[x3d-public] Sound v4 > channel nodes > node design comparison v4 vs Doug proposed; ChannelMerger
GPU Group
gpugroup at gmail.com
Mon Mar 6 05:05:50 PST 2023
I take that back, proposed also semi-bakes destinations into connections
and I don't have a tested code for un-connecting at this time.
-Doug
On Mon, Mar 6, 2023 at 5:00 AM GPU Group <gpugroup at gmail.com> wrote:
> Mind you v4 design only half-passes the SF routing threshold:
> Selections/sources. Destinations are baked in scenegraph.
> If the routing threshold was changed to MF, then proposed would have full
> runtime control over sources and destinations.
> One way to save proposed, is to promise utility nodes for SF to MF in v4.1.
> -Doug
>
> On Sun, Mar 5, 2023 at 3:11 PM GPU Group <gpugroup at gmail.com> wrote:
>
>> I don;t have a good answer for changing individual elements of an MFInt32
>> field.
>> I guess that use case kills the proposed method.
>> -Doug
>> more.. there was another browser I used about 15 years ago, and it had
>> some utility nodes where you route an MF field to the input, and it would
>> have 10 SF outputs, and another one going the other way - 10 SF inputs, and
>> routed and MF.
>> Seemed limited use with just 10. But would work for a 6 channel case.
>> Something more general like the IntegerSequencer, which takes 2 routings,
>> one for the index (next, previous, or index) and one for the value, but for
>> this MFInt32 case the input value would be SInt32. Then routing from the
>> MFInt32 field in Merger to the utility function, and routing back when
>> changed.
>> But that would take some prose to explain the use case even if there was
>> a utility node like that.
>>
>>
>>
>> On Sun, Mar 5, 2023 at 2:13 PM GPU Group <gpugroup at gmail.com> wrote:
>>
>>> Don,
>>> Thanks for feedback.
>>> Tuples needed. And web3d doesn't have tuple fields.
>>> If web3d had MFInt32_3tuple field type that would be great in this
>>> use-case.
>>> In this case the tuple needed is (destinationChannel, sourceChannel,
>>> audioStream)
>>> Next best I can think of 3 MFInt32 fields --and some prose to explain
>>> they should have the same number of entries.
>>> It's not in w3c because its another mechanism to convert procedural code
>>> to declarative.
>>> -Doug
>>> more..
>>> I showed the v4 design works under challenging scenarios/use-cases, but
>>> is awkward in some scenarios --requiring prose to explain. I couldn't
>>> understand how to use the nodes as the prose was written. Maybe more
>>> detailed prose would help.
>>> "The Merger.children field represents a list of destinationChannels ..."
>>> That's when I tried the 2-tuple. The prose for it would look like
>>> "The Merger.children field represents a list of sourceChannels"
>>> The 2-tuple didn't work with scenarios d) and e).
>>> The 3-tuple prose would look like this:
>>> "The Merger.children field represents a list of audio streams"
>>> Decoupling the children field from both destinationChannel and
>>> sourceChannel had benefits:
>>> - no need to DEF/USE the same Splitter (or non-splitter) audio stream to
>>> fill channel slots, if you have one splitter and/or audio stream, you can
>>> repeat the indexStream index
>>> - can do all the other permutations of merging and swapping channels
>>> with indexSource and indexDestination
>>>
>>> PROSE COMPARISON TEST?
>>> Perhaps a good test of design is which is easier to explain for all use
>>> cases
>>>
>>>
>>> On Sun, Mar 5, 2023 at 12:57 PM Brutzman, Donald (Don) (CIV) <
>>> brutzman at nps.edu> wrote:
>>>
>>>> Thanks for continued scrutiny and experimentation efforts Doug.
>>>>
>>>>
>>>>
>>>> Primary point from last Friday is that we are trying to align with Web
>>>> Audio API parameters closely and correctly.
>>>>
>>>>
>>>>
>>>> Design point: MFInt32/MFFloat arrays with distinct elements in the
>>>> array are quite difficult for authors to manipulate without scripting, and
>>>> even then tend to be quite error prone. If actually needed, it is much
>>>> preferred to define each parameter as an independent field. Looking over
>>>> what you’ve shared here, for example:
>>>>
>>>>
>>>>
>>>> - SFInt32 [in out] indexStream 0
>>>> - SFInt32 [in out] indexSource 0
>>>> - SFInt32 [in out] indexDestination 0
>>>> - SFInt32 [in out] channelCount 0
>>>>
>>>>
>>>>
>>>> Reference
>>>>
>>>> -
>>>> https://www.web3d.org/specifications/X3Dv4Draft/ISO-IEC19775-1v4-DIS/Part01/components/sound.html#ChannelMerger
>>>>
>>>>
>>>>
>>>> However am not seeing a match between what you’re suggesting here and
>>>> what W3C Audio API says, please advise.
