I’ve tested this on 2 lines of dialogue but haven’t tested it on 3.
Although I have a DialogueAudio_ID I never used it - could be useful but isn’t integral in this setup.
In the NPC Character BP, NPC_ID is 1 and Conversation_ID is 1. Also, go to Class Settings > Interfaces > Add BPI_Interact (or whatever you’ve called it) to access the Interact Event in the Event Graph.
There should be 18 images here… The image order’s a bit messed up. The top one with the mannequin in it shows you what you’ll ned to make int he Content Browser. Then scroll down to the bottom and go up.
I’m thinking of making a tutorial to show this tbh, if Ryan’s ok with me using his as a base.
Well I got all the dialogue setup and tied with quest system , I am just missing the ability to access both answer sounds as clicked. Is there a part of what you show that I can use to access the sounds without redoing the whole thing? thanks
I’m going to apologise in advance because I’m about to ask you a lot of questions but this is just so I can better understand what’s going on in your setup and how we can fix this
On the first page of this thread, you have the Sound Wave type added - what’s updating this?
In your screenshot on this page, you have the Get Audio node - how is that being updated? It looks like it’s separate from the Structure and that might be why it’s only playing one sound or only playing once - how is this Get Audio node getting updated with a new sound? (If it was an Array it could have multiple values, but if it’s not it’ll just have one value. If nothing’s coming through to update the Get Audio node elsewhere in your BP when you flip to the second line of dialogue, this could be the problem.)
How are you calling your Structure data from within the BP? Using the Get Data Table Row node will call all of the columned data on a specified row and might be the best solution.
How do you have your structures and tables set up? I can see 4 different Structures on your page 1 screenshots - are they all unique or are they controlled by a master Structure? Just trying to get an idea of how big a task you’re dealing with
I think the simplest solution is to store the audio within the Structure itself by adding a type Sound Cue and assigning cues to each line of dialogue in the table. That would use your existing setup.
Alternatively, depending on how much work is involved for you, if you had an independent Structure containing all the dialogue, text, character names, locations… whatever you need to use for each quest/area/character or whatever works for you, then you can use the Get Data Table Row node in a BP to call all of the data stored in that row of the Structure. That would include the corresponding Sound Cue, which you can hook up to a Play Sound 2D node (or Play Sound At Location, if you need to have it coming from the NPC).
Well that " get " node goes up to a segment, or I should say the “answers” array on the segment goes down to that " get " node.(the arrow points to the line going down to the “get” node that’s circled in the second pic.).But the “answers array” is for the answer text , not sound.There is a dialogue sound variable on that upper segment as well , but that is for the NPC speaking, and so there is only one sound needed for him. From there in " Event answer clicked" I just tied in the answer struct there. In the answer struct I tried an “answer index” (giving answer 1 an index of 1 and answer 2 an index of 2) , but I couldn’t get that to work. Also I tried adding two answer sounds on the answer struct , putting the first answer on “answer sound” , and the second answer on "answer sound 2 ", and using the switch on enum to try and play those separately, but that didn’t work either so far. Also I tried to use a couple Booleans on the answer struct as is shown on there (Answer 0? and Answer 1?.
Have you tried having separate rows for each Answer?
I know you don’t want to change too much of your current setup, but why not test this and see if it works?
I’m still not clear on how you’re getting the table data and I think you might be treating the entire table as one repository for everything you need? If you set up your table with Quest ID, Answer ID, Sound Cue field etc then you can use it for every conversation in your game. You could have a look at the quest system used in Skyrim; the modding tools let you see how the developers arranged quests and dialogue.
I think your current setup is trying to call data on one row?
That sounds amazing @Valkyrie_Sound, I would eat that up for sure. I’m a novice at using UE4 itself for sound design, and know absolutely nothing about implementing Wwise into it, so that would be super helpful!
I am having an issue with VOIP volume levels with some microphones. My friend’s mics sound fine in discord but they have to scream into their mics ingame for it to register at all. Do you know if I can create a slider that increases VOIPtalker volume input gain? I’m using VOIPTalker with steam subsystem
Hey @Resolve - the link below suggests changing the Set Mic Threshold node value to -1. I haven’t used the VOIP Talker myself, but they seem to be describing the same issue you’re having. Hope this fixes it!
So i do have a request for some insight or a tutorial: How do we go about implementing proper sounds for automatic weapons in Unreal, without using FMOD or WWISE. By proper i mean we should be able to divide gun sounds into three parts: First shot, looping part (in case weapon goes into full auto) and tail sound. RPM should be adjustable by parameters. Gun sounds should be framerate independent, meaning that the gun sound speed is not allowed to speed up / down during firing as this ruins everything. It should be possible to use the sound setup so that it works for guns that can switch between semi auto / burst / full auto. So if the trigger is pulled very shortly then only one shot should be heard, on burst the respective number of gunshots should be heard and on full auto a fullauto sound should be heard for as long as the trigger keeps being pulled. Ideally, to save on performance the looping part should indeed use a looping sound and not a single shot sound which is looped through BP / Sound Cue. I searched the internet up and down on this and no one apparently has a satisfying answer to this. So if you could let us know on how to do that it would be a singularity
The Ballistics FX pack by Cubit Studios has something similar to what you’re after, though it uses a single sound. It runs from mouse-clicks and not a Tick so isn’t frame-dependent. No single fire or burst modes though (you could add them), and the sound plays in full rather than being split into head and tail.
Tbh, this is something I’ve been meaning to look into more as well Single shot and burst fire are easy because you can set up the Sound Cue to only play one sound (using the Switch node in the Cue) or use the Concatenator node in the Sound Cue to play several sounds in a row.
