The Rocky Road Of Audio Software Development...
From the original Analog Industries archive. Some links and images from that era no longer resolve.
Well, it’s been an interesting day. Ever since we released Phase Two (way back in the day, it seems; was it only Sunday?) we’ve been discussing the transition to Universal Binaries, and how Audio Damage would deal with this rather extreme event (from a developer’s perspective, anyways.) We had pretty much decided on our plan of attack, then Steinberg dropped a bombshell on us this morning, in the form of a new VSTSDK, version 2.4.
This forced our hand in the decision-making process, which is both good and bad. One can assume that the two main VST hosts we deal with, Cubase/Nuendo and Live, will be VST2.4 compliant almost immediately, if they weren’t already. (Which they are, of course.) On the other hand, what I refer to as the Stupid Hosts, which is pretty much everything besides Cubase/Nuendo and Live, will deal with the VST2.4 spec in fits and starts. DAWs that don’t correctly host VST2.3 now can hardly be expected to host 2.4.
So, in order to be fully compliant with _all_ hosts and both flavors of AU, we’d be required to make and support eight different builds of every plugin we release, in the form of VST old and VST new on Windows, the same for both flavors of OSX, and AU for both flavors. This would have the effect of quadrupling our work, both during creation and during the support phase of a product. Now, it already strains our production cycle and support load to just do AU and VST (the latter is more or less the same on both OSX and Windows, currently.)
So, the Official Word from Audio Damage on the upcoming transitions to both 64-bit Windows and Intel Mac is thus: we will build our plugins using the current version of VST, whatever that is. (In this case, 2.4.) We will also offer all our Macintosh products as Universal Binaries. Thus, if you use a host that isn’t Cubase/Nuendo, Live, Logic, or Digital Performer, your mileage will start varying in the not-too-distant future, and you need to complain to them.
Since this is going to get complicated quickly, here’s our plan. We will update FuzzPlus 2 to current VST standards (which I’ve already done) and also add all the bells and whistles to it that our commercial plugs have, such as MIDI Learn and our new text entry gee-gaw, which was created specifically to please Suit&TieGuy, so he can stop bothering us about text entry fields. This way, if you need to know if your host is compliant with what our plugs need, you can just check it with FuzzPlus2. We will also release the long-awaited AU version in Universal Binary, for the same purpose.
Which brings us to the Universal Binaries. Here’s where the bad news comes in to play. All current Audio Damage products will be ported to Universal Binary. The crossgrade for each product will be US$5.00, which will go directly to covering the cost of the Intel Macs we have to purchase to do all this. Current products will all increase in price by $5.00, as well, except PhaseTwo, which will remain at its current price. All future products will be provided as Universal Binary installers of VST2.4 and AU for OSX, and an installer for 32- and 64-bit Windows hosts that are VST2.4 compliant.
Now, as for the timeline. This is going to take a bit of time for us to get everything lined up. We have to receive our new iMacs and then learn how to make Universal Binaries, then port our product line. I’m unable to provide an estimate for when this all will occur, but suffice to say that by the time Live, Logic, DP, and Cubase/Nuendo are all available for Mac Intel, our products will be as well. The one exception to this will be the Mayhem package, which will shortly cease to exist in its current form, as it is replaced with more sophisticated offerings.
I hope that answers all the questions that I’ve been receiving. Adam and I will field your thoughts in the comments section of this post.
19 comments
I think it's very reasonable of you considering all the new code, both VST and AU. I'll be buying plenty of your software when I get an Intel mac laptop in a few months, good luck.
I'm totally all for paying $5 for UBs for all my AD stuff. Hell, frankly I'd probably pay $10.
sounds more than reasonable to me and nice that you guys are so up front about it all :)
As a side note, I mention in the above post that I've built FuzzPlus2 with the new VST 2.4 spec. It's a reasonably time-consuming process, but not too complicated. I've now checked it in every host I have on my PC, which is pretty much every host, and it seems to work fine. Obviously, hosts that don't hook in to the 2.4 extra features (which is all hosts, currently) can't do, say, 64-bit processing or whatever, but it seems to be fine.
The reason this concerns me is that when Nuendo/Cubase is released for OSX Intel, it will _only_ support VST 2.4-compliant plugins, according to Steinberg. (This won't be the case with Windows.) So we need to port all our products to this new format regardless. We'll have to do it on Windows, too, even though it's apparently not necessary, as we work from a single code-base for all our products, and we can't go splitting the tree just to do the Mac side.
-CR
all of that is more than fair. i don't think i will be going for a new mac for some time, but when i do, i'll be happy to throw you guys 5 bones per update.
thanks for the info and all the great work.
well, i don't mind paying a few extra bucks even though i own most of all AD plugz... as far as i'm concerned, i'm more than happy to finally see the mayhem suite finally ported to audio unit...
5$ per upgrade is more than a reasonable fee for such a tremendous amount of work...
BTW, have you thought about renewing some work around the RTAS format? I'm glad to be able to use the Mayhem natively within ProTools... :)
five bucks!?!?!
are you kidding?!!?!
okay, sure. you want a tip too?
you know, Discord is pretty much the best plug in i've ever bought. thanks for the text entry fields. in fact, i stopped bitching about that shit last week yo.
yes, i do actually buy plug ins. NI B4 is an NFR though ...
It's a pretty cool solution Adam came up with. Now that we have it, I find myself wondering why everyone doesn't do it this way. What happens is you double-click on _any_ control, and a text field pops up over the control allowing you to enter a real-world value. So no messing up my UIs with stupid little text boxes, and no need to gingerly stir the pot to get that fine value.
We're still in the testing stages on this, but all future products will have it.
-CR
Speaking of Performer... I've not touched it yet. I'm currently taking a class based around it. I've done a little with Protools and Logic, but I'm more into Cubase. How is it? Performer, that is. Just as a general question.