>>>>
>>>>
>>>>
>>>> - https://www.w3.org/TR/webaudio/#channelmergernode
>>>>
>>>>
>>>>
>>>> *1.14. The **ChannelMergerNode
>>>> <https://www.w3.org/TR/webaudio/#channelmergernode>** Interface*
>>>>
>>>> The ChannelMergerNode
>>>> <https://www.w3.org/TR/webaudio/#channelmergernode> is for use in more
>>>> advanced applications and would often be used in conjunction with
>>>> ChannelSplitterNode
>>>> <https://www.w3.org/TR/webaudio/#channelsplitternode>.
>>>>
>>>> *Property *
>>>>
>>>> *Value *
>>>>
>>>> *Notes *
>>>>
>>>> numberOfInputs
>>>> <https://www.w3.org/TR/webaudio/#dom-audionode-numberofinputs>
>>>>
>>>> see notes
>>>>
>>>> Defaults to 6, but is determined by ChannelMergerOptions
>>>> <https://www.w3.org/TR/webaudio/#dictdef-channelmergeroptions>,
>>>> numberOfInputs
>>>> <https://www.w3.org/TR/webaudio/#dom-channelmergeroptions-numberofinputs>
>>>> or the value specified by createChannelMerger
>>>> <https://www.w3.org/TR/webaudio/#dom-baseaudiocontext-createchannelmerger>.
>>>>
>>>>
>>>> numberOfOutputs
>>>> <https://www.w3.org/TR/webaudio/#dom-audionode-numberofoutputs>
>>>>
>>>> 1
>>>>
>>>> channelCount
>>>> <https://www.w3.org/TR/webaudio/#dom-audionode-channelcount>
>>>>
>>>> 1
>>>>
>>>> Has channelCount constraints
>>>> <https://www.w3.org/TR/webaudio/#audionode-channelcount-constraints>
>>>>
>>>> channelCountMode
>>>> <https://www.w3.org/TR/webaudio/#dom-audionode-channelcountmode>
>>>>
>>>> "explicit
>>>> <https://www.w3.org/TR/webaudio/#dom-channelcountmode-explicit>"
>>>>
>>>> Has channelCountMode constraints
>>>> <https://www.w3.org/TR/webaudio/#audionode-channelcountmode-constraints>
>>>>
>>>> channelInterpretation
>>>> <https://www.w3.org/TR/webaudio/#dom-audionode-channelinterpretation>
>>>>
>>>> "speakers
>>>> <https://www.w3.org/TR/webaudio/#dom-channelinterpretation-speakers>"
>>>>
>>>> tail-time <https://www.w3.org/TR/webaudio/#tail-time>
>>>>
>>>> No
>>>>
>>>> This interface represents an AudioNode
>>>> <https://www.w3.org/TR/webaudio/#audionode> for combining channels
>>>> from multiple audio streams into a single audio stream. It has a variable
>>>> number of inputs (defaulting to 6), but not all of them need be connected.
>>>> There is a single output whose audio stream has a number of channels equal
>>>> to the number of inputs when any of the inputs is actively processing
>>>> <https://www.w3.org/TR/webaudio/#actively-processing>. If none of the
>>>> inputs are actively processing
>>>> <https://www.w3.org/TR/webaudio/#actively-processing>, then output is
>>>> a single channel of silence. […]
>>>>
>>>>
>>>>
>>>> v/r Don
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> *From:* x3d-public <x3d-public-bounces at web3d.org> *On Behalf Of *GPU
>>>> Group
>>>> *Sent:* Sunday, March 5, 2023 5:51 AM
>>>> *To:* X3D Graphics public mailing list <x3d-public at web3d.org>
>>>> *Subject:* Re: [x3d-public] Sound v4 > channel nodes > node design
>>>> comparison v4 vs Doug proposed
>>>>
>>>>
>>>>
>>>> correction to b)
>>>>
>>>> proposed
>>>>
>>>> <ChannelMerger channelCount='6'
>>>>
>>>> indexStream='0,1'
>>>>
>>>>
>>>>
>>>>
>>>>
>>>> On Sun, Mar 5, 2023 at 6:44 AM GPU Group <gpugroup at gmail.com> wrote:
>>>>
>>>> Here's an even better proposed design using a 3-tuple, and showing more
>>>> tested scenarios
>>>>
>>>> CHANNEL NODE DESIGN COMPARISON
>>>>
>>>> v4: channel nodes as written in v4 Spec Draft Sound Component
>>>>
>>>> - merger.children is a sequential list of output/destination channels
>>>>
>>>> proposed: add 3 fields to ChannelMerger
>>>>
>>>> - MFInt32 indexStream, indexSource, indexDestination
>>>>
>>>> - and eliminate ChannelSelector node.
