SFZ format documentation
- KVRAF
- 7059 posts since 19 Apr, 2002 from Utah
One thing that I think is definitely needed for SFZ is a lowest denominator standardization for compatibility. There needs to be a proposed and accepted basic minimum that everyone agrees on. For example: If I wanted to buy Rusty Drums SFZs to use with LinuxSampler, they wouldn't work. This is supposedly because of the scripting. So, there appears to be a level of SFZ support for LinuxSampler and a more advanced SFZ level of support for Sforzando. This is just an example, but I find it frustrating as a consumer of SFZ samplesets that I can't buy SFZ samples if my DAW is Linux and I'm using LinuxSampler. I bought Secret Agent Guitar last year, and haven't tried it yet, because supposedly it uses scripting as well.
What this all boils down to is that there MUST be a standardization in order for SFZ to remain a viable standard. The standard needs to be maintained and supported by VST Sampler instrument developers and by SFZ instrument content developers. If the group accepts SFZ level 1 support as the baseline compatibility, developers should not use opcodes that make their sample sets incompatible. SFZ should run on EVERY sampler that claims to be SFZ. Likewise, if SFZ level 2 support is decided by the consortium to be the base line of support, VST Sampler instrument developers need to support ALL of the new opcodes that Level 2 supports. If not, then they shouldn't claim full SFZ support.
We need to have a community meeting and a create a consortium accepted SFZ baseline.
Standardization is the only way.
People that I can think of that should be included:
Translation software teams
VST/AU/etc. Instrument software developers that write SFZ samplers and players
Sample instrument makers that record and create multisamples for the format that the samplers and players read
Anyone who is interested in setting an free and open standard for the industry
This, in my opinion is the ONLY way to move SFZ forward--it's too fragmented right now. We have Plenty of proposed and supported opcodes. (In fact, I'd say that we have too many--some of them are too extreme for a "bottom denominator" supported set.
Doing this, we consumers/users who buy the SFZ products will always know what we are buying, and developers will all have a standardized level that their products will be expected to meet to be called SFZ compatible.
SFZ instruments could be allowed a superset of the SFZ format, but would at the very least have to support the SFZ baseline to claim full SFZ compatibility. If it is an extended format (above and beyond the accepted SFZ standard), the format name needs to be called something else (possibly SFZ 3, SFZ+, etc).
Likewise, to be called SFZ compatible, multisample makers need to make their samples run at least on the baseline SFZ standard to be called SFZ. If they want to add extensions, they can call it something else possibly (SFZ 3 SFZ+, etc). But SFZ needs an accepted baseline and everyone needs to agree to it and stick to it.
It's the only way!
What this all boils down to is that there MUST be a standardization in order for SFZ to remain a viable standard. The standard needs to be maintained and supported by VST Sampler instrument developers and by SFZ instrument content developers. If the group accepts SFZ level 1 support as the baseline compatibility, developers should not use opcodes that make their sample sets incompatible. SFZ should run on EVERY sampler that claims to be SFZ. Likewise, if SFZ level 2 support is decided by the consortium to be the base line of support, VST Sampler instrument developers need to support ALL of the new opcodes that Level 2 supports. If not, then they shouldn't claim full SFZ support.
We need to have a community meeting and a create a consortium accepted SFZ baseline.
Standardization is the only way.
People that I can think of that should be included:
Translation software teams
VST/AU/etc. Instrument software developers that write SFZ samplers and players
Sample instrument makers that record and create multisamples for the format that the samplers and players read
Anyone who is interested in setting an free and open standard for the industry
This, in my opinion is the ONLY way to move SFZ forward--it's too fragmented right now. We have Plenty of proposed and supported opcodes. (In fact, I'd say that we have too many--some of them are too extreme for a "bottom denominator" supported set.
Doing this, we consumers/users who buy the SFZ products will always know what we are buying, and developers will all have a standardized level that their products will be expected to meet to be called SFZ compatible.
SFZ instruments could be allowed a superset of the SFZ format, but would at the very least have to support the SFZ baseline to claim full SFZ compatibility. If it is an extended format (above and beyond the accepted SFZ standard), the format name needs to be called something else (possibly SFZ 3, SFZ+, etc).
