Re: [fluid-dev] Support for SF4 (FLAC) format?

[Garth Hjelte]

At 06:31 AM 1/23/2020, you wrote:

>Well, my use-case is to reduce the loading time of soundfonts stored > on an 
>SD-card in an electronic musical instrument. [...] > Whether it actually has 
>any benefit with regard to loading times > remains to be seen. Turns out that 
>I don't have a use-case after all. I created some FLAC compressed soundfonts, 
>stored them in an SF3 file and did some tests on my embedded machine. Loading 
>times were significantly longer with FLAC than with uncompressed WAV. So my 
>loading times are not IO-bound... oh well. The soundfonts are about half in 
>size, so transferring new soundfonts onto my instrument will be slightly 
>quicker. But that doesn't happen often enough to be a good use-case. So, I 
>second Garth's opinion and would like to pose his question again: Does anybody 
>have a good use-case for FLAC compressed soundfonts (besides "it seems like a 
>sensible idea")? And is that use-case beneficial enough to justify opening up 
>the SF3/SF4 can-of-worms? Cheers Marcus 

Amen =) Well, the point is that if the issue is load time, you really are 
trading one problem for another. I do think that the decompression overhead is 
less than the loadtime overhead, but not enough to justify yet another file 
format. And in fact, at least in NI's case, it turns into a clever 
copy-protection scheme without admitting it publicly. 

Garth Hjelte
Sampler User