>>>>
>>>> - Merger.children is a list of audio streams.
>>>>
>>>> I have both v4 and proposed implemented in freewrl: if indexStream,
>>>> Source or Destination is empty, thunks to v4
>>>>
>>>> I tested both methods on 5 scenarios and they both work as follows:
>>>>
>>>> a) swapping 2 channels
>>>>
>>>> v4
>>>>
>>>> <ChannelMerger channelCount='6' >
>>>>
>>>> <ChannelSelector channelSelection='1'>
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> <ChannelSelector channelSelection='0'>
>>>>
>>>> <ChannelSplitter USE='Channel_Splitter'/>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> proposed
>>>>
>>>> <ChannelMerger channelCount='6'
>>>>
>>>> indexStream='0,0'
>>>>
>>>> indexSource='0,1'
>>>>
>>>> indexDestination='1,0'
>>>>
>>>> >
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> b) merging 2 channels into 1
>>>>
>>>> v4
>>>>
>>>> <ChannelMerger channelCount='6' >
>>>>
>>>> <Gain>
>>>>
>>>> <ChannelSelector channelSelection='1'>
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> <ChannelSelector channelSelection='0'>
>>>>
>>>> <ChannelSplitter USE='Channel_Splitter'/>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> </Gain>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> proposed
>>>>
>>>> <ChannelMerger channelCount='6'
>>>>
>>>> indexStream='0,0'
>>>>
>>>> indexSource='0,1'
>>>>
>>>> indexDestination='0,0'
>>>>
>>>> >
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> c) merging 2 channels into 2nd merger channel (skipping first output
>>>> channel)
>>>>
>>>> v4
>>>>
>>>> <ChannelMerger channelCount='6' >
>>>>
>>>> <Gain/>
>>>>
>>>> <Gain>
>>>>
>>>> <ChannelSelector channelSelection='1'>
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> <ChannelSelector channelSelection='0'>
>>>>
>>>> <ChannelSplitter USE='Channel_Splitter'/>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> </Gain>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> proposed:
>>>>
>>>> <ChannelMerger channelCount='6'
>>>>
>>>> indexStream='0,0'
>>>>
>>>> indexSource='0,1'
>>>>
>>>> indexDestination='1,1'
>>>>
>>>> >
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> </ChannelMerger>
>>>>
>>>>
>>>>
>>>> d) assign different audio streams to first and 2nd channnels (play beat
>>>> in left, cnote in right)
>>>>
>>>> v4
>>>>
>>>> <ChannelMerger channelCount='2' >
>>>>
>>>> <Gain gain='.5'><AudioClip url='"sound/beat.mp3"'/></Gain>
>>>>
>>>> <Gain gain='.5'><AudioClip
>>>> url='"sound/cnote.wav"'/></Gain>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> proposed:
>>>>
>>>> <ChannelMerger channelCount='2'
>>>>
>>>> indexSteam='0,0'
>>>>
>>>> indexSourc='-1,-1'
>>>>
>>>> indexDestination='0,1'
>>>>
>>>> >
>>>>
>>>> <Gain gain='.5'><AudioClip url='"sound/beat.mp3"'/></Gain>
>>>>
>>>> <Gain gain='.5'><AudioClip
>>>> url='"sound/cnote.wav"'/></Gain>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> e) assign different audio streams to first and 2nd channnels via
>>>> splitter take first channel (play beat in left, cnote in right)
>>>>
>>>> v4
>>>>
>>>> <ChannelMerger channelCount='2' >
>>>>
>>>> <ChannelSelector channelChoice='0'>
>>>>
>>>> <ChannelSplitter channelCount='2'><AudioClip
>>>> url='"sound/beat.mp3"'/></ChannelSplitter>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> <ChannelSelector channelChoice='0'>
>>>>
>>>> <ChannelSplitter channelCount='2'><AudioClip
