Huggybaby
Zebranky food

Offline
Posts: 33
|
I took Paul Reiche's new dialogue and added some effects. The dialogue he provided had lots of digital artifacts which I was mostly able to remove. One of the main problems with my version is that it sounds too clean. 
All three lines are here in sequence: New Mycon Dialogue.ogg (513.72 KB) http://www.mediafire.com/?zddmagrnxu0
I'm interested in your opinions as to how close I got, thanks!
|
|
|
Logged
|
|
|
|
Alvarin
Enlightened
    
Offline
Gender: 
Posts: 795

|
It differs from the original voice pack, sound like it lacks the echo effects or something. As you wrote, too clean. But, anyway it would be much better than nothing.
|
|
|
Logged
|
|
|
|
Cedric6014
Enlightened
    
Offline
Gender: 
Posts: 701

|
Well sounds pretty good to me too. A few other people have tried and made some reasonable efforts. See here: http://forum.uqm.stack.nl/index.php?topic=4166.15
Whether it is good enough to put in is a bit moot. I thik for some reason the core team are reluctant to insert this fix into UQM? Or maybe theyre gonna role it out with 0.7.
They're always pretty quiet on this.
Meep-Eep - speak to us!
|
|
|
Logged
|
|
|
|
|
Huggybaby
Zebranky food

Offline
Posts: 33
|
Actually I don't think I could improve on what VileRancour's done, that's a great fix and it includes more speech than the Paul Reiche stuff.
I listened to VileRancour's samples and tried to match the echo in that, but some of the speech doesn't even have echo. The coordinates part does though. I'll listen to some more samples and see if I can hear what you guys are hearing, and maybe try adding some noise.
|
|
|
Logged
|
|
|
|
Halleck
Enlightened
    
Offline
Gender: 
Posts: 750

Personal Text
|
I haven't been around recently so this is the first I've heard of a re-recording. Could you message me with the url where I can get the unprocessed originals? I'm an undergraduate sound design student with access to professional audio processing tools and could do cleanup and echo/reverb matching.
|
|
« Last Edit: May 04, 2010, 04:12:03 pm by Halleck »
|
Logged
|
|
|
|
|
Halleck
Enlightened
    
Offline
Gender: 
Posts: 750

Personal Text
|
Thank you!
I've had a little free time in which to take a crack at this. Here's my first pass at the "source of juffo-wup" line. It by no means matches perfectly and could use some tweaking but I think it's probably closer than anything else yet submitted.
I have also mixed it into the game dialogue in a similar manner to the files submitted in the bug report (Although the timing is different and therefore will require revised subtitle timing. I have also added chorus tails to the game dialog in order to make it transition smoothly. Here is a 44.1 kHz version of mycon043 with a chorus tail at the end, and here is an 11.025kHz version which ends abruptly as does the original file.
The trick to this one was using a chorus effect on top of the reverb. I couldn't identify this on my own but some nice folks at freesound.org suggested this to me and it works beautifully!
I run it through a fairly complicated processing sequence, but I will do my best to describe it here should anyone want to re-create it in the future: 1. Pitch shift dry dialogue to 87.08% of original pitch (-2.4 semitones) while perserving duration 2. Manually pull down peaks that are too loud with volume envelope 3. Compress with a threshold of -44.1 dB and gain up 29.8 dB 4. Apply a low-fi ('bit crush") effect to simulate 14700 hZ, 9 -bit audio - adaptive encoding with 63% antialiasing (this is probably not "accurate" to the original recording but it sounded right) 5. Send a copy of this new "dry" signal to an effects chain containing: 6. A 7 second decay reverb with 1ms delay 7. A chorus effect with an oscillator at 0.82 Hz 8. Another lo-fi "bit crusher" at 11025 Hz, 66% adaptive antialiasing, 10bit sample size (once again just going off sound with my plugin)
These "wet' and "dry" channels are both sent to a -3 dB fader. I also send one syllable of the tail of each original dialogue to the "wet" effects channel and fade its volume up to create a tail.
It would be much easier if you could just look at my Pro Tools session which I would be happy to post, but this is proprietary software using proprietary plugins and i don't think would be of much use to anyone other than professional audio people.
|
|
« Last Edit: May 07, 2010, 09:20:02 pm by Halleck »
|
Logged
|
|
|
|
Alvarin
Enlightened
    
Offline
Gender: 
Posts: 795

|
Very close! Very good job! on the negative, I think the inserted part main voice is bit "lower" pitched compared to original.Other than that, very VERY close.
|
|
|
Logged
|
|
|
|
Halleck
Enlightened
    
Offline
Gender: 
Posts: 750

Personal Text
|
Thanks! I was thinking that as well... I was playing around with different pitching options and none of them seemed to be "right".
Thankfully this is all non-destructive and easily repeatable so I can tweak the pitch and/or duration of the clip.
|
|
|
Logged
|
|
|
|
Angelfish
Enlightened
    
Offline
Posts: 568

|
Don't try to make it perfect. You have some artistic freedom in creating these voices, especially in the case of the mycon where it's essentially multiple personas speaking through one body, it's quite okay to have some differences .
|
|
|
Logged
|
|
|
|
Alvarin
Enlightened
    
Offline
Gender: 
Posts: 795

|
Angelfish is absolutely correct about the miltiple-personalities and there is one more thing to concider - anyone played the game already will know where to look and will probably not pay attention to this minute difference. Any first-time players will most likely miss this as well, because it will sound intentional or something. But I do like to have my comments productive (usually) therefore commented the difference. In absolutely no way it was disrespect or mockery or anything like that.
|
|
|
Logged
|
|
|
|
Halleck
Enlightened
    
Offline
Gender: 
Posts: 750

Personal Text
|
Angelfish: Good point. It's been a long time since I've played sc2 but I seem to recall split-personalities. Do the voices change too? Can you point me to some examples?
Alvarin: I posted a draft here in order to get feedback so yours is appreciated. 
Here is my version 2 - again not perfect but the pitch should be closer, this time it's 95.93% of the original pitch which is a very slight reduction (-0.72 semitones). Perhaps it is a bit too high this time- pitch is tough to get right because I cannot tweak it in real time unlike every other effect I'm using.
Scratch that, Here's version 3 at 91.5%/-1.54 semi. Maybe a bit low but I think it's closer.
I also added a short reverb to the "dry" channel (consider this step 3.5) and fiddled with the dry channel bit-crusher.
Not having to match would be great but I need to hear some examples of not matching in the game to see what my range could be... I don't want these 3 lines to be the only ones in the game that sound different!
|
|
« Last Edit: May 07, 2010, 11:23:08 pm by Halleck »
|
Logged
|
|
|
|
Huggybaby
Zebranky food

Offline
Posts: 33
|
That's cool, your first pitch shift was -2.4 semitones, and I settled on -2.5, so we heard the same thing there.
|
|
|
Logged
|
|
|
|
Alvarin
Enlightened
    
Offline
Gender: 
Posts: 795

|
The original parts begin with some backgroung effects, but the inserted part begin silent. Can you add some speech to the beginning for the effects and then remove leaving the actual speech only? It is kind of hard to explain, hope you understand what I mean.
|
|
|
Logged
|
|
|
|