are they dx now? their site says no...but can you use them with a vst-dx adapter? does it REALY work or will we have the problems sonar users talk about ?
The sonar problems are effectively over, and yes, they work. I'm presenting for Mackie at the NAMM show, showing the UAD1. The idea of DirectX is somewhat misleading, it's the wrapper that allows it to be used in Vegas. The wrapper has been updated and seems to be fine, although occasionally I lose automation in SONAR with the plugs. I've not taken the time out to try the automation in Vegas, as I'm focused on other Vegas features right now.
Also Pipe, you might want to check a post by Angus at KvR here concerning his VST-DX wrapper. There's a public beta available now with improvements that concern both Sonic Foundry applications and the UAD-1.
this thread is somewhat scary. The pluging delay troubles with the UA card, I think will surely be multiplied in vegas. The cubendo users speak of their plugin delay compensation ability, but we have no such thing....is this card just going to be a nightmare or what?
Hey Pipeline,
Read this thread from KVR-VST forum(FXpansion) concerning some testing I did(billybk1) using SONAR 2.1, the latest FXpansion VST-DX 4.2b2 Adapter, and a UAD-1(ver 2.26):
It should put some of those "plugin delay troubles" to rest, for good. I have not tested, with any SOFO products, because, I am pretty sure they do not have full automatic delay compensation built into it's programming architecture like SONAR does. This is the reason, when inserting and deleting effects while SONAR is playing, exhibits a momentary pause or "gapping". This is SONAR's audio engine resetting itself to stay in sample accurate sync at all times. If an app does not exhibit this behavior, it does not have full ADC, on all track inserts, Aux Buses and Main outputs. When doing my testing with SONAR, the adapter and the UAD-1, I found the only limitation was that you can't use multiple UAD-1 effects on a track insert that is being routed to an Aux bus that has multiple UAD-1 effects or you will get an instant dropout. It makes no difference how you set the buffers, this will always happen. Other than that, the ADC implementation is next to perfect. You can use the UAD-1 effects just like any other native VST/DX effect, in combination with each other and with DXi/VSTi instruments. No need to use plugin delay compensators plugins nor is it necessary to move/align tracks to fix timing errors. Everything works the way it should, staying in sample accurate sync, without having to think about it, at all times. The way it should be. My other observations included:
On my PIII 1Ghz CPU, each added UAD-1 plugin added less than 1% native CPU overhead(far less than the 10% the Bomb Factory guy was talking about) Typically (10) UAD-1 plugins added about 7%-10% overhead to my native 1Ghz CPU.
It is also important to set correctly the buffer settings, so that they all match, between your host app, the adapter, and the UAD-1, or you will get an instant error message stating that you have incorrect buffer settings and you will not be able use the UAD-1 plugins. I basically have two settings, one for mixing and multitracking(23.2msec latency,1024 samples), the other for using softsynths and live input monitoring(2.9msec latency,128 samples). I can easily switch between the two in less that a minute, by resetting the buffer sizes in SONAR, the adapter and the UAD-1 so that they all match. This will insure that my project is always in sample accurate sync. At least there is a DSP solution available, that works with and acts like normal VST/DX effects and has ADC through out the signal chain. Hopefully, once the version 3.0 software for the UAD-1 is released, with full DX ADC implementation, there will not be a need to set buffer sizes, it will all be automatic. But until then, the FXpansion 4.2b2 adapter does work and takes full advantage of native automatic delay compensation, at least in SONAR 2.1.
cool thanks...How is it turning bypass on and off on say a pultec in sonar? Does it have different delays? And I guess the big problem here is that vegas has no delay compensation that Im aware of because I run into many plugin delay troubles on just regular dx fx.
Using the full ADC in SONAR, everything stays in sample accurate sync at all times(no delays or out of sync tracks), except for when you insert or delete a plugin from an FX bin, during playback(and that goes for all types of plugins in SONAR, because of the native automatic delay compensation on all inserts, AUX buses and Virtual Mains) there is a slight momentary pause,as the ADC resets itself to stay in sample accurate sync. Otherwise, once an effect(native or UAD-1)is inserted into a track, AUX bus or Virtual Main, you can open, close, mute, unmute, adjust effect parameters, change presets, do envelope node parameter mixing, with no gapping or zipper noise in real-time with no perceivable delay. Like most VST effects, the UAD-1's effect parameters are also automatable using envelope nodes from the track view.