>>>> url='"sound/cnote.wav"'/></ChannelSplitter>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> proposed:
>>>>
>>>> <ChannelMerger channelCount='2'
>>>>
>>>> indexStream='0,1'
>>>>
>>>> indexSourc='0,0'
>>>>
>>>> indexDestination='0,1'
>>>>
>>>> >
>>>>
>>>> <ChannelSplitter channelCount='2'><AudioClip
>>>> url='"sound/beat.mp3"'/></ChannelSplitter>
>>>>
>>>> <ChannelSplitter channelCount='2'><AudioClip
>>>> url='"sound/cnote.wav"'/></ChannelSplitter>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> /CHANNEL NODE DESIGN COMPARISON
>>>>
>>>>
>>>>
>>>> On Sat, Mar 4, 2023 at 10:20 AM GPU Group <gpugroup at gmail.com> wrote:
>>>>
>>>> CHANNEL NODE DESIGN COMPARISON
>>>>
>>>> v4: channel nodes as written in v4 Spec Draft Sound Component
>>>>
>>>> Doug proposed: add 2 fields to ChannelMerger MFInt32 indxDst and
>>>> indxSrc and eliminate ChannelSelector node
>>>>
>>>> I have both v4 and Doug proposed implemented in freewrl: if indxDst or
>>>> indxSrc is empty, thunks to v4
>>>>
>>>> I tested both methods on 3 scenarios and they both work as follows:
>>>>
>>>> a) swapping 2 channels
>>>>
>>>> v4
>>>>
>>>> <ChannelMerger channelCount='6' >
>>>>
>>>> <ChannelSelector channelSelection='1'>
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> <ChannelSelector channelSelection='0'>
>>>>
>>>> <ChannelSplitter USE='Channel_Splitter'/>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> Doug proposed
>>>>
>>>> <ChannelMerger channelCount='6'
>>>>
>>>> indxSrc='0,1'
>>>>
>>>> indxDst='1,0'
>>>>
>>>> >
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> <ChannelSplitter USE='Channel_Splitter'/>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> b) merging 2 channels into 1
>>>>
>>>> v4
>>>>
>>>> <ChannelMerger channelCount='6' >
>>>>
>>>> <Gain>
>>>>
>>>> <ChannelSelector channelSelection='1'>
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> <ChannelSelector channelSelection='0'>
>>>>
>>>> <ChannelSplitter USE='Channel_Splitter'/>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> </Gain>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> Doug proposed
>>>>
>>>> <ChannelMerger channelCount='6'
>>>>
>>>> indxSrc='0,1'
>>>>
>>>> indxDst='0,0'
>>>>
>>>> >
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> <ChannelSplitter USE='Channel_Splitter'/>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> c) merging 2 channels into 2nd merger channel (skipping first output
>>>> channel)
>>>>
>>>> v4
>>>>
>>>> <ChannelMerger channelCount='6' >
>>>>
>>>> <Gain/>
>>>>
>>>> <Gain>
>>>>
>>>> <ChannelSelector channelSelection='1'>
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> <ChannelSelector channelSelection='0'>
>>>>
>>>> <ChannelSplitter USE='Channel_Splitter'/>
>>>>
>>>> </ChannelSelector>
>>>>
>>>> </Gain>
>>>>
>>>> </ChannelMerger>
>>>>
>>>> Doug proposed:
>>>>
>>>> <ChannelMerger channelCount='6'
>>>>
>>>> indxSrc='0,1'
>>>>
>>>> indxDst='1,1'
>>>>
>>>> >
>>>>
>>>> <ChannelSplitter DEF='Channel_Splitter' channelCount='6'>
>>>>
>>>> <AudioClip url='"sound/6_Channel_ID.wav"'/>
>>>>
>>>> </ChannelSplitter>
>>>>
>>>> <ChannelSplitter USE='Channel_Splitter'/>
>>>>
>>>> </ChannelMerger>
>>>>
>>>>
>>>>
>>>> /CHANNEL NODE DESIGN COMPARISON
>>>>
>>>> -Doug
>>>>
>>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://web3d.org/pipermail/x3d-public_web3d.org/attachments/20230306/0a09824d/attachment-0001.html>
More information about the x3d-public
mailing list