foo_dsp_continuator, crossfading plugin Options

[mobyduck]

Here's a direct port of the plugin, consider it beta:
http://stephan.kochen.nl/proj/foobar2000.html#continuator

Thanks a bunch, but I'm unable to modify Fade-out curve, Fade-out length and Fade target: I see the options but disabled. Is this an issue with Win2000?

Alessandro

[Silversight]

Thanks a bunch, but I'm unable to modify Fade-out curve, Fade-out length and Fade target: I see the options but disabled. Is this an issue with Win2000?

Alessandro

No, your Win2k has nothing to do with it. Uncheck "Disable track overlapping on seamless transition", and the options become available.

@G-Lite: Wow, that was fast!

This post has been edited by Silversight: Apr 11 2006, 15:29

[mobyduck]

No, your Win2k has nothing to do with it. Uncheck "Disable track overlapping on seamless transition", and the options become available.

Oops, yet another stupid post...

Thanks for your help.

Alessandro

[Silversight]

When I use foobar2000 to DJ (playing background music, et al), continuator keeps the music going seamlessly, and the directsound 2 crossfading allows me to switch tracks and stuff in a way that sounds very smooth and natural, as if I were actually crossfading in a mixer.

Is it somehow possible to implement this "manual track skip crossfade" in Continuator? That would make it kind of an all-in-one fading device (Germans have a funny name for those which roughly translates to "egg-laying wool-milk-sow").

This post has been edited by Silversight: Apr 11 2006, 16:46

[oshah]

Is anyone interested in my update of the plugin? I've removed all level 4 warnings from the code (it may not actually fix anything, but it could help with troubleshooting that STL assert).

This post has been edited by oshah: Apr 11 2006, 16:54

[manudevil]

Cool ! YOur Continuator seems to work well !
Thanks !

[Supacon]

I think this warrants a thread-split, or at least a new thread. This one's getting long, and offtopic.

[Supacon]

I'm testing out the new port of continuator. Mostly it works well but there are a few issues. One of them is that when it pre-loads the next track at the specified buffer length, it starts counting time played on the next track, even if it hasn't started playing yet. (especially noticeable with the "power" checkbox off)

I still haven't tested if this causes other problems, like the track prematurely ending. (I've just been fast forwarding to right before the buffer length, which is set to 15s before the track end)

The other problem is that it doesn't always seem to buffer the amount of time specified in the preferences. Perhaps I'm just not aware of how it's supposed to work, but it usually seems like it pre-loads the next track quite a bit after it should.

Other than those problems, things seem to work very well, and it sounds great.

This post has been edited by Supacon: Apr 12 2006, 00:36

[Lyx]

While Continuator is now very powerful regarding features, i think it has become a bit difficult to use.

For fb2k 0.8, there was a limitation regarding seamless-trackchange detection and fadeout-length. Is this still the case for 0.9?


Anways, here is my proposal on how it should work in an optimal world (please tell me if its technically possible):


PREFERENCES:
General options(always available): power, do nothing on seamless trackchange, Adjust Tracklength to CD-Blocksize

Overlap-Mode:
Look-ahead time (in ms)
Threshold (in dB)
Fade-Out (in ms - plus curve-style and checkbox to only apply if audio at the end)
Fade-In (in ms - plus curve-style)

Crossfade-Mode:
Look-ahead time (in ms)
Fade-Out (in ms - plus curve-style)
Fade-In (in ms - plus curve-style)

Automatically calculated:
Crossfade-Length (is always either the length of "Fade-Out" or "Fade-In" depending on which is larger)
Buffer length=fadeout/fadein(depending on which one is larger) plus Look-ahead time
(Yes, i am aware that this results in a minor feature-loss, but i think the simplification is worth it)


CONTINUATOR-TAG (backwards-compatible):
overlap (uses default preference-settings)
overlap,threshold_in_dB (uses default fade-lengths)
overlap,threshold_in_dB,fadelen_in_ms_for_both (sets fadelength for both fadeout and fadein - both will have the same value)
overlap,threshold_in_dB,fadeout_in_ms,fadein_in_ms

crossfade (uses default preference-settings)
crossfade,fadelen_in_ms_for_both (sets fadelength for both fadeout and fadein - both will have the same value)
crossfade,fadeout_in_ms,fadein_in_ms

gapkill (alias for "overlap,60")

off (disable continuator for this trackchange)



DROPPED FEATURES:
empty buffer at overlappoint (the user should not need to mess with this)
crossfade-time (minor feature-loss, but simplifies prefs and allows to make continuator-tag more useful)


UNSURE:
Should fade-in or fade-out be prioritized(order) in the preferences and continuator-tag? (needs to be decided now, because prefs should have the same order as continuator-tag because of consistency)


Ideas, questions, proposals, etc. welcome

- Lyx

This post has been edited by Lyx: Apr 12 2006, 01:36

[DeepDose]

Thx G-Lite, seems to work well, I have to turn the value up slightly high to get a effect, and maybe someone can tell me this but this plugin doesn't allow the track to fade out on hitting the "stop" button does it? That was directsound 2.5 wasent it? thx anyways.

