Imitone -- wow! Most embarassing post I ever started

[bungle]

This guy has more reasons for why something is old fashioned/out of date/pointless/you are doing it wrong/no this square wheel is better than the round one than gol haha

[trimph1]

This guy has more reasons for why something is old fashioned/out of date/pointless/you are doing it wrong/no this square wheel is better than the round one than gol haha

Good thing I no longer drink anything perusing this forum.

[klinik]

I´m a kickstarter backer too, and I jumped because the VST version of imitone.
I run a studio, I can´t use USB mics or such, I run on Pro interfaces and the built in audio cards are deactivated to keep my system stable.
Also I plan to use imitone to double vocal takes, I mean dubbing with a s external or internal synth/sampler vocals already recorded, and thats something that only can work with a VST version inside the DAW.

For me the VST portion of it is the only option. I downloaded the beta 0.8 today, but apart to play 5 minutes with it in my laptop, It will remind unused as the other betas until a VST versiom, even an alpha one, comes out....

[interactopia]

Regarding PC versus Mac, you could say I'm just bitter about Windows' terrible audio subsystems.


Regarding plugins, a note I meant to make in the last post:

Part of my reasoning in delaying a VST version so far has been my doubts about the usefulness of earlier versions. imitone 0.5 was promising. 0.7 was viable, but flawed. 0.8, finally, is actually good (though not yet magic). So it's worth much more serious consideration...

I expect maintaining plugins will be quite a lot like maintaining additional ports -- meaning that pushing out new versions will involve more overhead. I appreciate Compyfox's suggestion about building a small VST, but the way I do porting generally involves jumping right in.

One interesting note I'm taking here is that there's good reason to prioritize the Windows plugin over the Mac one, given that the situation is more constrained for Windows musicians. I have run into trouble (audio corruption) with device sharing on Mac, but I expect buffersize and samplerate settings to help with this.

b) the home-page is so scarce on information it's not even funny
c) return policy is a bit wonky

I'm open to more detailed feedback on these points.

[bioroid]

Maybe you could look into use the JUCE library which may make it easier to release plugin versions (VST, AU, etc) in addition to desktop and mobile versions. Don't re-invent the wheel if you don't have to.

http://www.juce.com/features

[deastman]

I've just spent a while playing around with 0.8.0. I can't get it to even remotely reliably track my voice at all.

I suspect a malfunction and would be interested in more details. Are you singing in the vocal fry register (deep and throaty)? What kind of microphone are you using? Is auto-adjust enabled? What's the sensitivity mode? Is the environment noisy? If you could make a recording and send it my way, I could run it through my testbed and diagnose the issue.

I've been having professional friends test this and they've been reporting very good results other than the "flutter" effect. (A bug which induces brief notes of the same pitch on attack and release.)

I was just using the microphone built into my Dell laptop, in a quiet environment. You're correct, however, in that I have a cold right now and was singing with an atypically deep baritone voice. I'd record a demo for you, but I can't stop coughing now! I did spend a while experimenting with all the different parameter combinations, but never found anything to work very well. I was surprised that I didn't get better results out of it, since as I mentioned, I had pretty good success with my Nord Modular years ago. That patch was pretty simple, relying on the built-in zero-crossing frequency counter, combined with several stages of carefully tuned amplification and filtering. I can't recall if you've described your technique in any detail, but I'm guessing that it is based around a resonant filter bank.

[ThomasHelzle]

Regarding PC versus Mac, you could say I'm just bitter about Windows' terrible audio subsystems.

Another thing you won't have to worry about with a VST...

Like klinik above, the internal sound in my PC is deactivated - I never use it.
I have good microphones that are XLR based and connected to my audio interface (Focusrite 18i20). Everything is low latency and working very well via ASIO...

The vocal backing klinik talks about above is a really strong argument IMO. This is one of many extended uses over the "sing in what you can't play" that musicians will come up with. And this is a major pain to do with a standalone (even on Mac).
For example: I use JamOrigin Midi Bass not to replace my Bass, but to add interesting vibes to it:
First in the device chain in Bitwig Studio I place Midi Bass to translate audio to midi. It allows to output both the original audio AND the generated Midi notes.
Next in line is for instance Native Instruments Prism FX, a tool that uses "Modal Banks" of resonators on the incoming audio and you "tune" them via the incoming Midi.
And this allows to enrich the Bass sound tremendously with fascinating tones...
This would open up many uses for imitone as well...

