Previous Page | Next Page

by haspor at 3:17 AM EDT on March 30, 2014
Ritzierhades, there is no generic way to know if a song has loop-point/fade.

Fade-length is a bit nasty to create/implement. This can be done later tho. Would it help if i put back the default songlength selector with override? Since i implemented the generic looping, this time it could actually work as expected.

For the record, LazyUSF overrides everything, it's priority #1 #2 and #3. People wanted USF player, now they are gonna get it ( probably :) ) , and it comes with a price :(
by RitzierHades63 at 9:45 PM EDT on March 30, 2014
With a price? That sounds...creepy...

Anyway, fade length isn't really what's bothering me right now (five seconds works). What's bothering me is that all my timed files are timed to end right at the loop point of the song. With these files, the current fade option makes the song fade out right before the end of the loop. I'd like the fade option to also add five seconds to the end of the song's timer, so it can fade out during that. That would make my timed files fade out right after the loop point, which is the way most OSTs have it.

And yeah, I understand that USF is the priority. If this isn't a quick and easy add, by all means save it for later. :)

edited 12:38 AM EDT March 31, 2014
by haspor at 1:27 AM EDT on March 31, 2014
price as shitload of work :( NOT related to money, no worries :)

Even tho, if someone wants to donate me Nexus 5, I won't mind :)

After LazyUSF, I'll see what I can do about this.



by Level99 at 4:46 PM EDT on April 2, 2014
Awesome to see such dedication to the USF project. Keep up the great work!

And I'm still waiting for a donation link so I can buy you a six-pack or twelve.
by kode54 at 8:06 PM EDT on April 14, 2014
Regarding the recent addition of Adplug to the development source, I would recommend borrowing the emuopl interface from my foo_input_adplug component, based on dbopl, because that emulator is much more accurate at some things, even compared to the so-called "most accurate OPL3 emulation" included with the SDL port of AdLib Tracker 2. For instance, compare the rhythm sounds from Operation: Body Count's nubc.imf, or Bloogton Mfg., Incorporated(28).imf from the Commander Keen series rip.

(Also, that one from Keen requires 1:1 sampling. My version of dbopl and the two emulators in my modified Adplug set the timing ratio to 1.0 if it's already near enough, such as if you ask for OPL2 with 3579545 and 49716Hz, or OPL3 with 3579545*4 and 49716Hz. Then you can use my fairly speedy and decent quality sinc resampler to downsample that to 44100Hz for output. Or to save processing power, you can pre-construct a series of phase tables. Ooh, I can do that too. I'll give it a try shortly.
by haspor at 7:15 AM EDT on April 15, 2014
All right, the adplug plugin is almost done, i just need to add the database thingy and the resampler, i wasn't aware of things that might happen if the database is not included.

edited 7:16 AM EDT April 15, 2014
by kode54 at 8:21 PM EDT on April 15, 2014
The database includes speed overrides for some IMF files, among other things.

If you use the dbopl thing, it should be stereo from the get go, as long as you initialize it with the stereo parameter set to true. Otherwise, it will always be mono.

The resampler is the best way to have sample accurate emulation of some effects, as simulating it at 49716Hz with a 1:1 clock ratio and downsampling produces exactly the right amount of aliasing, whereas running the emulator at a different rate can result in unwanted aliasing that breaks some effects or songs.
by haspor at 1:06 AM EDT on April 16, 2014
Done with database thingy. I will make that lanczos resampler use ARM NEON. Android can't play higher than 48000hz so the resampler is needed.
by kode54 at 3:26 AM EDT on April 16, 2014
You have updated my source, right? It's using resampler.c now, which is a sinc resampler using a Nuttall 3 term window, not Lanczos.
by haspor at 6:58 AM EDT on April 16, 2014
I was planning to update the lanczos_resampler.c found from foobar_input_adplug, is it correct?

Previous Page | Next Page
Go to Page 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Search this thread

Show all threads

Reply to this thread:

User Name Tags:

bold: [b]bold[/b]
italics: [i]italics[/i]
emphasis: [em]emphasis[/em]
underline: [u]underline[/u]
small: [small]small[/small]
Link: [url=http://www.google.com]Link[/url]

[img=https://www.hcs64.com/images/mm1.png]
Password
Subject
Message

HCS Forum Index
Halley's Comet Software
forum source