Likewise, to be called SFZ compatible, multisample makers need to make their samples run at least on the baseline SFZ standard to be called SFZ. If they want to add extensions, they can call it something else possibly (SFZ 3 SFZ+, etc). But SFZ needs an accepted baseline and everyone needs to agree to it and stick to it.
It's the only way!
Vendor‑Dependent Copy Protection: Customers lose. Pirates win.
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
- KVRAF
- 7059 posts since 19 Apr, 2002 from Utah
One way to start would be to get everyone to state the opcodes used in their product currently, whether it be VST Sampler instruments/players or whether it be multisample developers. This would also go for Translation software--we would need to know what opcodes they support. Then, we could create a baseline that most developers currently use. If the major developers would be willing to divulge their level of opcode support, then that would be a great start.
Developers/products that come to mind:
LinuxSampler
Sforzando
Bliss Sampler
Carla
Redux
ChickenSys (Translator)
Awave
Karoryfer
SampleScience
Samples From Mars
Wave Alchemy
michael picher music
Jeff G
etc...
Developers/products that come to mind:
LinuxSampler
Sforzando
Bliss Sampler
Carla
Redux
ChickenSys (Translator)
Awave
Karoryfer
SampleScience
Samples From Mars
Wave Alchemy
michael picher music
Jeff G
etc...
Vendor‑Dependent Copy Protection: Customers lose. Pirates win.
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
- KVRAF
- 7059 posts since 19 Apr, 2002 from Utah
Maybe we could start by sending out an excel spreadsheet survey which includes all of the known opcodes and get everyone who develops (and wants to have a voice) to do the survey and give their input. Those who want to participate would be welcome to do so. In the end, we could have a set "Official" standard that everyone could follow/aspire to, and no one would be held back if they want to go beyond as long as the baseline level is supported. We could have a consortium name/badge that all developers who meet the standard could advertise their product to support, which would help with sales of the product as well. 
Imagine a product being able to brag: "SFZ Consortium" Compliant! or "SFZ Cooperative" Compliant! or "SFZ Roundtable" Compliant! or etc., etc., etc... whatever the group chooses.
Imagine a product being able to brag: "SFZ Consortium" Compliant! or "SFZ Cooperative" Compliant! or "SFZ Roundtable" Compliant! or etc., etc., etc... whatever the group chooses.
Vendor‑Dependent Copy Protection: Customers lose. Pirates win.
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
-
- KVRAF
- 3352 posts since 19 Mar, 2008 from germany
@audiojunkie:
You are right - a standard is a capital point. As far as I know there are three
standard levels at the moment:
1. sfz 1
2. sfz 2
3. sfz 2+ (Plogue Sforzando, Aria engine)
The levels and the classification of all levels are already documented at
http://www.sfzformat.com - thanks to DSmolken!
Missing is: The sensitive use and the labeling of these levels by developers. Some
developer only create "sfz 2+" sample sets and some create very basic "sfz 1" sample sets.
That can be confusing - indeed.
You are right - a standard is a capital point. As far as I know there are three
standard levels at the moment:
1. sfz 1
2. sfz 2
3. sfz 2+ (Plogue Sforzando, Aria engine)
The levels and the classification of all levels are already documented at
http://www.sfzformat.com - thanks to DSmolken!
Missing is: The sensitive use and the labeling of these levels by developers. Some
developer only create "sfz 2+" sample sets and some create very basic "sfz 1" sample sets.
That can be confusing - indeed.
free mp3s + info: andy-enroe.de songs + weird stuff: enroe.de
-
- KVRAF
- Topic Starter
- 2210 posts since 20 Sep, 2013 from Poland
Yeah, it can be confusing, and hopefully someday we'll get an official SFZ 3.0 standard. I could go through the older Karoryfer instruments and see which ones are just SFZ 1 or SFZ 2, but the more recent ones I do label as requiring Sforzando, so at least there's that. But, yes, at least now we have a site with almost all the opcodes - almost, because there are some such as decim and bitred which were used in Cakewalk stuff, didn't make it to ARIA, and which I still need to document.
