(the product formerly known as) Epoch

[taoyoyo]

Things are getting strange,
I'm starting to worry
This could be a case for Mulder and Scully.

[colin@loomer]

It's still definitely coming; it's just a very big, very complex program! I can assure you though that it is being worked on full-time (although sometime work on fixes and updates for previous Loomer plug-ins do take priority.) Bug fixing and finishing the UI are the only substantial portions outstanding for the first 'to the public' release; there are a few things I'd like to get done before v1 also, such as supporting the new MIDI Audio Unit standard, but I'm holding off on until post-beta in the interests of not being crucified!

[pinki]

there are a few things I'd like to get done before v1 also, such as supporting the new MIDI Audio Unit standard, but I'm holding off on until post-beta in the interests of not being crucified!

Careful what you don't wish for.............!

Seriously though I do think it's 'now or never' for this product. The wait is really long now. Kinda gonna back fire if you leave it any longer. Not crucifixion...heaven forbid...but it's clear that people are starting to loose patience with the waiting. And yes of course nobody is 'owed' anything at all. That is a truism. (But some of us have been here an awful long time now )

Happy coding.

[colin@loomer]

I absolutely agree, and am working as hard as possible to bring this to you all as quickly as possible. To be honest, the occasional "are we there yet" serves as a great reminder that you want this, and I think without that, I'd likely have lost motivation for a project that has taken a lot longer than originally anticipated.

And just to give you an idea what I'm doing literally right now: I'm fixing some issues that had snuck into the multiple time signature support, as well as tidying up the way that the overall arrangement metre works if you're just using a single time signature and tempo for the entire piece; the goal being to make more esoteric features, multiple sigs being one such, as invisible as possible to people who don't care about them, whilst still keeping them easy to use for those who wish to explore that area.

[treebeard]

Hey Colin,

Loathe as I am to keep you from your work, I would really like to ask you something software related since I feel you are the best guy to have an idea on this.

There is a program I have which uses sequencing objects for midi. It works great except for one thing. When it is running and synced to an external midi clock, every now and then from 2-30mins, seemingly randomly the objects will send out multiple very short notes very close together. It is completely smooth apart from this little stutter/hiccup effect.

What is this likely to be the result of - a poor master clock, poor slave implementation (algo) or something else?

I don't want to bother you for a detailed analysis but just what your first hunch would be! Maybe you have come across this kind of behaviour before?

[News From The Sky]

This has been an extraordinarily long development process - I'm going to call the end result Eventually no matter what Colin ends up naming it - but let's remember that it is also an extraordinarily complex piece of software being written by one person. Reactor, Max/MSP, and Numerology all took time to develop too.

[colin@loomer]

What is this likely to be the result of - a poor master clock, poor slave implementation (algo) or something else?

It could be either of what you suggested; or a dodgy MIDI driver; or something else altogether. I'd run a MIDI Monitor and try to narrow it down. I like Snoize's MIDI Monitor on Mac, although I'm sure there are equivalents on others OSs if needed.

I'd eliminate the master clock issues first, as it's easily checked: with a MIDI monitor, track MIDI Clock messages, and ensure that they are evenly spaced (in terms of milliseconds between them) when these rogue notes appear.

[treebeard]

What is this likely to be the result of - a poor master clock, poor slave implementation (algo) or something else?

It could be either of what you suggested; or a dodgy MIDI driver; or something else altogether. I'd run a MIDI Monitor and try to narrow it down. I like Snoize's MIDI Monitor on Mac, although I'm sure there are equivalents on others OSs if needed.

I'd eliminate the master clock issues first, as it's easily checked: with a MIDI monitor, track MIDI Clock messages, and ensure that they are evenly spaced (in terms of milliseconds between them) when these rogue notes appear.

Excellent, thanks Colin, I will try to monitor the clock.

The weird thing is that I use an MPC1000 as my masterclock for 5 other bits of gear (HW and SW) and this particular program is the only one which has this problem. Still, I shall investigate its clock signal to be sure.

