I have trouble choosing the right audio playback technology. There's a ton of technologies to use on the iPhone, it's so confusing.
What I need to do is this:
start playing short sounds ranging between 0.1 and 2 seconds
high quality playback, no crackle (I heard some of the iPhone audio playback technologies do a crackle sound on start or end, which is bad!)
ability to start playback of a sound, while there's already another one playing right now (two, three or more sounds at the same time)
What would you suggest here, and why? Thanks :-)
There are basically four options for playing audio on the iPhone:
Audio Toolbox. Easy, but only good for playing sound effects in applications (sample code).
Audio Queue Services. Very powerful, can do anything. C API, pretty messy to work with. Callbacks, buckets, pain.
AVAudioPlayer. About the easiest option. Can play compressed audio, with a simple wrapper you can easily play multiple instances of the same sample at once (non-compressed audio only, as there is only one HW audio decoder). Starting to play a sound with AVAudioPlayer seems to lag about 20 ms, could be a problem.
OpenAL. Decent compromise between complexity and features. Sounds do not lag, you can play multiple sounds just fine, but you have to do a lot of the work yourself. I’ve written a sound engine called Finch that can help you.
Don’t know much about cracking, never experienced it. I think there were some issues with playing seamless compressed loops with AVAudioPlayer, can be overcome by saving the loop without compression.
Related
I've been searching for a while and can't come to a good conclusion.
I am trying to create an app that can "record" beats that a user makes on a 4x4 button array. Each button has a sound tied to it and after they hit record, I want to mix the audio that gets played and save it to a file so they can listen to it and play over it later.
What makes this even trickier is that there will be a metronome playing and I do not want to mix the metronome sound into the audio that is getting saved.
From what I have found, the only way to go is Audio Units for these features, but I am reluctant to since it seems a little overkill and somewhat complicated to learn. Can Audio Toolbox make this any easier?
Thanks!
In generally, using a AudioToolBox easily implements.
more information, see below sample code. it's a lot of help.
MixerHost
I am developing one application. In that I want to record the sounds and I want to play that recorded sound file. I know the frameworks for doing this. But how to develop programmatically by using that frameworks?
You can refer to this link:
I have implemented this code in one of my apps and it works completely fine.
How do I record audio on iPhone with AVAudioRecorder?
For Playing the sound you have option to use AVAudioRecorder.
Hope this helps.
The best way to do it - and I am talking from painful experience here - is with the RemoteIO audio unit. You can also do it with AudioQueue, but it has a higher latency, and the queue type approach becomes very problematic.
So, I think that they are really different tools for different jobs. Note that you won't play a sound file as such. You will play the contents of a buffer held in memory. As long as the buffer is not too large, this should not be an issue.
So, going with RemoteIO, you will find this blog and tutorial very useful. It includes code samples.
Using RemoteIO audio unit By MICHAEL TYSON
I'm trying to find the best way to play a seamless loop of audio, that the user can switch out for another at the shortest possible notice, with a decent number (30-150) of very short loops being available. Will OpenAL be sufficient for this, or do I need to delve into Audio Units? The Apple Documentation says that for real-time feedback like an instrument, Audio Units is the right choice.
I just want to get the community's opinion on this, and any links and sample projects would be greatly appreciated.
You can use AVAudioPlayer to seamlessly loop a compressed audio file (numberOfLoops = -1). I suggest using IMA4-encoded CAF files, as these are rumored to benefit from hardware decompression (saving CPU cycles for other things).
To keep file size down, you can lower the bit rate (try 96 kbps) and/or use mono.
Note that AVAudioPlayer does not allow you to change the tempo or frequency of playback.
this probably doesn't really answer your question, but have you ever looked at Finch?
Just looking at the source might provide some pointers.
Johannes
If you need to be able to switch to another audio sample with no playback delay, you'll need to use OpenAL. AVAudioPlayer has a delay before it starts playing.
You can minimize that delay by calling prepareToPlay, but it won't always eliminate the delay completely. As well, if you have 30 to 150 samples that the user selects for playback, you won't know beforehand which samples need to be preloaded.
Here's a rundown of the pros and cons between OpenAL and AVAudioPlayer: http://kstenerud.github.com/ObjectAL-for-iPhone/documentation/index.html#choosing_sec
I ended up using Cocos2D's audio library for this, and it was far more performant than I'd expected. AVAudioPlayer ended up being a good bit lower-level than I'd needed.
What is the difference between OpenAL and AVAudioPlayer on the iPhone? It seems that both can be used for playing/recording audio. When would you use either? What advantages/features does each provide?
Thanks!
-MT
OpenAL is an advanced audio playback engine. You have to get a lot more "down-and-dirty" with it than you do with AVAudio player.
OpenAL is often used in games. OpenAL provides 3D sound positioning which is where things like distance related parameters come into play. It is a C based API while AVAudio is Objective C. If you don't know medium to advanced C programing than you are going to struggle with OpenAL programming.
AVAudioPlayer is best if you just need basic playback and recording. If you need more control than OpenAL or Audio Queues is the way to go (though Audio Queues are also a C-based API). Many people seem to prefer OpenAL over Audio Queues as it's a cross platform library and works similar to OpenGL which game programers already are quite familiar with.
In most caes outside of gaming or advanced audio synchronization situations, AVAudio is the way to go and works great. Even in games, I'm often seeing people use a combination of OpenAL (for sound effects) and AVAudio (for music playback).
OpenAL provides more control if you want to change the audio parameters like distance related, patter-based playback. AVAudioPlayer is a pretty basic one, but you can achieve similar with AudioQueues and AudioUnit using low level api's. In a nutshell, OpenAL does few things with snap of finger but then again it needs what you are looking at? Hope this helps.
I'm making an opengl game for iPhone. And I'm about to start adding sound effects to the app. I wonder what's the best framework for this purpose.
Is AV foundation my best option? Any others I'm missing, like Open AL perhaps?
General strength/weakness summary of iPhone sound APIs from a game perspective:
AVFoundation: plays long compressed files. No low-level access, high latency. Good for theme song or background music. Bad for short-lived effects.
System sounds: plays short (think 0-5 sec) sounds. Must be PCM or IMA4 in .aif, .wav, or .caf. Fire-and-forget (can't stop it once it starts). C-based API. Appropriate for short sound effects (taps, clicks, bangs, crashes)
OpenAL: 3D spatialized audio. API resembles OpenGL and is a natural accompaniment to it. Easy to mix multiple sources. Audio needs to be PCM (probably loaded by Core Audio's "Audio File Services"). Pretty significant low-level access. Potentially very low latency.
Audio Queue: stream playback from a source you provide (reading from file, from network, software synthesis, etc.). C-based. Can be fairly low-latency. Not really ideal for a lot of game tasks: background music is better suited to AVFoundation, shorter sounds to system sounds, and mixing to OpenAL or Audio Units. Can record from mic.
Audio Units: lowest public level of Core Audio. Extremely low latency (< 30 ms). C, and hard-core C at that. Everything must be PCM. Multi-channel mixer unit lets you mix sources. Can record.
Be sure you set up your audio session appropriately, meaning you declare a category that indicates how you interact with the rest of audio on the device (allow/disallow iPod playback in the background, honor/ignore ring/silent switch, etc.). AV Foundation has the Obj-C version of this, and Core Audio has somewhat more powerful equivalents.
Kowalski is another game oriented sound engine that runs on the iPhone/iPad (and OSX and Windows).
You might want to check out Finch, an OpenAL sound effect engine writter exactly with games in mind.