[Gambas-user] OpenAL (gb.openal) example code

[Kevin Fishburne]

On 10/20/2015 05:45 PM, Benoît Minisini wrote:
> Le 20/10/2015 07:40, Kevin Fishburne a écrit :
>> On 09/15/2015 01:49 AM, Kevin Fishburne wrote:
>>> Currently using Alure.BufferDataFromFile to load sound files into
>>> OpenAL buffers I'm unable to get the buffer properties using the
>>> Al.GetBufferi function. I don't know if this is because I'm using
>>> Alure versus the regular Al method, but this code should work and
>>> doesn't (see attached project):
>>>
>>>     ' Load audio data from .wav file into audio buffer.
>>>     Print "Load buffer data from wav file:" & Gb.CrLf & " ";
>>>     Print Alure.BufferDataFromFile("./test.ogg", AudioBuffer[0])
>>>
>>>     ' Display audio data properties.
>>>     Print "Audio frequency in Hz:" & Gb.CrLf & "  ";
>>>     Print Al.GetBufferi(AudioBuffer[0], Al.FREQUENCY)
>>>     Print "Audio buffer bit depth:" & Gb.CrLf & "  ";
>>>     Print Al.GetBufferi(AudioBuffer[0], Al.BITS)
>>>     Print "Audio channels:" & Gb.CrLf & "  ";
>>>     Print Al.GetBufferi(AudioBuffer[0], Al.CHANNELS)
>>>     Print "Audio data size in bytes:" & Gb.CrLf & "  ";
>>>     Print Al.GetBufferi(AudioBuffer[0], Al.SIZE)
>>>
>>> It returns zeros instead of the proper values. In particular I need
>>> the Size property returned, as this can be used to calculate an offset
>>> from which to start playing a sample once attached to a source. This
>>> is necessary because multiple sources created and played
>>> simultaneously using the same buffer will result in the waveforms
>>> being combined at the same sample point, which creates a metallic,
>>> flange-like effect. If I can find a buffer's size I can offset the
>>> position at which a source begins to play randomly between zero and
>>> the buffer's size like this:
>>>
>>> Al.Sourcei(Ship[ShipIndex].AudioSourceVelocity[0], Al.BYTE_OFFSET,
>>> Rnd(0, Al.GetBufferi(AudioBufferVelocity[0], Al.SIZE)))
>>>
>>> For now I'm going to have to use something like Rnd(0, 100000) for the
>>> offset and simply hope it doesn't exceed the buffer's size, which
>>> silently returns an error and plays the source at zero (no offset).
>> Having not gotten a response I'll try again. There's now a second
>> critical use case where I need to retrieve the properties of an OpenAL
>> buffer. According to multiple Internet posts such as this one:
>>
>> http://stackoverflow.com/questions/7978912/how-to-get-length-duration-of-a-source-with-single-buffer-in-openal
>>
>> I'm using the correct method with the possible exception of relying upon
>> Alure to load the buffer (which you'd think would be irrelevant since a
>> buffer is a buffer is a buffer and all). Please see the OpenAL example
>> in the GAMBAS software farm to test the failing code. I'm using Mint
>> 17.2 amd64 and GAMBAS from the stable PPA. There (as usual) is a
>> possibility I'm just doing something really stupid, or there could be a
>> legitimate problem with the OpenAL implementation in Alure or GAMBAS.
> Sorry, I forgot to look at your problems...
>
>> On an unrelated note, I finally watched Benoît's interview. A few
>> things... First Benoît's English was good enough that I could understand
>> about 80%+ of what he was saying. Second, the two guys doing the
>> interview didn't seem to know their assholes from their elbows,
>> especially the older Mac guy. Benoît, you were outstanding and your
>> passion for both GAMBAS and BASIC really shone through; nice work and it
>> was awesome seeing and hearing you talk about your (and our) passion.
>> Also around the 23:00 mark you gave me a shout-out by name. I nearly
>> shit myself. :)
>>
> ...maybe this is the reason why I mentioned your name: I knew
> unconsciously that I had something to do relative to you. :-)
>

Haha, perhaps. Maybe from now on I'll start submitting formal bug 
reports. That way they'll stay there waiting, forever, looking at you 
with their sad, compound eyes until you can provide them a merciful coup 
de grâce.

-- 
Kevin Fishburne
Eight Virtues
www: http://sales.eightvirtues.com
e-mail: sales at ...1887...
phone: (770) 853-6271