Jump to content
The Dark Mod Forums
Sign in to follow this  
Deadlove

Grumbler with reverb

Recommended Posts

Contribution!! Not a FM, but sound revisions...

 

 

 

Here's a link to a zip file with the grumbler sounds revisited, there's slight reverb + compression.

 

 

LINK REMOVED UNTIL SOUND SET FIXED

 

 

INSTRUCTIONS:

Download and extract the .ogg files into the tdm_sounds_vocals4/sound/voices/grumbler and it will replace all the current sounds with the new ones. I'd recommend backing up the original .ogg files first, so you can revert back to them. Run the game and it will play the new sounds.

 

I demoed on "Business as usual" because I remembered there is a guard with the grumbler voice right outside the door in the beginning.

 

INFO:

There are A LOT of lines that vary in amplitude, so the reverb density was also a bit different for some odd number of the 363 files. It fluctuated usually between -19dB to -21dB with the very low lines / whispers given -15dB for rounding out. The compression / leveling was added to get all sounds to mostly float somewhere in between -2dB to 0dB, so they will be louder than most other sound effects, as most sound effects are from who knows where and recorded with gain stages all over the map. The reverb was a simple room tone.

 

 

 

I plan to use this setting for all dialog. For outdoor sound effects, etc., I'll journey into some ambient and nonlinear reverb processing, not only for realism, but for distinguishment of sound.

 

 

 

Hopefully I'll be able to knock off another character in the next day or so, to do the entire sound bank will take some time.

 

Any and all feedback in welcome! Please share opinions on what you guys think!

 

 

 

Random, who did the grumbler voice? It sounds like the same guy who did the zombie...

 

 

Any suggestions for the next character to do? I'm trying to pick relatively popular vocal sets...the grumbler was one that I heard a lot. I think I might do the guy who's like "what's somebody doing hiding in here anyway...up to no good I bet" or he says something close to that...

Edited by Deadlove

Share this post


Link to post
Share on other sites

Hmm, gave them a listen and I think they sound a bit better this way, though I'm not completely sure. It's fairly subtle, but I think it does sound a bit more 3d. There was one odd moment where they overlapped vocals and it sounded really weird, but that only happened once. I'd like to hear from someone else.

  • Like 1

Share this post


Link to post
Share on other sites

Thanks for the positive input Springheel.

 

Five other people downloaded the files, hopefully they'll have come time to try it out soon and give their opinion.

 

Even just for my own enjoyment, I find adding the reverb does add to it for me. I'd like to imagine I have it in me to do to all the sounds just for my own pleasure playing. Of course, I'll share all progress with the board, in the form of downloadable files that all can load up, if they wish to.

Share this post


Link to post
Share on other sites

The reverb sounds pretty pleasantly neutral. You've got a good ear for balancing both wetness and decay time. :)

 

I think you could be a bit more gentle with compression. You don't necessarily need to normalize all the barks to around -0 dB. My particular set, "Drunk", lets idle/muttering peak between -10 and -6, whereas shout/combat peaked near -0. Then again, every voice is different and benefits from different approaches to compression. For example, I thought mine sounded a bit better when I more aggressively tamed the peaks, rather than "digging deep".

 

From the technical side, one issue I note: when one bark is interrupted by another bark, the first is instantly muted. When the original recording is relatively dry, this isn't too much of a problem. With the introduction of a couple seconds of reverb, this cutoff sounds awkward. This is why some of us wanted to wait until run-time software-driven EFX effects would get implemented in the engine itself.

  • Like 1

yay seuss crease touss dome in ouss nose tair

Share this post


Link to post
Share on other sites

Thanks Mortem! Glad you find the balance accurate.

 

In regard to the compression, it was added twofold reasons. One was to bring the overall loudness up, and the other was to add the general compressed sound to voices that is common in video games. Voices in games tend not to be as natural sounding as real life, so that was my main focus with the compression. I wasn't too focused on the louder barks than the quieter barks, I was shooting for a basic balance between them so regardless how low the line (muttering and whispering) was or how loud (combat and shouting), it would not be that too far off in volume, even if that is unnatural in the real world.

 

I do think you raise a very good point about overall the sounds reaching closer to 0 dB, as the vocal sets should probably be no louder than -3 dB and louder sound effects like mechanical/machinery would go to closer to 0 dB to give the sounds in the world a more dynamic feel and the player would be more immersed.

 

I agree about the barks being cut off halfway through. For example, if a searching bark is playing and the AI spots the player and moves to a found player bark before the other is complete, it does cut off abruptly jumping to the next bark. Only thing I think to do for that is wait for software environmental FX, as the sounds would still have the environment to carry the bark verb trailing off.

 

All in all, I do think the gentle reverb/compression does add to it, in the meantime before the EFX gets implemented into the code. Since also that the "raw" Thief sounds had reverb/compression on them as well, that was what I'm using as a model.

 

Hoping to get at least one more positive response before I officially decide to make up my mind to set out to do ALL the sounds. If I do it for myself, I have no deadline to meet, but if others enjoy it and think it will benefit the mod, then it would be more motivating to work on it more often.

Share this post


Link to post
Share on other sites
I was shooting for a basic balance between them so regardless how low the line (muttering and whispering) was or how loud (combat and shouting), it would not be that too far off in volume, even if that is unnatural in the real world.

 

What's the benefit of this if it's unnatural? Why wouldn't we want muttering and whispering to be quieter than shouting?

Share this post


Link to post
Share on other sites

According to the book of AAA, benefit would be that no lines go missed by playing by being very low.

