After reviewing the contents of the dumped tapes and talking to various people at Revolution, we've learned there are 2 sets of recordings.
The 1st set, which Revolution recorded and supervised (these are the tapes we have and are, as far as dialog is concerned, mostly complete - but some of the voice-acting is quite poor and often different from the material used in the final game), and the 2nd set consisting of (a lot of) re-recorded takes done by Virgin (because they didn't like the quality of the original recordings and quite possibly the fact that it sounded too 'british' to appeal to e.g. the US audience) - these are the recordings we're currently missing and are looking for.
This project unfortunately can't proceed without them, and is thus on hold until we manage to somehow get ahold of these recordings (they are being looked for, however the fact that Revolution are in the middle of moving offices and Virgin UK is no longer in business doesn't help us!).
Well you could always just wait and see if the missing recordings do turn up, there's no rush to get it done within a certain timeframe AFAIK. It's just a pain that another gremlin has crept into the process.
I don't know too much about audio algorythms, but are there any enhancing filters like there are for gfx? (like build a table which sample is spoken by which actor, then apply one set of filters to one speaker, and for example another set on Joey) - I guess this could be either done on the fly or external and ScummVM loads those files then...
I don't remember anyone discussing something like that yet
clem wrote:like build a table which sample is spoken by which actor, then apply one set of filters to one speaker, and for example another set on Joey
What would be the point in doing that on the fly? Joey still will and is supposed to sound the same throughout the game.
On the fly effects would make sense if you need different versions throughout the game, for instance the same text spoken in different rooms could use different kinds of reverb. Some more advanced game engines can do that.
I meant more in the sense of keeping joey the same, and enhancing (interpolating? like I said I don't know what is possible with audio) the human speakers so the audio quality seems better
different filters depending on rooms is a neat idea, I can almost hear the echo in the huge exhaust shaft behind the cathedral already
clem wrote:I meant more in the sense of keeping joey the same, and enhancing (interpolating? like I said I don't know what is possible with audio) the human speakers so the audio quality seems better
Sure, you can interpolate sounds. This is usually called "resampling" or when your target rate is an integer factor of the source, it's also called "oversampling". This is done in order to get rid of mirror frequencies (and alias frequencies during recording), though it can't restore information which weren't present in the first place. A 11kHz sample will never sound as good as a 44kHz sample, providing they are recorded properly.
different filters depending on rooms is a neat idea, I can almost hear the echo in the huge exhaust shaft behind the cathedral already
Though the parameters for that must be set in the scrips, which would require to change the game data, and also enhance the engine in order to do that.
cpt. wesker wrote:do the DAT tapes have the missing parts?
Did you bother to read the thread?
We're waiting/hoping for the missing recordings to be found (either on DAT, an old harddisk, tape backup, or whatever), as we can't do what we originally intended without them.
Been following the BASS Audio Restoration for a while, had a bit of time on my hands and processed Joey's voice using LogicPro. It's in the same order as the original mp3 further back in the thread. Just threw it together, stuck a reverb that would be similar to something like what the first room in the game would sound like (hangar/warehouse-ish). What do u guys think?