Float Invert??

rl · 14 Apr 2010

Bug!

or do you use trigger blockers somewhere?

McBarGig · 14 Apr 2010

No I have since confirmed that it is not a bug in the float invert, it happens the same using just using math. Perhaps the sliders this is connected to, but what is weird is that it amounts to half??

MichaelBenjamin · 14 Apr 2010

does it show the correct value if you refresh the readout? (ie deconnect and reconnect or click on connection)

oddson · 14 Apr 2010

I've noticed some trigger anomolies in the most recent update. I didn't load any of the betas and I'm pretty sure the problems are quite recent.

I've not reported as bugs yet becasue I'm still trying to isolate what the circumstances are but I've had to resort to some workarounds where I 'kick' the primitive with a trigger from further back in the stream to make sure it's updating at the right time.

infuzion · 14 Apr 2010

oddson wrote: I've noticed some trigger anomalies in the most recent update. I didn't load any of the betas and I'm pretty sure the problems are quite recent.

I am thinking triggers have changed sometime also. Try having the trigger pass through (eg have a receiver, no dead-ends).

TIP: it is a bad idea to use a Float or Int primitive; not only it can make your filesize larger compared to alternatives, but I think it uses more system memory & CPU, since the "display" uses more resources, & I'm also unsure how the values are saved. Use an Float/Int Add or Sample & Hold instead, depending on your needs.

McBarGig · 14 Apr 2010

Yeah I'm sure its a trigger issue...there is no other reason for this to be happening...?

Granted I am still a novice compared to you guys, but, I know when something is connected and a value in that connection changes, it should change everywhere, unless blocked. I am afraid this is looking more like a bug..

Throughout our development we have had to drop features, for many reasons, one being the complexity to impliment vs. its uses. This is a new feature I have been working on for quite some time and it CANNOT be dropped. Essentially what this feature does is re-arrange the guitar head schematic is a "Pseudo" autonomous way and is part of the SHRED 1.5X release. Because of the fact that you can select any one of the components in our software and combine it with any of the HEADS, there ends up being millions of combinations. 240 million in Shred 1.06 alone. In 1.5X we are safely into the billions. Because of the number of combinations and to make it easier on the user, I started by re-arranging the head schematic in a very simple way with a simple switch marked Vintage or Modern. When the user clicked Vintage it would rearrange values and the order of things to create more of a vintage amp sound. This grew into a full blown ideal that without this error, is really quite cool. The problem is, that like everything else, it has become a lot more complex than first intended(story of my life). Some things I have to leave up to the users, I mean we have done the arrangements and created the new head on the fly. But, things like output levels(voltages) etc have to be adjusted by the user. What if for example I take a knob from a 1954 type amp, and combine it with the gain section of a newer amp, like a 5150? At this time Shred has also decided to place the tonestack before the gain section? The result is a really terrible sounding amp. So, in this case the user needs to be able to adjust the output level of the tonestack, gainstage, etc...This is what this area does, it holds the output voltages of the tonestack to a useable level, so, without being able to do that, this entire section cannot be used, and, there's another month I'll never get back:)

I just wanted to clarify what this does... I suppose I should notify MALC..

Thanks guys..

trogluddite · 14 Apr 2010

oddson wrote: I've noticed some trigger anomolies

infuzion wrote: I am thinking triggers have changed sometime also.

It's not just my imagination then! I've had a few bits of sequential logic I've been playing with recently - despite all sorts of trigger order and blocking shennanigans, I just couldn't get the expected results and resorted to a state machine (mostly integer equals seemed to be giving me extra triggers from an untraceable source). Sadly, I trashed the original files after I got the new version working, but I'll try to recreate it to post up and send to malc.

McBarGig · 14 Apr 2010

Trog can you explain to me a little how to use that state machine? Thanks

oddson · 14 Apr 2010

[old link]

It's not likely needed or helpful for what you're trying to do... but it's worth knowing about for that time down-the-road when you will suddenly remember it exists and go 'ah State Machine, that will work!' 😉

If you have anything upstream with an explicit 'trigger' output that corresponds to when you want the value to finalize you could just substitute a sample and hold in place of the data primitive (which infuzion rightly points out is not the way to do things... and may even be involved in the specifics of the problem your experiencing) and fed it that trigger to ensure the result is what you want before it propagates downstream. This kind of work-around is what I was referring to above when I said it needed a 'kick' to resolve the value state.

McBarGig · 15 Apr 2010

Yeah its like if there was a way to send a mouse click message or something then that would probably be a work around. So, what you and Tom are saying is not to use the Float, instead use a sample and hold? Do I have that correct? Thanks

edit: nevermind just read it again, ok thanks for that tip. I'm gonna make those changes today and see what I get, use all sample and hold.. Thanks again

EDIT: Yeah you're right Leif, I don't see too many uses for that state machine. In what I'm doing anyways...

infuzion · 15 Apr 2010

McBarGig wrote: Yeah its like if there was a way to send a mouse click message or something then that would probably be a work around. So, what you and Tom are saying is not to use the Float, instead use a sample and hold?

No exactly; you should just bypass the Float, if you use an Add or whatever to do so is up to your schematic.

McBarGig · 15 Apr 2010

Well I went in and changed all the floats in the schematic, which was some great advice, thank you.

It did not fix this issue though. I still get exactly the same thing...So, back to square 1..

infuzion · 15 Apr 2010

McBarGig wrote: It did not fix this issue though. I still get exactly the same thing...So, back to square 1..

As a temp workaround, you have to route triggers directly. I am contacting Malc directly now about this bug.

McBarGig · 15 Apr 2010

Yeah I did too. Thanks, I'm just gonna stop working on it until the issue is solved. I have plenty of other things I could be doing anyways.