadda ya know .... the life had returned !

I recall in the past overriding the GUI with updated versions, but can't remember playing LEGO with the engines themselves.

Not sure of the consequences of this action in the wider scheme and whether there is some tight relationship between the engine and GUI within each release, but it's opened my ears to an alternative approach to testing. Hopefully this technique doesn't open up a can of worms .. Peter has enough hair pulling at the moment keeping track of changes across versions without the further complication of us creating hybrids.

Food for thought hopefully

Cheers,

Russ

Hi GerardA,

I noticed your signature indicates you use Q2,3,4,5 settings of all 15's. If they are still current, may I suggest you try 30,30,0,0 and see if the "blanket" is removed.

One thing I discovered early on when the later Q settings were implemented, that if you had the first couple at max ( ie. 30 ) you DO NOT want to move the Q4 & 5 at all. Doing so will suck life and dynamics from the music.

Give that a try, especially with 0.9y-2a

Cheers,

Russ

My initial impressions are that 0.9y-3 has had some life sucked out of the percussion compared to 0.9-2a. Cymbals, triangles, cowbells etc. seem to have lost some of that "metallic ring" I had grown accustomed to since the start of the 'Y' versions.

I have a couple of pieces of music I use as reference that have very simple percussive content, and I've gone back and forth between these last two versions many times this evening, and discounting Mr Placebo, I still prefer the dynamics and excitement that 0.9y-2a brings.

On a more positive note, the gallery issues all seem to resolved ... B.B. King etc. have now joined the rest of the ensemble.

Will continue with more comparisons tomorrow.

Cheers,

Russ

'twas just an idea Peter. If it proves too difficult to implement or you find some logical reasoning to chuck the idea in the reject bin, then feel free to do so .. haha

I've been flipping between 0.9y-2a and 0.9y-3 in the last couple of hours doing comparisons, and I've been trying to figure out an easy way to make sure I always had the right 'pathing' defined within AHK for the version of XXHE I was using. I ended up simply closing and restarting AHK in between version swaps each time. A little tedious but effective.

Cheers,

RGB

Happy to help guys

If I was an astute dotnet programmer I could do something real clever like ... oh ... maybe check for the existence of AutoHotkey running before I started up a particularly fine example of a program that converted 1's and 0's into musical bliss, and if AHK wasn't running , start it up for you. But then I'm not a clever programmer!

I wonder if there is one of those lurking around here somewhere 

Cheers,

Russ

Hi Eric,

If I read between the lines here, it seems you are unaware, or may have forgotten, of the unique function of the 'Startup' folder within 'All Programs' that is displayed when you click on the START button ( bottom LHS ).

Any programs residing in the 'Startup' folder will get loaded automatically when Windows boots.

I will assume you already have a shortcut on your desktop to AutoHotkey.

Navigate to that 'Startup' menu (via the START) button, and highlight it, then click on the right mouse button to bring up a context menu, and choose the 'Open' option. A small explorer window should open showing you the contents of that 'Startup' folder.

If you now drag (using the RIGHT mouse button) the AutoHotkey shortcut (on your desktop), and DROP it into the right panel of the explorer window you just opened up, you will be presented with an option to 'Copy here'.

If you do that, then go and recheck the contents of the 'Startup' folder, you should now see the presence of AutoHotkey there.

Next time you boot Windows, Autohotkey should automatically load for you.

You might also want to look here ... http://www.phasure.com/index.php?topic=753.msg6655#msg6655

In a similar manner you could also autoload XXHE this way should you wish.

Cheers,

Russ

Hi Leon,

try putting the executable for 0.9y-2a over the top of your current install.

See here .... http://www.phasure.com/index.php?topic=888.msg7026#msg7026

I also noticed you have Volume set at -120db in the screen capture ??

Cheers,

Russ

Here's a few  more examples from my M1 harddisc.

The following artists names DO cause problems:

B.B. King
G.E.N.E
J.J. Cale
Jay B. Jay
K. D. Lang

while the following do NOT cause issues.

B.B. KIng & Diane Schuur
B.B. King , Big Mama Thorton & Muddy Waters
Booker T. & The MG's
M.J. Harris & Bill Laswell
Sammy Davis Jr. & Laurindo Almeida

Sort of reinforces the theory that if there is a dot in the folder name, it needs at least 6 characters (including space) following the dot for things to work as desired.

Cheers,

Russ

Hi Peter,

I have attached a few screen captures to try and demonstrate the coverart display bug I am seeing.

Coverart_Bug_01: shows the folder structure of my music. I don't currently have them split my genre's, but simply stored as artist\albumname format under a high level folder of M1 , M2 , etc, which corresponds to the disc volume they reside on.

Coverart_Bug_02: shows the initial layout when I click on the upper level M2 entry, with the first entry in the middle panel highlighted to show the other artwork in the righthand panel. Note that "A.C. Marias" , "A.R. Kane" and "a silver mt. zion" are shown.

Coverart_Bug_03: from the last picture I clicked on the entry on the left panel, and, as expected, the contents of the "1000 Homo DJ's" were displayed correctly.

Coverart_Bug_04: following the last picture I then clicked on "A.C. Marias" entry on left panel. Coverart for album appeared in middle panel.

Coverart_Bug_05: clicked on the album entry on left panel under A.C. Marias ( it expanded to show track listed as expected ), however the coverart was now missing in middle panel, although the path text was showing. The artwork showing in right panel was a leftover from previous 'click'.

Coverart_Bug_06: re-clicked back on the main A.C. Marias entry on the left panel and now there is no coverart showing, which differs from the "Coverart_bug_04" scenario above.

Coverart_Bug_07: starting from the view I got in "Coverart_Bug_02" , I then clicked on "A.R. Kane" and no change occured. I would have expected the contents of the A.R. Kane folder to be have shown now. Note: this is the same scenario if I was highlighting my "B.B. King" entry on my "M1" disc.

Coverart_Bug_08" If I click on the little 'expand' triangle to the left of the A.R. Kane album, and then highlight the "I" album, the coverart shows as expected.

Coverart_Bug_09: Just prior to this screenshot I had highted the "A Certain Ratio" entry on left panel. This pcture shows a null response when highlighting the "a silver mt. zion" entry. In a similar scenario to just above, if I expand the entry using the little triangle, and then highlight the album entry underneath, the covertart appears as expected.


Common theme in this bug is the presence of a dot (.) in the album name, and it maybe just co-incidental, but there are only 4 characters following last dot in the name _where you get NO initial response_. The positioning of these artist names within the folder structure may be having an influence either side of the actual entry.

Good luck in tracking it down Peter! It's likely another quirky m$oft gotcha where the pathing names are being mangled somehow.

Cheers,

Russ

Just realised the 'Played.dat' isn't a hidden file ( no need to panic now Peter ... haha ). I was looking at the XXDat0000.dao also and mixed up the attributes in my 'ol brain.

RGB