Page 205 of 299

Re: lekt player

Posted: Fri Sep 26, 2014 10:21 am
by jesuscheung
lekt wrote: ...
MOD2: softness better, vibration better. seems OK, only drum/bass body worse, not enough height.
YES. body is worse.
i think first MOD maybe wrong. clean, but, lost fake or actual details?

Re: lekt player

Posted: Fri Sep 26, 2014 10:25 am
by jesuscheung
nopt man/nopt64 has strong power as needed for Barbra Streisand. also, vibration is kept NOT HARD at all.

2.98x has strong power but not vibrate easily

---
think need to slowly adjust the power

Re: lekt player

Posted: Fri Sep 26, 2014 10:52 am
by lekt
v2.99.4 256 c7 , v2.99.4 256 tck and v2.99.9 256 xx now are good for me.
tck have best vibration/soft, only weight lesser, ie. male vocal not enough bass layer. other things very good.
very hard to beat its vibration.

need more test v2.99.9 xx, new method, almost rewriten all code based on c7 code structure. all variables used numeric data type, not pointer * data type. think can be optimized more.
------------------------------

i think machine code understands only numbers. assembly and C, C++,... made them to pointer/object/class... for easy making code. may not good for audio player.

Re: lekt player

Posted: Fri Sep 26, 2014 10:54 am
by jesuscheung
remove exception is very confusing... different effect for different versions!

for c7, i think it is incorrectly dirty.
dirty makes water.
when dirty is gone, water also goes away, so sound becomes drier.


now, i remove exception section for nopt man, the sound is less power excess. goood water.

----
power must be diffiult to tune. too much -> weighty. not enough -> dry.


---
i say c7 has good power. not enough
nopt man is full power. maybe more than norm.
what you think?

Re: lekt player

Posted: Fri Sep 26, 2014 11:53 am
by jesuscheung
uploaded 2.99.4 256 c7 MOD5

-deleted exception
-deleted relocation
-relocation stripped = true //NEED THIS. OTHERWISE NO SOUND.

all other things untouched

seems original bass + body + water is kept. cleaner/transparent sound.

-----
uploaded 2.99.4 256 c7 MOD6

same as MOD5. different method of deleting section.

prefer MOD5

Re: lekt player

Posted: Fri Sep 26, 2014 9:27 pm
by goon-heaven
lekt wrote:try find and remove some metadata or "quantiumlization" as you called, keep only needed things.

uploaded:
lekt.exe v2.99.9 256 nopt64
lekt.exe v2.99.9 256 nopt man

compiled manually, feel more organic, better flow. i feel good emotion.
Sorry, both exit immediately back to cmd prompt here - without playing any sound.

Re: lekt player

Posted: Sat Sep 27, 2014 2:53 am
by jesuscheung
2.99.4 256 c7 MOD4 is fully modded
MOD5 only deleted sections

actually much like MOD4... bass not kept. still cannot tell fake or real details was gone

here all MOD has similar height/stage/weight/power...

they differ in body. also differ in treble/bass balanace.
MOD4 has highest ease, cleanest sound while evironment is kept.

Re: lekt player

Posted: Sat Sep 27, 2014 4:08 am
by jesuscheung
think c7 MOD4 maybe correct. now i use realtime on c7 MOD4 -> quantity, refinement, dynamic range... all things improve.
realtime doesn't do much on original c7.

let me test more versions....

----
stripping sections works for mqn2.2. same effect

large address = disable broke mqn2.2 -> threw an exception... damn it


---
uploaded exenopgomemcpy MOD -exception
deleted exception section

starting to enjoy this version, extraordinary bass is slighty different to original

Re: lekt player

Posted: Sat Sep 27, 2014 4:30 am
by jesuscheung
in CFF_Explorer, think if rebuild image, rebuild PE will worsen the sound

maybe why stud_PE always has an negative effect

Re: lekt player

Posted: Sat Sep 27, 2014 4:44 am
by lekt
goon-heaven wrote:
lekt wrote:try find and remove some metadata or "quantiumlization" as you called, keep only needed things.

uploaded:
lekt.exe v2.99.9 256 nopt64
lekt.exe v2.99.9 256 nopt man

compiled manually, feel more organic, better flow. i feel good emotion.
Sorry, both exit immediately back to cmd prompt here - without playing any sound.
v2.99.9 256 nopt and v2.99.9 256 nopt ehsc works for you?

all nopt xx use same code, code changed for nopt64.

ok, will review, seems your memory addresses of wasapi functions doesn't align 8, need use +8 more safe than ^8.