Honestly I personally have a hard time even finding one good reason to not push the VST.
Sure the algorithm should shine, but what if it takes years until you reach your goal of the perfect recognition? Or you never reach it but somebody else does?

Give your backers the chance to work with what is there now in a way that makes sense.
I personally would not have backed the standalone without the VST option.

Create a prototype with an ugly GUI first, start to learn the ins and outs of VSTs, let it grow over time...

Cheers,

Tom

[Doug1978]

Regarding PC versus Mac, you could say I'm just bitter about Windows' terrible audio subsystems.


Regarding plugins, a note I meant to make in the last post:

Part of my reasoning in delaying a VST version so far has been my doubts about the usefulness of earlier versions. imitone 0.5 was promising. 0.7 was viable, but flawed. 0.8, finally, is actually good (though not yet magic). So it's worth much more serious consideration...

I expect maintaining plugins will be quite a lot like maintaining additional ports -- meaning that pushing out new versions will involve more overhead. I appreciate Compyfox's suggestion about building a small VST, but the way I do porting generally involves jumping right in.

One interesting note I'm taking here is that there's good reason to prioritize the Windows plugin over the Mac one, given that the situation is more constrained for Windows musicians. I have run into trouble (audio corruption) with device sharing on Mac, but I expect buffersize and samplerate settings to help with this.

b) the home-page is so scarce on information it's not even funny
c) return policy is a bit wonky

I'm open to more detailed feedback on these points.

To be blunt, there are experienced people here who can help this product develop in an efficient manner.
For whatever reason, the developer doesn't want to cede control.
Which would be fine but, as evidenced in the above post and elsewhere, there's way too much over-thinking and procrastination, and a reluctance to listen to or implement customers' requests.

It's the difference between the amateur wannabes and conscientious companies who actually deliver in a reasonable time.

A poor, time-wasting example of a kickstarter campaign, which makes people more skeptical about backing others in the future.

It makes you wonder what all the beta money (such as mine) has actually gone towards?

The snipe at PC audio versus Mac audio was also poorly judged, inaccurate and is a heads-up to future time-wasting problems.

[No_Use]

@ Evan,
another option which might be worth looking into for easy building and maintaining multi-platform plugins apart from already mentioned JUCE framework could be WDL-OL, another multi-platform plugin framework.

https://github.com/olilarkin/wdl-ol

http://forum.cockos.com/showthread.php?t=91811

Once set up it can build standalone, Win / Mac VST 2 /3, 32 / 64 bit, AU, Pro Tools (RTAS, AAX) plugins and even iOS format practically "with one click".

I was able to do some simple plugins with it after just a few days getting into it and that as a complete beginner in the (C++) plugin department, I think it's a quite nice framework and easy getting into.

(I have no comparison to JUCE though or any other similar frameworks).

edit:
See here for some examples what can be done with it:
http://forum.cockos.com/showthread.php?t=122276

[ThomasHelzle]

That is the kind of feedback you can get here Evan!
Awesome links No_Use.

Your life may actually get easier if you work more with pros...

And for the procrastination part:



Cheers,

Tom

[Compyfox]

Regarding PC versus Mac, you could say I'm just bitter about Windows' terrible audio subsystems.

Sorry, but it's not terrible - it's just a bit inflexible. Depending on the program you use. And ASIO never really allowed driver sharing. That is the main issue.


There are workarounds, yes. But they are not really great.

Some people for example say that ASIO4ALL can be used to link several drivers into one - but I once tried it, found it absolutely unreliable and a nightmare to set up. Still admire the effort to bridge the sharing limitation.

Rewire might be an option, but not many companies used this, and I don't know of any additional fees.

The other/most common workaround is using the PC drivers and a cheap mic, as you mentioned already. But it's like a video game where you need a specific controller to even use it/appreciate it (Armored Core anyone?!, GuitarHero/RockBand? Bemani games?). And this is the biggest turnoff in the debate here.