I have almost no experience with DP. I know that people that like it are quite militant about it, the way I am with Nuendo. I also know that its built-in plugs are on par with the Logic built-ins. Other than that, it's outside my realm of expertise. I am a fan of MOTU hardware, so how bad could the software be?
Viz. our products and RTAS, don't hold your breath. We got in a bit of a tiff with Digidesign. They asked us how our RTAS stuff was doing, and we said "well, it sells about a tenth of what VST/AU sells, yet it is ten times the trouble to make" and they yanked our dev license. I'm not making that up. So, on top of the fact that I _really_ dislike PT in a general sense, I have an axe to grind against the company that makes it. There'd need to be an incentive for us to do RTAS. Like, we were stranded on a desert island, and the only way we could be rescued is if we ported one of our other products.
As I mentioned in the post, Mayhem is about to be obviated by other products, and will cease to exist on this temporal plane in its current form. I'm sorry about that, but it's getting long in the tooth, and needs to be addressed.
-CR
DP rocks it hard and forever.
How do minor updates to existing products (for example gain control for the A907) fit into the timeline? Are you gonna port to UB/2.4 first or add features first?
"It's a pretty cool solution Adam came up with. Now that we have it, I find myself wondering why everyone doesn't do it this way." - CR
yeah actually that rocks pretty damn hard.
good solution.
v-knobs look pretty but i find them totally unusable. but then again, not everyone spends most of their mixing time tabbing around text fields in Performer and entering values manually.
did i mention the coolest thing about the eventide Hx000 boxes is the numeric keypad? when the unobtanium optical encoder goes out on my 3000SE i won't miss a thing. it's a habit i picked up from the XP-80 i used to own whose encoder went haywire 2 weeks after the warranty ran out so i just unplugged it and lived without.
fucking digital, baby! totally. fucking. digital.
<quote>On the other hand, what I refer to as the Stupid Hosts, which is pretty much everything besides Cubase/Nuendo and Live, will deal with the VST2.4 spec in fits and starts. DAWs that don't correctly host VST2.3 now can hardly be expected to host 2.4.</quote>
Them's fightin' words...
You make a lot of sense, and I admire your devotion to your host, but as a majorly happy Sonar user I take umbrage with the characterization of it as "Stupid."
(Well, not really, but umbrage is fun to type.)
Bad news, Puffer. Sonar isn't a VST host. It's a DirectX host with a (well-written, I'll take pains to point out) wrapper. So it doesn't count.
-CR
Chris, I'm not a plugin person, so might not ever buy one of you (even though the Discord 1.5 demo sounds the part!), but I have to give you props for the timely, honest and upfront way you've come out with this. Not only that, you've kept the prices of upgrades very reasonable!
Greetz,
characterstudios
A couple of amplifications on what Chris said, and some responses to what you all said:
-Mayhem's future is currently undecided. The biggest problem with moving it forward (as in Intel and VST 2.4) is that it's written in CodeWarrior, a development system which is discontinued and no longer viable for creating software. That means updating the Mayhem family would take more work than almost all of the other updating projects put together. Hence we're tempted to introduce new products instead. We have a good concept of what we'll do with Crush, and are kicking around ideas for MasterDestrukto. FilterPod and TimeFnk are less clear, particularly since the latter is arguably replaceable with Ronin or DubStation. In any case, regardless of what we do, we will take steps to make our Mayhem customers happy.
-RTAS is a dead end for us. Chris's story of what happened is correct. It's one thing to put a huge amount of effort into developing a product and not have it sell well. It's another thing to have a company in charge of a proprietary plug-in format slam the door in your face, after they were the ones that opened the door in the first place, simply because you mention that your sales have been disappointing. To have those two things happen together, well, Chris pretty much said it all. We're really unlikely to do any more RTAS development, particularly now that fxpansion sells a VST/RTAS wrapper.
-Big thanks to everyone who has voiced words of support and understanding here. It's really encouraging in the midst of all of this to get this sort of response.
-Regarding the gain control on 907A: I hear ya. It really needs it. After further discussion, it's likely that we're going to make a sort of combined/case-by-case approach. In some cases, like PhaseTwo, it will simply be a UB/VST2.4 update. PhaseTwo's barely out the door, so there's no point in changing it yet (if ever--I mean, it's already as close to the hardware as we could possibly make it; Chris will probably do me bodily injury if I start tweaking the DSP code again). In cases like 907A, there will probably be a more complicated event. I'm gonna be cagey here and say that this event could involve much more than a gain control...
-Cakewalk products are generally not stupid. I'll agree with that.
-No promises or even speculations on dates for this stuff yet. Even in ideal situations it's pretty much impossible to predict how much time any software project will take, and this is far from an ideal situation. We just don't know enough yet about what we're getting into. So far it's looking fairly straightforward, but unforseen things can happen, obviously. (For example: during the latter stages of development of PhaseTwo my main Mac's main hard drive went completely dead one morning. I mean, *dead*. Total hardware failure. All of Audio Damage stuff is replicated and backed up so I didn't actually lose any of that, but I lost several days of work to replacing the drive and rebuilding its contents. Then, right as we were in the final pre-release preparation phase, my SO fell down the stairs and we spent half the day in the ER. You get the picture.) The only date I know for certain is that Amazon says they'll have an Intel iMac on my doorstep on Wednesday.
--Adam
A question concerning the Fuzz+ 2 update:
Should it work in hosts that are still only VST2.3? I haven't looked too closely at the changes Steinberg implemented, so what should we be looking for in testing it out?
I'm gonna start a new post re: FuzzPlus2 and what I've discovered about 2.4 over the weekend here in a couple hours.
-CR
Comments are closed on this post.