Really appreciate the advice - its driving me nuts!

oof - I'm on PC here Cubase 5's midi monitor does not show clock messages :C Ill try Midiox ;D

[bungle]

So I don't think it's unrealistic to expect it to be "slightly" better at it...
Cheers,

Tom

Actually it is completely unrealistic to 'expect' anything
1 There is actually no proof at this point that this project even exists (Yes it does, we know it does, Loomer is not a fly by night, but the fact remains)
2 Expecting a new plugin to be able to compete with one of the most powerful modular systems ever created again is complete nonsense
3 Expect nothing until the product is out and usable, all can change

I think the word you are looking for is 'Hope' not expect

Great programming is by no means restricted to certain big companies alone. Infact, as every software developer knows, every individual end-user will have his or her own take on what defines good or bad software. It's great that you think so highly of Reaktor but for many it just isn't held in such high regard - different strokes and so forth. What others can "expect" is also up to them to decide, but based upon the sheer amount of time(lol) and user feedback so far from Colin it's fair to say this will be a much more focused program than Reaktor as regards midi sequencing. The "expected" release date however, that is another fishy kettle...

but maybe your right and it doesn't even exist except inside the head of a madman who finds great pleasure in lurking about on forums, pretending to be a developer and teasing us all over the years with his wild fantasies, why he's probably not even called Colin, why would we expect that to be true, has anyone seen his birth cert? Pacefalm

First off, i was just pointing out how stupid it is to expect anything from an unreleased product, it is utter stupidity in its most basic form, second please point out where i said that this does not exist (i didn't) or stop putting words in peoples mouths.

If i where Colin i would be seriously worried that the self entitled fanboy attitude is already kicking in without ever even seeing said product, these are always the people that are most vocal when their 'expectations' are not met.
Again i reiterate, expect nothing, he owes you NOTHING

[treebeard]

I'm very very sorry Bungle! I didn't mean to put words into your mouth or be basic and stupid or a self entitled fanboy or assert that Colin owes anybody anything.

I was just trying to make a stupid light hearted joke and support Colin by saying I believe he is capable of making something great - I guess it came out badly or something and missed the mark.

Once again, truly sorry for being an idiot, my mistake - my rubbish humour is easily misunderstood!

[ThomasHelzle]

And just to give you an idea what I'm doing literally right now:

What about making this a weekly occurrence?

Cheers,

Tom

[colin@loomer]

Good idea.

I'm currently polishing the arrange window, which is the view for creating linear clip sequences. Whilst nice and speedy when using only a hundred clips or so, once we move into the thousands and tens-of-thousands, it gets a touch slow. And whilst this may sound like an extreme case, Logic 8 - which is my current target in terms of how responsive things should be - handles this situation with ease.

With this many clips, it also adds a noticeable start-up time, something I've tried my hardest to avoid. So I'm currently profiling away, and looking at ways to optimise this. I had to perform a similar optimisation with the rendering of the step sequencers, where once you exceeded a certain threshold of sequencers all running together (over a thousand, admittedly), zooming in was a quite unresponsive. I suspect the same improvements I made there will have the same vast improvement in responsiveness here.

I've also been finalising the quantise in the arrangement window. Snapping is now both absolute and relative to the clip offset. This may sound odd, but feels just right. For example, with the quantise set to 1 bar, if you dragged a clip that started at bar 2, it would snap at bar 3, bar 4, etc. However, if you dragged a clip that starts at 1.2, it would snap at bar 2, as well as position 2.1, bar 3, 3.1, 4, etc. And this all plays nicely with multiple time-signatures, where 1 bar the 4/4 section isn't the same as a bar in 7/16.

And so it goes: polishing, refining, and improving the whole 'feel' of the UI and program in general.

[ThomasHelzle]

Ha - over a thousand sequencers and thousands of clips - wow.
Not something I would ever need - I'm a minimalist - but good to know it holds under heavy fire

But from what you describe I also gather that the scope of !Epoch has indeed gone far from the initial idea of a node based midi toolbox and now sounds more like a complete modular host with sequencers, clips, arranger and VST hosting.
I guess I have to re-adjust to that idea.

That bigger scope also suggests a stand-alone mode?

Wow, no wonder you took "slightly" longer than expected...

I wish you all the best motivation in the world then!

Cheers,

Tom

[bungle]

Problem is that the first thing people will want are audio tracks haha

[colin@loomer]

Yes, it runs as a standalone app as well.

I have already been asked about audio tracks, but I'm afraid that suggestion is definitely on the far-off-distant request list!

Good progress has been made on the arranger sluggishness though: I rewrote the way that it internally manages clips, and it is now literally a thousand times faster. Even on arrangements with tens of thousands of clips, it all responds immediately: a big relief.

Now, I just need to run through some regressions and fix what I inevitably broke whilst making this change!