Just like DrumsAnywhere - I have my eye on this tool, but two major things turn me off:
a) price
b) features

In case of DrumsAnywhere, the feature I'd love to see the most (MIDI out to connect to another host) only works with the Pro Version, and that is 120USD. The demo omitted that feature. So no test is possible. That is a huge risk I don't want to take. Especially considering the driver sharing problem on a PC.

In your case, we have to merely invest 25USD - but this is the Alpha/Beta price. So I assume it will be 49USD on RC/Release, and the Prime/Studio version will be 129USD. There is no demo available, and the return policy is not mentioned on your page. Also, the advertising is heavily focused on Macs. In this case, I'd also be hesitant what to do/go for.


Summed up:
Your think-tank goes the wrong direction IMO. You want people to keep interested in your tool, then you have to change the direction. And ASAP!

Regarding plugins, a note I meant to make in the last post:

Part of my reasoning in delaying a VST version so far has been my doubts about the usefulness of earlier versions. imitone 0.5 was promising. 0.7 was viable, but flawed. 0.8, finally, is actually good (though not yet magic). So it's worth much more serious consideration...

Sorry - but that is a mere excuse.

I told you, I'm a beta tester for various companies. And even with Alpha/Beta versions that have not reached v1 yet, you can already do a lot. No matter how broken the software is. The emphasis is "getting stuff done either way".

If you change the concept on the long run, improve things, change some parameter. Well, that's to be expected during a beta (this is why these programs exist!). But you could already do things! Appreciate the workflow, etc.

A famous example on KVR Audio is the Audjoo HELIX Synth. There was a huge outcry the day it started to cost 150USD. Why? Because the beta version was so good, it was a secret treasure in the community. And the full version couldn't top that much (actually, the full version was criticized for it's GUI design, can't remember if it was upgraded on features as well).

And something that you can relate to in the video game realm:
Satellite Reign, The Last of Us, Sheltered, Wasteland 2... four games that were successful early access games and worked right out of the box. And the developers did not hesitate to improve things, listen to their customers, improved on that.

So why is it so hard with your tool?
Just because you work alone? Or you're frustrated of the Windows Platform compared to the Mac?

I expect maintaining plugins will be quite a lot like maintaining additional ports -- meaning that pushing out new versions will involve more overhead.

Yes, it will be just as much overhead as programming stand-alones for Windows, Macintosh and Linux (should you go that route). Each with their own drivers, OS irks, etc (also: 32bit/64bit, MacOS also needs "sandbox checks" IIRC). However, as mentioned a couple of times here in this thread, there are frameworks that ease up on things. Frameworks that render multiple plugin versions at the press of one single button.

The rest is down to the beta testers and early adopters to find the bugs, and you and your code documentation to find fixes/workarounds.

If it's too much workload (on either end), then involve another person.

I appreciate Compyfox's suggestion about building a small VST, but the way I do porting generally involves jumping right in.

Then jump right in! All the better.

I merely recommended to create a VST Breath Controller as test environment, especially since you weren't sure if a Plugin Version does even make sense, or works like you planned. On top of that, the added bonus would be that people get an appetizer, a tool that they can use to control strings and brass with. Such tools usually need a dedicated controller (there are USB Breath Controllers). And you could make it freeware where people only need a mic and breathe into it.

Think outside the box!

Else, the Plugin (Imitone) doesn't even need to be special, or VST3 or whatever. It just needs to work on an audio channel, analyze the incoming audio signal, and output a MIDI signal that the host can then pick up (routing of channels). Much like a gate that can send trigger signals via MIDI (Cockos Reaplugs ReaGate comes to mind).

One interesting note I'm taking here is that there's good reason to prioritize the Windows plugin over the Mac one, given that the situation is more constrained for Windows musicians. I have run into trouble (audio corruption) with device sharing on Mac, but I expect buffersize and samplerate settings to help with this.

Stop with the excuses - try it, do it!

The standalone version was a test - the test has been going on for a couple of months now. You know what is working, what needs to be changed in order to work better. You could present your initial concept at fairs and events.

Now make it usable for everyone!

b) the home-page is so scarce on information it's not even funny
c) return policy is a bit wonky

I'm open to more detailed feedback on these points.