At last, a high-end DSP solution that works, like native FX plugins, in a DX software environment. I can't believe how cheap($475.00-$599.00 brand new,street) these UAD-1 cards are nowadays, I may get a second one, now that the latest 3.0 version allows for the use of multiple cards in one PC.
I spoke with Cakewalk, they do no ADC with DX, because, as I have stated before, there is no defined mechanism to do this. They do deal with non-inplace plugins - which is what a plugin that inserts a delay essentially is. Any DX host MUST deal with non-inplace plugs. For the record, SF has been dealing with them since we first release Sound Forge 4.0a.
What is happening with the UA-1, is the VST to DX wrapper they use will account for the delays caused by the UA-1 and hide this delay from DX hosts.
NOTE: There is a industry group that is attempting to define delay compenstaion as an extension to DX, but so far nothing has been finalized. If and when there is some standard proposed, I assure you that will we do the right things. I still don't agree with the DSP purists out there for the need for this, but it is an issue that wont go away is seems.
"I spoke with Cakewalk, they do no ADC with DX, because, as I have stated before, there is no defined mechanism to do this. They do deal with non-inplace plugins - which is what a plugin that inserts a delay essentially is. Any DX host MUST deal with non-inplace plugs. For the record, SF has been dealing with them since we first release Sound Forge 4.0a."
Hi Peter,
If that is the case, then we should be able to use the same FXpansion VST-DX 4.2b2 Adapter,(the same type of wrapper we've used to convert VST plugins to DX in SOFO apps all along), with the UAD-1 in SOFO apps, with the same results too, like in SONAR. There has to be something else going on, besides just the adapter, or it would work with SOFO apps too. Ron Kuper has stated on more than one occaison, when specifically asked about ADC, that indeed SONAR had ADC as part of it's DX architecture. Here is also a direct quote from a senior tech from Cakewalk answering a question about ADC from the SONAR newsgroup, last year:
"SONAR will automatically compensate for any latency caused by realtime effects plug-ins. This applies to track inserts, aux busses, and virtual mains. The technology for this is a component of DirectX and is therefore directly integrated into our audio engine. Basically, each packet/buffer of information transmitted to the mixing engine is time-stamped, which allows us to play back different audio streams in sample accurate synchronization."
Hope this helps,
Jesse Recinos
Sr. Technician
Cakewalk
It sounds like they are doing something with DX and ADC,in SONAR, to make it work. Also, Universal Audio is supposed to be coming out with a DX implementation on the UAD-1, in their next 3.0 update (hopefully @ NAMM this week)which will provide ADC for native DX hosts. Here is a direct quote from the UAD-1 forum on December 28th, 2002:
"We'll be showing DX versions of the UAD plugins at NAMM. These will have full delay compensation within Sonar and other DX hosts. We hope to include these in our 3.0 release, but we'll know for sure by the show. In the mean time, the latest version of the FXpansion wrapper has been tested with Sonar and works very well with the UAD-1."
-Joe
Joe Bryan
VP Engineering & Technology
Universal Audio
Santa Cruz
Will SOFO apps, like in SONAR, be able to provide full ADC on all inserts, Aux Buses and Main outs when using the UAD-1?
If yes, great, that will be good news!
If not, then why not?
Peter, for one thing, this card has a reverb that is supposed to be decent, and that is a FAR cry from anything offered in DX or VST native...accoustic mirror is AWESOME, but takes unreal power to run, Mirror on a card would be cool. Also, apps including vegas get real funny with different plugs running on it, some just dont get along. This isnt something that happens in the real world, so just trying to add some stability in the virtual one. But then again, it sounds like this card will add a few of its own instabilities.
I always wonder, WHERE is this all going? I guess naievely, I thought from the beginning that all apps were just starting on a road towards complete professional useability and integration like any other studio device would have. I realized we were a LONG way from there and that there would be tons of bumps in the road. I thought that there would come a time, where we could use apps like any other tape recorder; monitoring schemes that people were used to, the ability to route out to other devices WITHOUT sacrificing functionality, someday an end to the effects of latency, even though latency itself may never be vanquished....