The tricky part is full auto. If you want to use discrete head/tail assets I’ve found you can end up with clipped audio or some artefacts - but it is over a year since I looked at this and I know more now. You could use the Get Sound > Duration (Sound Base Reference) which will get the longest length of the Sound Cue - but this won’t be effective if you have several sounds in the Cue that aren’t exactly the same length.
Epic’s Shooter demo uses a looping sound for the weapon but I can’t see any of the logic for the weapon audio - not sure if it was written in C++ but the BPs are empty.
I’ll look into this though! If I can figure out a solution I’ll post a video up next week
All right, thx for getting back to me. Quartz sounds cool, however i won’t be able to dig into it in the near future as it would be yet another system i’d have to learn before getting onwards. In case you come up with a cool solution for the gun problem i’d be happy to hear about it!
No worries, and sorry! The standard audio engine nodes aren’t loading in the tail of the audio at the right point in time every time, creating the occasional gap. I’m sure there’s a way to do this in UE4… If you’d prefer to stick just with UE4 and not use middleware it might be worthwhile thinking about using the whole sound, spawning a sound cue for each bullet.
If I remember rightly from what I’ve read, this was the approach in Counter-Strike originally and to this day - they fade out or cull the previous sound and play the new one for each bullet.
Have a read about - it can be useful to see how other sound designers working on other engines handled their problems
In the meantime I’ll come back if I find a UE4-only solution for this - and you do the same please :'D
@Valkyrie_Sound I will do for sure m8, and thx for looking into that for me anyways. One other thing: I just now realized that you do have some cool tutorials up there and what freaked me out was to see that you have some tuts to recreate the sound mood of Alien Isolation. Just like you, i think this game has one of the - if not the - best sound design ever in a computer game (and it’s so **** awesome in so many other ways…).
Also my initial question about the gun sounds originates from my current personal project, which is a VR game based on the universe of the second movie Aliens. Even though Aliens has an entirely different mood and style than Alien, i also love the second part of the franchise. After finishing the whole gripping, weapon operation and shooting logic for the PulseRifle in VR, i came to the audio problem we are talking about here. Sound wise the PulseRifle is an especially tricky gun to do, since it does not really have a single shot mode, nor does it really have full auto mode (while the specs still say that is a full auto weapon). You more or less only ever hear short bursts in the movie. After analyzing and reversing some of the ACM game sound design i found, that it does not implement real full auto sounds for the PulseRifle as well. If you go full auto with that weapon they simply play some concatenated random burst sounds as long as u hold the trigger. So i decided to go for the following route: If the weapon is in burst more, the gun fires bursts and i play a respective burst sound. If the weapon is in full auto mode and the player taps the trigger i also play a burst sound and the weapon fires 4 consecutive bullets (as defined in the specs for the PulseRifle). If the wepaon is in full auto mode and the player holds the trigger, i release full auto bullets while playing concatenated burst sounds like in ACM. I tried to create a real full auto loop for the PulseRifle using some flanger and pitch modifications on original PulseRifle sounds, however my skills in any DAW are limited and it never sounded right. Anyways, as soon as i got the PulseRifle sounds close to being OK i’ll send you a snippet in case ur interested.
And since proper spatialized sound makes VR games even more immersive i am very keen to get my sound design right and i will definetly look into your stuff on AI for that.
*Alien: Isolation *is such a cool game - hellishly scary, mind haha, but super cool - and the sound design is just fantastic. It’s like being in the first movie. Aliens is awesome too though - it’s like Terminator 2, imo, one of the few franchises where the second film was as good as, and in some ways better than, the original.
Interesting solution there as well. If understand it, is the 4-shot burst the minimum number of shots for your rifle? Are you still separating the shots into head and tail?
How are you wanting your effects to change the sound in the DAW?
Indeed, James Cameron seems to understand how to make sequels Unfortunately he didn’t do Predator 2 which is my other favourite franchise next to Aliens and Terminator. Every Predator sequel sucks compared to the original imho. Really do not understand how you can mess something as cool as Predator up so badly. Not even Shane Black managed to do it right and he even played Hawkins in the first movie. Also i think i will never be able to like a Predator movie without Arnold. Him and the Predator…coolest antagonists in history of mankind Anyways, i always get carried away when talking about my most favourite movies of all time.
So yeah, my PulseRifle only has two fire modes: Burst or Full Auto. The minimum number of shots i implemented is 4, since it’s the number of shots that are written down in the (fictional) specs of the weapon for burst mode. Since as far as i know no one ever heard a PulseRifle fire a single round and since i can not even imagine how a single shot would sound like from that weapon, i decided to go with bursts instead of single fire. A tap in full auto mode also releases a burst, bullet count wise as well as sound wise. And yes, i am separating into heads and tails for both burst and full auto. However, like i mentioned before, the full auto sound is a loop of a concatenation of bursts. This is subpar imho, but when i realized that even the guys from ACM (who had full access to the original movie sound recordings) were not able to implement a real full auto sound i gave up my hope on doing it myself.
However, if i were capable in DAW-stuff, i’d try to go down the following route:
I do have heads and tails already. What i am missing is the body of a full auto sound.
That said i do have a monotone full auto PulseRifle sound, which is however missing the signature pitch and flanger effects of the original sound, you know the ups and downs in tone, almost as if the weapon was screaming. So i would try to create several versions from that looping sound, each version with a different pattern of pitch/flanger/whateverisneededtogettherightsignature. Let’s say 4-8 versions of the full auto loop. One might go constantly up in tone, one constantly down, one might start going up and then down again, one doing vice versa, … Of course the modifications would not be allowed to change the tempo / RPS of the original full auto loop.
So in the end i’d have 4-8 varying full auto body pieces that i could use randomly in full auto fire of the weapon. Hope that makes sense anyhow, english is not my native language