Some things might be possible to redo so they'd work with just SFZ 2 but I end up using things beyond that because the ARIA extensions do make things like pan, volume and tuning controls much easier, and using include and master to make things more modular is a huge time-saver when dealing with thousands of samples.
Some things might be possible to redo so they'd work with just SFZ 2 but I end up using things beyond that because the ARIA extensions do make things like pan, volume and tuning controls much easier, and using include and master to make things more modular is a huge time-saver when dealing with thousands of samples.
- KVRAF
- 7059 posts since 19 Apr, 2002 from Utah
I can agree with that.
I just wish all instruments/players supported sfz 2+ so that all content (even the basic content) works. I'm not a fan of Sforzando (even though it is free). It is not cross platform and doesn't have any controls without special permissions and such. It DOES have DFD support though.
LinuxSampler has DFD and is cross platform, but supports less opcodes than Sforzando.
Bliss sampler supports SFZ natively and has cool VST + effects sampling (for SFZ) , but it doesn't have DFD. It barely supports Level 1 opcodes.
Carla has basic SFZ support and is not cross platform compatible.
Redux barely supports SFZ and is not cross platform compatible.
If Sforzando became cross platform to include Apple and Linux and provided at least basic ADSR control, or if LinuxSampler supported Level 2+ opcodes, or if Bliss supported level 2+ opcodes then any one of them would work, and content developers wouldn't have to worry at all.
I just wish all instruments/players supported sfz 2+ so that all content (even the basic content) works. I'm not a fan of Sforzando (even though it is free). It is not cross platform and doesn't have any controls without special permissions and such. It DOES have DFD support though.
LinuxSampler has DFD and is cross platform, but supports less opcodes than Sforzando.
Bliss sampler supports SFZ natively and has cool VST + effects sampling (for SFZ) , but it doesn't have DFD. It barely supports Level 1 opcodes.
Carla has basic SFZ support and is not cross platform compatible.
Redux barely supports SFZ and is not cross platform compatible.
If Sforzando became cross platform to include Apple and Linux and provided at least basic ADSR control, or if LinuxSampler supported Level 2+ opcodes, or if Bliss supported level 2+ opcodes then any one of them would work, and content developers wouldn't have to worry at all.
Vendor‑Dependent Copy Protection: Customers lose. Pirates win.
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
- KVRAF
- 7059 posts since 19 Apr, 2002 from Utah
I like your products! I don't want you to get me wrong.DSmolken wrote: Wed Feb 27, 2019 7:58 pm Yeah, it can be confusing, and hopefully someday we'll get an official SFZ 3.0 standard. I could go through the older Karoryfer instruments and see which ones are just SFZ 1 or SFZ 2, but the more recent ones I do label as requiring Sforzando, so at least there's that. But, yes, at least now we have a site with almost all the opcodes - almost, because there are some such as decim and bitred which were used in Cakewalk stuff, didn't make it to ARIA, and which I still need to document.
Some things might be possible to redo so they'd work with just SFZ 2 but I end up using things beyond that because the ARIA extensions do make things like pan, volume and tuning controls much easier, and using include and master to make things more modular is a huge time-saver when dealing with thousands of samples.
Vendor‑Dependent Copy Protection: Customers lose. Pirates win.
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
-
- KVRAF
- Topic Starter
- 2210 posts since 20 Sep, 2013 from Poland
Yeah, that's understandable. Doing a Linux version would take a few extra evenings of work (a lot less than a Kontakt version, at least), so then it becomes a question of how many potential sales I'm losing out on by not doing one. With free samples at least I can say "the wavs are included, map'em for whatever sampler you want" but if expecting people to pay, that doesn't really fly.
And at least now if somebody does want to make instruments that will work in LinuxSampler or just need basic SFZ 1 or whatever, the information needed for that is all on one site. That's a big step forward in theory. Whether it's actually resulting in more people making SFZ instruments in practice, I don't know, but we'll see.
And at least now if somebody does want to make instruments that will work in LinuxSampler or just need basic SFZ 1 or whatever, the information needed for that is all on one site. That's a big step forward in theory. Whether it's actually resulting in more people making SFZ instruments in practice, I don't know, but we'll see.