I would write a new homepage at this point. This top-to-bottom design is good for mobile devices, but it barely tells me anything new.

For example:
- on the page are a couple of videos that show me how this thing works. But as mentioned over and over, it's focused on Mac. It doesn't show how to set it up on PC. Actually, it is still one(!) video from the Kickstarter campaign. Your other video releases are not even hot-linked.

- speaking of hot-linking, I don't know where the forum is, how to access it even. This would show user interaction and that you care(!) to get things done.

- there is no link to the Kickstarter Campain, only a mention.

- your names are just names - I can not relate who you are. So I'd love to see an about section, read up on what projects you were involved so far, etc.

- what is the name of your company?! Or is Imitone the company name and the plugin is called different? I'm confused...

- no links to social media (I found you by chance on Twitter a couple of months back, remember?), no link to write mails (I found you through an elder project and contacted you via this route).

- you mentioned several times that the prices may rise, so I'd love to see that addressed. Same with the beta version number so that users know "ah ha - it might still take a while". Basically, more info as to what is going on.

- there is no mention of a return policy on your page, especially since you don't offer a demo, and how many days we have to return our software. Also, since you go via PayPal and/or BitCoin, fees for returns are involved, no?

- a free accessible manual is always great for interested users. A lot of companies disregard that fact.


Summed up:
The main issues I have with this page, is that the core information is somewhat there. But to get to know the project more, or the people behind it, I need to heavily utilize Google or any other search engine to scratch together the information I want, need.

One of the Google results is this KVR Audio thread. And the opinions are absolutely mixed. Actually, I do have the impression that they are gearing towards "negative" as Wikipedia would say about movie reviews. So this is also adding up to the whole thing.



Once more:
I do understand that you're a one-man-army, you have a small company, you have a certain route planned out for your tool. But think of ripples in a pond... they can turn into tidal waves (I think I've been watching a bit too much Doctor Who lately).

[whyterabbyt]

Notwithstanding the fact that you're not currently presenting Imitone as a MIDI device to the OS in the first place, but rather something on the 'other' end of a virtual MIDI cable...

You need virtual midi cables

It's down to the user what type of loopback device he/she is using, unless you already provide the MIDI drivers/loopback system like a lot of other companies do lately.

There seems to be some misinformation here: The imitone beta has never required a separate virtual MIDI cable to operate on Windows or Mac.

Im installing Imitone on Windows. Ive installed it on a system without and with the provided loopback driver. There's no associated MIDI device available to any of the audio software on this system if I install without it. I only get that if I do install the provided loopback driver
So when you say 'The imitone beta has never required a separate virtual MIDI cable to operate on Windows or Mac' do you mean it doesnt require a virtual MIDI cable at all', or 'it doesnt require the user to install one themselves, because one ships with it'?

Because my question still stands in the latter case. And it doesnt work here in the former case.

When using a DAW with automatic input detection (like Logic), imitone can be used with zero set-up. It does get to be a pain with old-fashioned DAWs that don't support hotplugging, though.

TBH, comments like this make me slightly concerned that your view of the state of DAWs is filtered through what Logic does and doesnt do. To me that doesnt bode well for proper plugin development, as you're not really looking beyond Logic's capabilities in terms of the bigger picture eg

Regarding the plugin format, I am aware of no DAW on the market which can treat an audio-to-MIDI plugin as a MIDI interface for purposes of workflow.

Compare that with the fact that many DAWs (apart from Logic, AFAIK) are able to route audio to a plugin which can generate a source of MIDI that can be routed to another plugin.

Again, if you're not designing around the workflow/architecture that scales across what DAWs are actually capable of from the ground up, I think you're seriously undermining the possibility of adding it later.

All the workarounds with multiple audio devices and interface types etc, that you're having to suggest to get this working, they're all 'sorted' by using the capabilites of the DAW.

Or are you saying you want your system to be viewed as a MIDI controller but dont expect it to be used in conjunction with MIDI hardware? Or that you think its best if your software isnt a plugin, because its only going to drive plugins, not external hardware?

Every beta version of imitone can be used to drive external hardware.

Directly? How?