Do you see a day when we can patch in say, an LA-2A hardware unit onto a track's insert, with full delay compensation making up for the drivers, AD and DA stages? Supposedly Nuendo 2 will do this, but Im not holding my breath. I mean this is the kind of thing were all eventually aiming for right?
I just tried the Samplitude 7 demo and I put a DX plug called UltraFunk Gate R3 on it on a bass track. I then put another copy of that same bass track on it at the same point in time but without the plug-in. I set the gate to do nothing but look ahead 40ms. When I played them together I heard the track, just louder. Trying the same thing in Vegas, I hear 2 bass parts or a bass part with 40ms delay on it. What gives?
I asked Ron specifically about this. He said no ADC goes on. It isn't defined, so unless Cake is using a proprietary mechism - which Ron said they are not - then some other external mechinsm is doing this for SONAR.
There is no ADC capability defined for DX. Ron knows this and we have discussed it many times.
I am fully aware of time stamping of samples with DX. It is not the same thing as ADC.
As far as what the UA-1 is doing, they don't have a native DX impelementation, but rather they contracted FXpansion to update thier wrapper to supply DX functionality with their VST based architecture.
The delay compensation is handled by the wrapper prior to the data beind delievered to the host.
We do the exact same thing that SONAR does and have been doing LONG before anybody else has. There is currently nothing additionaly we can do.
Make 2 tracks one with the gate on it, set to lookahead the max 40 mS, and see if they echo. Maybe even my vegas setup is messed up, can someone see if the same problem I have happens in vegas for you?
Peter, it compensates perfectly in Sam7 so there must be something.
I only got the sam 7 demo and it wont let you save. I just dragged and dropped a track into channel 1 then drag and dropped a track into channel 2. I set the ultrafunk gate on channel 1 and at first it had a delay, but upon hitting play the second time and everytime thereafter, the delay was gone
I just spoke with the UltraFunk guys in your response to your post on their forum. They recanted very quickly when I asked about their response. They were not aware of any issues with SONAR so they "assumed" it worked.
I am still trying to get Samplitude demo to run. (Doesn't seem to like my system.)
I will let you know what I find.
If you could in Samp7: Start playback and then adjust the look ahead parameter in realtime.
Ok, I got Samp7 demo to run. Same issue as Vegas and SONAR when you use the DX plugin no matter how many times I playback.
When I run the VST version of this plug, Samp will eventually compensate for it using the VST delay compensation mechanism. However, if you change the look ahead, things get out of sync very quickely.
Same here peter, I changed the lookahead in realtime and nothing I did, not even " force latency" worked to make it right again....now the question is is it just ultrafunk? Im gonna try the ' trackplug" demo too and see what happens..is it even possible for lookahead to work without the delay happening?It seems to me that since the data is already on the hard drive, these plugs could read ahead all they want right?
I just tried it with the vst version and if I change it while running it starts screwing up, but as soon as I hit stop then play again, even with the 40mS lookahead it works fine. I know you guys dont want to stick VST in but...Is sidechain any more possible in VST than DX? or will it be the same troubles as the DX apps that sidechain ?
Nope, this is not how things work. The plug in doesn't read the file. The host app is doing all file i/o. The host reads the data, then sends it to what ever "mix paths" the app has defined, in this case to a plugin.
However, a plugin that does want to do a look ahead can force the host to provide it the look ahead buffer BEFORE it produces any data based on the input. The ultrafunk does not do this. These types of plugs are refered to as non-inplace plugins. Our Acoustic Mirror plugin is a worse case scenario of a non-inplace plugin. (Not how MANY hosts will NOT work with such plugins. This is why many plug in vendors wont implement their plugs in this way.)
If the plugin is set up such that it wants to look at X number of ms before it does anything, then the plugin has two choices as DX plug:
- Output x ms of silence
- don't output anything until it has recieved x ms of audio from the host
The UltraFunk plugins do the former. Since there is no means for a DX plug to communicate this information to the host, you get the results you are seeing.
The correct thing to do is the later. This will assure that all source material stays in sync.