Guides Audiobundles: assetbundles for sounds

[atani]

atani submitted a new resource:

Audiobundles: assetbundles for sounds - How to create audiobundles

View attachment 509309​

Audio assetbundles

I'll start with this topic because this is the one you might find strange, why turn audio files into assetbundles? This advice comes from MacGruber's comments when I was perusing the forum.
If you use audio files in a assetbundle, loading is almost immediate and - I haven't compared yet, sorry - might be much softer on available RAM...

Read more about this resource...

 

[jpxrz]

Thanks for your tutorial! Now I finally know how to pack my dialogs to an assetbunble.
But I have a further question: I don't understand why the file size became much lager after compressed in Audio Options (Step 3B).
For example as below screen shot: I selected the Decompress On Load and PCM format, the Information box shows the Ratio is 74.06% which means the Imported Size will be from 8.3MB to 6.1MB

However, after I clicked the 'Apply' button, the file size enlarged from 8.3MB to 36.1MB, of which the Ratio is 437.29% (as below screenshot)

 

[atani]

I don't know much about the topic so I'm going to say my current understanding:

• PCM is uncompressed, so perhaps everything is explained just on that point. "Decompress on load" is probably meaningless when chosing PCM
• The first info box shows the current situation if you just pack as when imported and don't apply any changes (vorbis is default I think)
• The second box is the result of changing your original file to PCM, which is uncompressed, and possibly you start with a compressed format like mp3 or ogg

On the guide there's a link to this page explaining a lot more:

Unity Audio Import Optimisation - getting more BAM for your RAM

A guide to using Unity's audio import settings to help improve game performance.

www.gamedeveloper.com

 

[jpxrz]

I don't know much about the topic so I'm going to say my current understanding:

• PCM is uncompressed, so perhaps everything is explained just on that point. "Decompress on load" is probably meaningless when chosing PCM
• The first info box shows the current situation if you just pack as when imported and don't apply any changes (vorbis is default I think)
• The second box is the result of changing your original file to PCM, which is uncompressed, and possibly you start with a compressed format like mp3 or ogg

On the guide there's a link to this page explaining a lot more:

Unity Audio Import Optimisation - getting more BAM for your RAM

A guide to using Unity's audio import settings to help improve game performance.

www.gamedeveloper.com

Thank you for your kindly help. I've read the article and been aware of what impact those setting mean, I just curious why did the file size enlagred times original.
Anyway, I will try what actual impact of those different settings in practice.
Thank you again for your toturial, it did solve my trouble.

 

[atani]

It's because you chose PCM.
Honestly, for speech lines using Vorbis will keep the size much smaller and the CPU cost and audio quality difference will be negligible, I would use that.

 

[BossTanaka]

Just a quick question. I'm currently trying to get your quickspeech preset to work. Do I need this for using your Audio bundles or is this tool only necessary if I want to create my own audio / assett bundles?

 

[atani]

TLDR:
the audio assetbundles (audiobundles) can be used independently from quickspeech stuff by using the SoundfromAB logicbrick.

Detailed:
There's several things working in combination:

1. VAR with audio assetbundles (1A) and empty atom presets (quickspeech presets) (1B)
2. scene quickspeech atom setup

For 1, the audiobundles (1A) are just audiobundles. Still in 1, the quickspeech presets (1B) is an empty atom preset that links the audiobundle with some plugins added to an empty atom in my scenes called "quickspeech". This preset works in tandem with 2.
Item 2 is also a empty atom in my scenes that has plugins and triggers that connects 1 with the person and what's happening in the scene.

Essentially, if you want to use 1A on their own you can, they're just assetbundles with audio files (see TLDR on top). If you want to have a scene structure that uses the audiobundles for looping, automated audio in a scene, that is what the quickspeech stuff is for.

 

[BossTanaka]

TLDR:
the audio assetbundles (audiobundles) can be used independently from quickspeech stuff by using the SoundfromAB logicbrick.

Detailed:
There's several things working in combination:

1. VAR with audio assetbundles (1A) and empty atom presets (quickspeech presets) (1B)
2. scene quickspeech atom setup

For 1, the audiobundles (1A) are just audiobundles. Still in 1, the quickspeech presets (1B) is an empty atom preset that links the audiobundle with some plugins added to an empty atom in my scenes called "quickspeech". This preset works in tandem with 2.
Item 2 is also a empty atom in my scenes that has plugins and triggers that connects 1 with the person and what's happening in the scene.

Essentially, if you want to use 1A on their own you can, they're just assetbundles with audio files (see TLDR on top). If you want to have a scene structure that uses the audiobundles for looping, automated audio in a scene, that is what the quickspeech stuff is for.

Ok, I guess a clear picture is slowly growing. ... I was (and still I'am a bit) confused on what to use. I read your guide about quickspeech https://hub.virtamate.com/resources/quickspeech-quickly-load-and-switch-random-audio-speech.30482/ here.


So as I understood I have to basically do these steps if I want to add behavioural speech to a scene.

Requirements to be able to use the presets and my audiobundles:

• Female person that speak must be named "Female"
• Female person needs a logicbrick "SoundFromAB" plugin in the fourth position (#3) of the plugins list

Steps to include quickspeech on a new scene:

1. Name the female person as "Female"
2. Add a logicbrick "SoundFromAB" plugin in the fourth position (#3) of "Female"
3. Add 2 empty atoms and name them verbatim as "Controlquickspeech" and "quickspeech"
4. Using the VAR in downloads, load the corresponding preset for the atoms above, starting with the atom "quickspeech"

That basically comes down to what you described with the 1 option you wrote above, right?

But I also have to load (obviously) quickspeech as described in step 3:

• Add 2 empty atoms and name them verbatim as "Controlquickspeech" and "quickspeech""

So I would have to always do step 2 as you described as:

• scene quickspeech atom setup

I'm still alittle confused.

 

[atani]

Ok, I guess a clear picture is slowly growing. ... I was (and still I'am a bit) confused on what to use. I read your guide about quickspeech https://hub.virtamate.com/resources/quickspeech-quickly-load-and-switch-random-audio-speech.30482/ here.


So as I understood I have to basically do these steps if I want to add behavioural speech to a scene.

Requirements to be able to use the presets and my audiobundles:

• Female person that speak must be named "Female"
• Female person needs a logicbrick "SoundFromAB" plugin in the fourth position (#3) of the plugins list

Steps to include quickspeech on a new scene:

1. Name the female person as "Female"
2. Add a logicbrick "SoundFromAB" plugin in the fourth position (#3) of "Female"
3. Add 2 empty atoms and name them verbatim as "Controlquickspeech" and "quickspeech"
4. Using the VAR in downloads, load the corresponding preset for the atoms above, starting with the atom "quickspeech"

That basically comes down to what you described with the 1 option you wrote above, right?

But I also have to load (obviously) quickspeech as described in step 3:

• Add 2 empty atoms and name them verbatim as "Controlquickspeech" and "quickspeech""

So I would have to always do step 2 as you described as:

• scene quickspeech atom setup

I'm still alittle confused.

This is about the quickspeech so I'll answer in the quickspeech guide instead.