[Supacon]

i've uploaded continuator version 0.3.0, which adds the following new features:...

...additional delay dependant on speed of volume drop, to give tracks with sudden ends some 'air', before the next track starts (currently not configurable)

Hmm... I'm not sure that I always like this behavior. It would be nice to have a checkbox or something to toggle this behavior.

I know this is old... due to the lack of built-in documentation for this plugin, I was reading old posts. That's another enhancement that could be made... have some context help, or at least tooltips to describe what some of the features do. Perhaps a button linking to an HTML page would be very helpful as well.

Another idea... that's too bad that fadeout has to be disabled if seamless track transition detection is on. Is there not some way of predetermining that before the song even starts? Or would that be terribly complex and well outside of the domain of this component? I guess you might be able to use the tag instead, if this were to happen.

Uhm... could you use the continuator tag to identify songs as seamless, so that continuator doesn't affect playback even if "seamless detection" is turned off?

[Chaser]

Could someone please share his settings?

[Stuart60611]

All that this plugin is missing, in my opinion, is the ability to beat match during crossfade.

[gr0under]

Please make support auto disable crossfading for CUE audio files.

sorry my bad english

[Supacon]

Please make support auto disable crossfading for CUE audio files.

An interesting feature proposition, but personally I don't only use cuesheets for seamless CDs (like DJ mix CDs, Live albums etc). I often rip a CD as one file with cuesheets just so I can preserve the exact layout of the CD, but the actual tracks aren't seamless.

It might be a handy switch to have in the options though, for many.

[Supacon]

I found another bug in continuator that occurs when playing very tiny files.

I was using foobar2000 to listen to some small wav files (under 10KB) with a sampling rate of 22 KHz in mono. As soon as I attempt to play them, foobar crashes when using continuator.

I thought I should bring this to the attention of developers and such...

The two files causing the problem are here:
http://fullspec.ca/misc/sounds/

[oshah]

I'm quite aware of this error. It's basically caused by improper use of STL iterators, as mentioned in this previous post by G-Lite:

Technical note: taking care of the foobar related stuff was easy, but I encountered some STL errors related to vector iterators. There might be some more I haven't bumped into yet, but the plugin is working in both modes from what I've tested so far. Visual C++ 2005 has gotten alot more strict regarding STL.

I think the fix requires the iterators needing to be reinitialized after modification, but I haven't had the time to check or test yet.

This post has been edited by oshah: Apr 15 2006, 21:14

[yourtallness]

Could the continuator be set up to emulate the behaviour
of the skip_silence plugin?

I mean, to have it start the next track when the current
sound level falls under a selected threshold, while also
stopping the current track instead of fading out?

This post has been edited by yourtallness: Apr 16 2006, 14:27

[NogginJ]

Could the continuator be set up to emulate the behaviour
of the skip_silence plugin?

that would be pretty cool.

ive never used this plugin before, is there a way to set it up to just skip silence between tracks...so for instance it could play files seamlessly? i would think that would be the point of the plugin, but i cant find an option that sets any number below 10ms, but i might be reading the options wrong. that and the feature listed above would really make this a nice all in one silence manipulation plugin. well i guess itd have to have ability to add silence too but now im talkin crazy

[Melomane]

G-Lite , can you do terminate port of continuator to foobar 0.9?
regards, melomane

[Squeller]

G-Lite , can you do terminate port of continuator to foobar 0.9?

G-Lite offers 0.9 ports only.

Question to anybody: Is the option "Disable track overlapping on seamless transition" working?

I tested on lots of gapless classical stuff and, it's fading as hell. The option does not seem to work.

This post has been edited by Squeller: May 6 2006, 16:24

[Melomane]

Squeller, port of continuator is not complete!
Look here
and it is possible that is the reason so that "Disable track overlapping on seamless transition" not working.

This post has been edited by Melomane: May 7 2006, 00:16

[Squeller]

Squeller, port of continuator is not complete!
Look here
and it is possible that is the reason so that "Disable track overlapping on seamless transition" not working.

Correct. I did some testing on the same tracks/ with the same settings with the 0.9 and the latest 0.83 compile: They behave differently on the issue. 0.83 works fine, 0.9 doesn't.

[G-Lite]

Squeller, port of continuator is not complete!
Look here
and it is possible that is the reason so that "Disable track overlapping on seamless transition" not working.

Correct. I did some testing on the same tracks/ with the same settings with the 0.9 and the latest 0.83 compile: They behave differently on the issue. 0.83 works fine, 0.9 doesn't.

I didn't change any of the dsp code, and the plugin should be equal in features with the 0.83. I don't know alot about dsp though, not sure I can fix that.

[Melomane]

g-lite, without make any change to dsp code, can you make "preset interface " work?
same as you do for v-level.

it is possible this is the reason for different behavoir enter 0.83 and 0.9.

regards, melomane