Custom Sounds.OXP

[milinks]

Ive put a oxp up to oosat2, which is based on the customsounds.plist in the new install, made available by Giles Hard work. It contains speech and audio / music files based on the list in ogg format. It is complete but at this stage is more of a WIP / example of the possibilities that Giles has created. the ogg files are generally 3-5 layered, and this is really where i want feedback. i have altered the volume of some parts of the layers as i have seen fit, but playing it, i havefound that sometimes speech can be drowned out by gunfire Any feedback on this issue, i.e do people favour speech volume over music/audio etc? Or are some of the speech files too long / short. Its kind of a new area, especially for windows users who do not have speech synth etc. the ogg files / speech etc are very easy to do with the software i have, so any ideas etc please feel free ( ever get the feeling you should never ask to be insulted so easily? )

The oxp is approx 1.8 meg, so quite small. I was thinking of tinkering with the music and putting voicecredits and stuff on there, such as "Oolite, the future of space trading by Giles" etc, but didnt want to get told off by Giles or NoSleep.

[aegidian]

Wow!

You've really gone above and beyond with that. It's exactly what I was hoping someone would do with this - and more.

Thank you.

[Wolfwood]

Whoa! That's excellent!

This really enhances the immersion aspect of the game! Job very well done, milinks!

A bug I noticed: when you engage the hyperjump engines and then disengage them (cancelling the jump) the speech still goes on with the countdown etc. Prolly requires an update to the code to fix this.

Also, it would be nice to also have speech effects when you engage (and/or disengage) the witchfuel injectors or at least the in-system jump engines.

And while we are talking about sound effects and music, would it be possible to have some sort of background music while you are docked on a station? I really liked how the docking music fades away slowly when you have docked and I almost expected it to go on and on...

[Rxke]

Kewl!

one little niggle or two so far:

Code: Select all

Refusing to load auto_on.ogg because its uncompressed size is over 4194304 bytes.

And when using 'c' to engage autopilot and then disengage, the voice is really silent...

Also, I'm not sure that's intentional, but sometimes I hear garbled voices in the background of the 'main voice' while exiting the station... In fact, I liked that, sounded just like the busy background of a flight-control centre

[milinks]

Thanks for the feedback, i tried to multi channel the audio so that you had various levels of chatter. Its this volume that i need to sort out, and / or removethe "background chatter" if people dont like it, but i'll try and get the balance better. I'll look at "auto_on", thanks for that Ryke, I'm going to create quite a bit more, but if anyone has any ideas please let me know as i have time on my hands at the moment.

[aegidian]

The best bet would be to use a simple message for auto_on and replace the docking music with your version by sticking it as BlueDanube.ogg in /Music (files from Music are streamed whereas Sounds are preloaded).

[Murgh]

cool stuff

have some comments, but they could have more to do with the current shape of customsounds.plist than the way these sounds are assembled..

it's a little weird to get a 'welcome to Xsystem' when you're leaving a station..

is there a distinction between IDcom lock and target lock? because it's not so nice to always be told you've locked missiles on someone you just want to identify..

re the multi-layer issue, it works on the whole, but sometimes I'm more interested in the lower voice in the background. if there was a vote, I'd put mine in for more simple, single layered messages..

looking forward to its evolution.

[Rxke]

Been playing it a bit more, thru my stereo, full tilt, and it's great!
Agree with Murgh about the 'missile-lock' message, would be better if it was simple "I.D. locked on" or something more neutral

Don't remove the backgroud chatter, it's great!

One thing that I noticed, on full-range speakers: some audio has quite a bit of low frequency rumble in the background, not very noticeable w/ computerspeakers, but a small blemish to the OXP of the year, heehee.

I bet you'll see it top-ranked @ OOSat in no time!


Again: grrrrreat addition!

[JensAyton]

Kewl! 8)

one little niggle or two so far:

Code: Select all

Refusing to load auto_on.ogg because its uncompressed size is over 4194304 bytes.

Time for some technical documentation, I think.

The Mac audio code limits sounds to 2^20 frames, or four megabytes.¹ I think this is a reasonable limit to in-memory sounds, if not a tad high, but I’ll look into switching to streamed audio (i.e., reading from disk and decompressing on the fly) automatically for this case, instead of the current artificial distinction between streamed music and in-memory sound effects. This won’t be entirely trivial since sound effects have to support being played at more than one offset at once. Also, sound effects are mixed down to mono, while music is played binaurally (discarding any extra channels). The mono mixdown is because of the intention to use positional audio, but this probably won’t happen for some time, so I’ll look into allowing stereo sound effects; we’d probably want to keep that ability for non-localised sounds even when positional audio is implemented.

I don’t know what the technical limitations of the other builds, which use SDL_mixer, are.

Incidentally, the Mac implementation currently supports Ogg Vorbis only. I want to add FLAC support at some point, but it’s not a priority unless someone asks for it. SDL_mixer doesn’t seem to support FLAC at the moment, so it’d require writing two importers.

¹In this context, a frame is one sample for each channel. Since sound effects are mixed down to mono, that’s one sample. The native audio format for OS X is 32-bit floating point, which is four bytes per sample, so 4 × 2^20 = 4 megs.

[JensAyton]

Incidentally, 2^20 frames is 23.8 seconds at 44.100 kHz, or 47.6 seconds at 22.050 kHz. This should be sufficient for most sound effects. :-)

[Lucidor]

What would be the advantage of using flac instead of ogg?

[JensAyton]

FLAC is a lossless format, which is traditionally preferred over lossy formats. It is also much faster to decompress, to the extent that it makes sense to store it compressed in memory and decompress it at the latest possible stage of playback, thus significantly reducing memory use (and cache hits).

[Rxke]

Incidentally, 2^20 frames is 23.8 seconds at 44.100 kHz, or 47.6 seconds at 22.050 kHz. This should be sufficient for most sound effects.

Weird how I knew that by heart. Used to be the max memorysize of my sampler, back in the days fitting 4Mb cost almost as much as the sampler itself....

[dajt]

Having just tried the OXP (cool) and been playing Freelancer for a few days (stupid, but some good visual and sound effects), I'm thinking there are a couple of other things we can do:

1. Add a playSound action to scripts. This would be a start of allowing random background chatter and other ambient sounds. The script timer and other checks like whether docked, or how close to the system station could be used to select groups of sounds.

2. Add some more events like when an AI ship is on final approach to or launches from a station, or a convoy leader is about to make a witchspace jump.

[dontfearthetuba]

Nice work, this definitely adds another level to an already fantastic game.

The only thing I can find being an issue is when my lasers overheat. I get a spoken message along the lines of "missiles overheated". Seems a bit odd since I'm not using missiles and they can't overheat anyways.

Otherwise, kudos, on a very worthy addition!

Currently running the v 1.65 beta for Mac OS X.

DFtT