-
- KVRAF
- 2436 posts since 5 Jan, 2006
A possible solution for sample developers is to create two sfz programs: one for Sforzando with custom controls and advanced features and another one simple "universal" sfz that can load in any sampler (just mapping)
- KVRAF
- 7059 posts since 19 Apr, 2002 from Utah
That's really what I'd like to see.PTV wrote: Wed Feb 27, 2019 10:12 pm A possible solution for sample developers is to create two sfz programs: one for Sforzando with custom controls and advanced features and another one simple "universal" sfz that can load in any sampler (just mapping)
Vendor‑Dependent Copy Protection: Customers lose. Pirates win.
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
- KVRAF
- 7059 posts since 19 Apr, 2002 from Utah
I know that if I were personally a content creator of SFZ multisamples, I'd be wanting as many users as possible to buy them, and that would mean making my SFZ patches work on as many systems and instruments as possible. But at the same time, I can fully understand how some developers want to have specialized scripts that allow advanced features such as strumming, drum swirls, key switching, etc. Advanced features such as these can add value and sales as well. The key would be to provide support for both on the content provider side, and further opcode support on the sampler/player development side.
Vendor‑Dependent Copy Protection: Customers lose. Pirates win.
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
- KVRAF
- 7059 posts since 19 Apr, 2002 from Utah
Another consideration is now iOS an Android now have some SFZ supported applications and lots of interested users. These tools currently don't support beyond level 1 opcodes. While I see no problem in supporting advanced opcodes, catering to the largest userbase by providing sample sets using SFZ level 1 will surely be more financially viable.
Also, Awave only supports SFZ level 1 conversions, and Translator (I believe) only supports up to level 2. None support SFZ 2+
Also, Awave only supports SFZ level 1 conversions, and Translator (I believe) only supports up to level 2. None support SFZ 2+
Vendor‑Dependent Copy Protection: Customers lose. Pirates win.
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
- KVRAF
- 7059 posts since 19 Apr, 2002 from Utah
By the way, that's exactly what Impact Soundworks did with Shreddage X. They have an advanced Sforzando (SFZ 2+) version and a (LinuxSampler) (SFZ 1) version patch set. The advanced patch set has key switching and multis and such. The basic basic set has separated patches and no key switching.audiojunkie wrote: Wed Feb 27, 2019 10:24 pmThat's really what I'd like to see.PTV wrote: Wed Feb 27, 2019 10:12 pm A possible solution for sample developers is to create two sfz programs: one for Sforzando with custom controls and advanced features and another one simple "universal" sfz that can load in any sampler (just mapping)![]()
Vendor‑Dependent Copy Protection: Customers lose. Pirates win.
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
-
- KVRAF
- Topic Starter
- 2210 posts since 20 Sep, 2013 from Poland
Might be worth trying that with one instrument and seeing how it affects sales. Sure, Impact apparently haven't felt the need to do it since Shreddage X, but that was years ago and maybe the Linux slice has gotten bigger. I kinda associate people who use Linux for music with wanting to avoid spending any money, ha. But I'd be happy to prove myself wrong.
- KVRAF
- 7059 posts since 19 Apr, 2002 from Utah
You might be right about the Linux people. Hehehe! For me, it's about having my tools stay viable for a long time. I don't mind buying software (just bought Diva), but I hate things like Challenge/Response or ILOK or operating systems that dictate how I do things or might go out of business and leave me high and dry.DSmolken wrote: Fri Mar 01, 2019 6:51 am Might be worth trying that with one instrument and seeing how it affects sales. Sure, Impact apparently haven't felt the need to do it since Shreddage X, but that was years ago and maybe the Linux slice has gotten bigger. I kinda associate people who use Linux for music with wanting to avoid spending any money, ha. But I'd be happy to prove myself wrong.
P.S. I already own Secret Agent Guitar, and would LOVE to use that with LinuxSampler (and Bliss, and other samplers), so if you are willing to do one, that would be the one I would wish for first. Please!
Vendor‑Dependent Copy Protection: Customers lose. Pirates win.
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
(Also: I'm Accused of lying about Linux—it boots, runs my pro audio workflow, stays stable, updates--though yearly dismissed as “niche”. Yet I'm the deluded one.)