[Shabdahbriah]

Notwithstanding the fact that you're not currently presenting Imitone as a MIDI device to the OS in the first place, but rather something on the 'other' end of a virtual MIDI cable...

You need virtual midi cables

It's down to the user what type of loopback device he/she is using, unless you already provide the MIDI drivers/loopback system like a lot of other companies do lately.

There seems to be some misinformation here: The imitone beta has never required a separate virtual MIDI cable to operate on Windows or Mac.

Im installing Imitone on Windows. Ive installed it on a system without and with the provided loopback driver. There's no associated MIDI device available to any of the audio software on this system if I install without it. I only get that if I do install the provided loopback driver
So when you say 'The imitone beta has never required a separate virtual MIDI cable to operate on Windows or Mac' do you mean it doesnt require a virtual MIDI cable at all', or 'it doesnt require the user to install one themselves, because one ships with it'?

Because my question still stands in the latter case. And it doesnt work here in the former case

Ditto, here. Several months ago, I installed it on a non-workstation, and did get some interesting results experimenting with several mics {cheap "pro-sumer" types, worked fine, btw}, "playing" synths... a "scat" style "singing" worked for me.

[tonedef71]

And ASIO never really allowed driver sharing. That is the main issue.

I believe you are correct. To provide high quality and low latency audio, the ASIO driver sacrifices the capability to be shared.

There are workarounds, yes. But they are not really great.

I have found two ASIO driver wrapping tools that more than adequately provide the ability for multiple audio applications to stream to a single ASIO driver simultaneously.

The free option is JACK Audio Connection Kit for Windows and the commercial option is O Deus ASIO Link Pro.

With either of those two ASIO mapper drivers, Windows users can route audio from several different audio applications to the same ASIO driver (such as ASIO4ALL) and if desired, let them simultaneously play all mixed together. I used JACK for a while and thought it was a godsend. However, I now use the O Deus product because I wanted a commercially supported product with regular releases of improvements and bug fixes (JACK has not had a new release for quite some time; however, one can download the source from github and build it himself).


By the way, I just noticed that a new version of ASIO4ALL was released on November 5th!

[interactopia]

Oof, spent a few hours writing a massive reply and KVR ate it.


Research and development on new algorithms is hard and takes a large, unpredictable amount of time. Software development is easy and predictable by comparison. Ergo, porting to VST won't take nearly so long as these improvements to the system did. Lots of good points have been raised here, so I'll begin investigating a VST port alongside the next update -- but I will make no guarantees about when it will surface.


deastman -- imitone is designed to detect periodicity, meaning the inharmonic vocal fry register does not and probably never will track well. If you want to control super-deep sounds, stick to the modal register and transpose down as needed.

Zero-crossings can work on vocal fry by counting the clicks over a period of time, but you'd never be able to achieve realtime latency with that approach. imitone's algorithm was prototyped before I knew what a resonant filter was, and its key element is the SAHIR -- a harmonic isolator unit that rapidly changes its fundamental frequency to "lock onto" and "follow" tones.


compyfox -- Many very good points about the website, which I individually and lovingly addressed in my deleted reply. I have taken them to heart and copied them to my notes.


whyterabbyt --

Again, if you're not designing around the workflow/architecture that scales across what DAWs are actually capable of from the ground up, I think you're seriously undermining the possibility of adding it later.

The app exists on the principle that with nearly any DAW, it's easier to use a MIDI keyboard than a MIDI-producing VST -- especially as a novice. This isn't to discount the many conveniences and possibilities offered to more experienced users by a plugin -- but novices are very important to me.

So when you say 'The imitone beta has never required a separate virtual MIDI cable to operate on Windows or Mac' do you mean

The VirtualMIDI driver is part of imitone, requires no configuration and is not a loopback because it does not bounce input: like a hardware device, it fires its events directly to the kernel. It doesn't induce the polling latency that a loopback service does, and it doesn't clutter the OS' device list after imitone is closed.

Every beta version of imitone can be used to drive external hardware.

Directly? How?

To control any external MIDI device, plug it into your computer, then open imitone, then select it as imitone's MIDI output.



I'll be getting back to work this week and as such I may be monitoring this thread less actively.