A very strange problem !

Tintin72

Hi all,

Lately, I build my own grand piano soundfont (5 layers, size: 80 Mo). I loaded it with no problem.
I don't know exactly how it happend, but later on, out of nowhere some cracking occured while playing the Piano Soundfont.
I tried to play the soundfont in Sonar and then in Vienna, but I could hear cracking in both of them.
I rebooted 2 or 3 times... cracking still remained.
I realized that when I export some notes from Vienna and I read the wav file, cracking is a part of the file!!!!
I really don't know how it just could happen.
So I had to delete and reassign all the corrupted wav files
I prefered to reduce the size of the soundfont from 80 to 66 Mo (4 layers) just in case, and now it seems to work.
But, I'm wondering if that problem could happen again.
has someone experienced this kind of pb before and knows its reason ?

My PC:
AMD Barton 2600
Motherboard Asrock K7S8XE
Chipset SIS
RAM 1.5 Go
Soundcard SB Audigy Platinum eX
KX Drivers 3537
Windows XP

Thanks for advance

TravelRec.

Are you runnig USB devices on your PC? Try to update to a newer beta release of drivers (3538m full version) to verify your problem.

Tintin72

I have an USB key that I plug sometimes.
Where can I get this beta release ? (I can't find it on the KX site)
What pb does this version is supposed to fix ?

thomasabarnes

Hi Tintin72:

Get the latest beta news and releases from the General Discussion forum (usually in a sticky at the top of the forum).

Here's the link: http://driverheaven.net/showthread.php?t=128197

Maddogg6

Im more inclined to think this error was from Vienna than simply loading a SF in KX or Sonar. I know Vienna 2.3 had issues either in XP or with KX or both maybe??...But makes more sense to me the problem occurred when WRITING the file, not reading it...

\btw, Vienna 2.3 was pretty much unusable for me with all its bugs, but I can't say I have done a whole lot of SF editing lately with the new version - much less for ones Ive read (from Lex I think??) wont be usable all at one time - I think the > 32MB may load but is cached or something... so I keep them < the 32MB limit - and just dynamically load the ones I need - but obviously not an option for so many layers.

*wonders why so many layers are needed?*

@OP: did you try the newer version of Vienna (2.4)

Tintin72

Thanks Maddogg6 for your answer.

I think you're right, pb may occured when writting the file, not reading it.
(I thought about it before).
Do you mean that a soundfont larger than 32 Mo is loadable but not
well managed by the soundblaster card ?
Why 32 Mo ? Have you got more infos about it ?
The more the instrument is layered, the more it sounds realistic (and the more you've got pb as you can see )

Maddogg6

Yes, its my understanding. - Im pretty certain a very smart guy name Lex here posted it some where. What thread no ?? Im not finding it.

This is more of fault of the sample recording (more specifically the mic and pre-amp used as well as mic placement and room acoustics) IMO. As, a single wave signal outputted is the same mixed/layered or not.
BUT Lex will also describe the issues the SB hardware has with sample playback and will suggest using something else for SF2 playback. Ive tried SFV (freeware VSTi SF2 player) - and well, he is right - you can hear a slight difference - he mentioned was in the transient response of the hardware (IIRC)

So maybe that SFV will help you ... ??? and use less layers anyway is easier to make with less layers ??

Tintin72

Actually you can find a clue on the kx project faq:
They don't say why 32 Mo, but I think we can trust them for sure, they know what they're talking about.
I also found an article wich explains more precisely what's happen in the card:
Creative Labs Audigy Sound Card

At least, I discoverded a GREAT soundfont reader named Sfz+:
rgc:audio software :: sfz+ Professional SoundFont Player
it really does a great job. No matter if your soundfonts are larger
than 32 Mo, it works !

Hmmm, I'm not sure that we're talking about the same thing.
When I said layers I mean velocity.
For example, if you press lightly the key of an acoustic piano, you'll get a weak and soft sound. Now if you press hardly this same key, the sound you'll get will be loud, brightly and percussif. So the trick is to record the piano at different velocities and use them
as layers in Vienna
Now, if your soundfont as several velocity layers, (the first one trigging between 0 and 70, the next one trigging between 71 and 90 etc...), you can reproduce (more or less) on your midi keyboard what you would get on a real acoustic piano.

Best regard

Maddogg6

Yes I mispoke - thats the same one I was talking about too.


Ahh ok - those I know as 'Splits' (key range and/or velocity). 'Layers' being played at same time and reduce polyphony (the total # of notes that can be played at one time).