 

Just like with music, gaming seems to be on a compression kick where most everything is close to the same amplitude. I did it to make the sounds closer to what gamers are used to hearing and not so dry.

 

As Mortem brought up, the lines for the drunk have 10 dB difference in some places, and I am under impression it would be better to close that gap to around 4 dB. I am NOT flat lining / hard compressing because that sucks all the life out of a given performance, whether film, music, game or what have you.

 

It's not really a right or wrong, but that this is the general sound formula everyone is used to hearing so I used it. Sounds in video games are very unnatural to the real world. One of my past time professors did sound for games and was shot down by his producers because his sound for a machine gun wasn't "real enough" when it was the exact machine gun recorded. Their translation was that even though it needed to sound bigger and badder, because it's a game that needs to be competitive within the market, and not the real world.

 

If all sounds were done, another thing that stood out to me is that missions with custom sounds, conversations, etc. would not have the reverb that is in the main .pk4 files.

Share this post


Link to post
Share on other sites

I'm quite familiar with that practice, and it's done that way to compensate for the fact that people will play their games on crappy-sounding TVs, crappy-sounding laptop speakers, mediocre headphones, all the way up to HiFi surround systems.

 

Bioshock Infinite is probably a good example of this practice. Elizabeth's vocal lines, whether shouting or almost whispering, pretty much have been brought up to the same amplitude. Ditto for Deus Ex:HR. Half Life 2 also compressor/limiter'd the hell out of their vocal lines and Foley SFX. Doom 3's SFX exaggerated it even more. It's common practice, and it's usually very beneficial for action games even down to old-school adventure games.

 

Whoever did Thief 3's audio design followed most of these action game conventions to a T. So whenever you heard an AI shout or bark, it was extremely jarring. Good! But!-- if you heard an AI merely yawn or stretch or mutter or even whisper, it was also just as extremely jarring. :wacko: That's just poor sound design.

 

If you were to ask any mixing engineer, they'll tell you that too much compression is very fatiguing on the ears. And Thief & TDM both require sitting quietly or patiently sneaking. To avoid such ear fatigue, a lot of TDM's elements of sound aren't heavily compressed or normalized, and many of our SFX aren't heavily compressed or layered. (Personally, I wish some of them were run through a fast limiter, just to get a few more dB of "punch" in them). I'll totally agree that you can address that large volume gap in the vocals, but I would do it more gently than AAA action game sound design would demands.

 

As for foleys, your prof's machine gun probably didn't sound "real enough" until he layered the actual machine gun recording with a book slamming, a few other gunshots, a burst of pink noise, a wooden plank 'crack', and an explosion. That's just the way foleys work -- If it's recorded real, then it probably needs something more to make it sound flattering. If it sounds bigger than real, your brain thinks must be real. I've even done that with a kick drum while mixing music: It needed more "punch", so I slammed a book on a wooden crate. :laugh:


yay seuss crease touss dome in ouss nose tair

Share this post


Link to post
Share on other sites
According to the book of AAA, benefit would be that no lines go missed by playing by being very low.

 

That runs counter to the way we do sound in TDM. Some lines are supposed to be missed, potentially, because the AI is not projecting them out to be heard. When people mutter to themselves, they don't project their voice like they do in a conversation, and it would sound very artificial if they did. In fact, I had to have an actor re-record almost their entire vocal set because they started off doing this.

 

 

If all sounds were done, another thing that stood out to me is that missions with custom sounds, conversations, etc. would not have the reverb that is in the main .pk4 files.

 

Hmm, that is a wrench in the works.

Share this post


Link to post
Share on other sites

Some lines are supposed to be missed, potentially, because the AI is not projecting them out to be heard. When people mutter to themselves, they don't project their voice like they do in a conversation, and it would sound very artificial if they did.

The fact that you can clearly hear a builder acolyte muttering some prayers only when very close serves to heighten the intimate danger of sneaking around--and TDM earns more tense and exciting gameplay moments because of it.

  • Like 1

yay seuss crease touss dome in ouss nose tair

Share this post


Link to post
Share on other sites

Thanks for the feedback guys! Glad to hear that with TDM we approach things in a more realistic manner, as opposed to the standard corporate / professional practice. For the next set of sounds, I will definitely keep those things in mind. Finding a balance between the realism and making it sound the best I can for whatever set of speakers any given player is using. I can't imagine getting enjoyment from sounds playing on a laptop without headphones. It's crazy to me how many people are used to sounds this way, or on a smart phone. I prefer dedicated speakers with a larger frequency range response. Also glad to be able to communicate on a technical level with someone like Mortem, as that makes things all the more easy.

 

I also agree about the fast limiting, I had put a compressor/limiter on the master channel for this exact reason, to add that punch.

 

And yes, Mortem, you're exactly right with my professor. He had to add some layers, and then once the sound was larger than life, the producers were very happy with the result. It's amazing the foley work that fools the mind. One such instance with Thief that stands out to me was with Viktoria's vines moving about in the cutscene when she steals Garrett's eye. It sounds extremely pleasing for effect, but is nothing more than leather gloves being pulled up on and twisted around on the hands! Also have heard many a broken celery stalk to simulate bones breaking, among other layers added in.

 

Very cool story about the book slamming for mixing with the kick drum. Two of my most used tools for drums are Drumagog + Sound Replacer. The majority of work I've done, people just don't want their drumsets to sound real. They love the highly compressed & triggered sounds. Also, I work with more metal / rock than anything else, so I've come to learn it's to be expected.

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Sign in to follow this  

×
×
  • Create New...