Video Microwrave WIP

Discussion of modular and standalone video generating/processing techniques and associated hardware.

Moderators: Joe., lisa, luketeaford, Kent, daverj

Post Reply

Would you purchase a Video Microwrave? $175-250-ish

Poll ended at Sat Aug 17, 2019 1:49 pm

Yes
13
100%
No
0
No votes
 
Total votes: 13

wrave
Learning to Wiggle
Posts: 42
Joined: Tue Mar 31, 2015 1:48 pm
Location: Raleigh, NC

Video Microwrave WIP

Post by wrave » Wed Apr 04, 2018 1:49 pm

[placeholder post for one day when i'm actually finished]

wrave
Learning to Wiggle
Posts: 42
Joined: Tue Mar 31, 2015 1:48 pm
Location: Raleigh, NC

A Year In The Works

Post by wrave » Wed Apr 04, 2018 1:53 pm

The Video Microwrave started on the 4th of July 2017. I'd had a Raspberry Pi on hand for a year or so. One of the things I knew that a Pi could do is run simple GLSL shaders at fullscreen using glslViewer or OpenFrame, both written by Patricio Gozalez Vivo

I knew that the Raspberry Pi didn't have analog inputs, but I'd bought an MCP3008 at the recommendation of an Adafruit tutorial. So it seems plausible: I could run a GLSL shader and send control voltages to it. Computer-generated visuals that still bear a kind of synthesis by the simple fact that GLSL is like mad science math wizardry, and the results are often beautifully complex, never repeating, and infinitely expanding in to the distance.

I didn't have any trouble building a proof of concept. After some
well-deserved social media mystery, later that afternoon on July 4 2017 I had a shader-based video synth.

I've been at it for a total of one year now, and just last night I had a bunch of electrical components on my desk, about to put together the first real eurorack-powered prototype, and I thought I should let the rest of the wigglers know that I'm out there making this thing.

Demo 1

More pictures and stories and screenshots coming later. I'm trying desperately to finish it to a point where I can demo it at Indie Grits, Columbia SC April 12-15

User avatar
NathansModular
Common Wiggler
Posts: 91
Joined: Thu Feb 15, 2018 10:03 pm
Location: Nashville

Post by NathansModular » Wed Apr 04, 2018 2:56 pm

I can think of many different ways this could be implemented. As someone with graphics programming experience that sort of scares me. If its just a hand full of shaders with the modulation set up in a limited way I would be interested at 150-200 provided it was a quality product in other ways. If you manage to make me feel like I am interfacing with glsl away from a computer through a plethora of options and enough meaningful modulation I would pay 400. It wouldn't even have to be extreme just enough that I experienced the machine as having infinite potential and could spend a significant amount of time getting lost in it to justify the cost. Frankly I might even pay more if it was a engineering marvel.

Regardless put me down for one I am good for it. I think you had a brilliant idea.

Don't forget me when it comes time sell unit 2
Last edited by NathansModular on Wed Apr 04, 2018 3:13 pm, edited 1 time in total.

User avatar
BTS
Common Wiggler
Posts: 139
Joined: Sun Jun 14, 2015 12:24 pm
Location: USA
Contact:

Post by BTS » Wed Apr 04, 2018 3:11 pm

Neat!

Would an upgraded Pi like the new B+ give better performance or more shaders?
My site with links to my work: Box of Tapes Sound Recordings

Twitter @BTS0

wrave
Learning to Wiggle
Posts: 42
Joined: Tue Mar 31, 2015 1:48 pm
Location: Raleigh, NC

Post by wrave » Thu Apr 05, 2018 9:51 am

BTS wrote:Would an upgraded Pi like the new B+ give better performance or more shaders?
Yes absolutely. I'm pretty limited right now. Anything that needs to loop over multiple iterative passes doesn't really fly. And I've actually gone the opposite direction and picked up Pi Zeros. But what's really interesting is that the GPU power of the Pi 2B I'm testing on and the Zero is pretty much identical. Better Pi models don't always guarantee more graphics processing. And just like with a video game console, there is a wealth of beautiful things that can be done with a "weaker" GPU. I'm excited to explore a lot of the things that we're currently using LZX hardware to do, but in a single thin module.

Somewhat related, there's that whole cryptocurrency gold rush nonsense going on right now, so the cost of GPUs has inflated well beyond reason. I'd really like to squeeze a lot of life out of a cheap chip first.

That said, there's got to be a beefy GPU out there that I could use instead of a whole Pi. I know nVidea started making prototyping single board computers. Because, damn, some of the shaders I've been trying out run better on my iPhone.

And yes, I've done all my homework. standard rca/rgb video input should be possible as well; i'm going to hold off on adding any more features or enhancements until the basic hardware and software are working.

One of my last frustrating hurdles is getting the application to start as soon as the operating system boots. You'd be amazed at how hard that is to do. Lousy computers... We should ditch all this nonsense and go back to the good old days before... oh wait.

User avatar
BTS
Common Wiggler
Posts: 139
Joined: Sun Jun 14, 2015 12:24 pm
Location: USA
Contact:

Post by BTS » Thu Apr 05, 2018 1:38 pm

wrave wrote:
BTS wrote:Would an upgraded Pi like the new B+ give better performance or more shaders?
Yes absolutely. I'm pretty limited right now. Anything that needs to loop over multiple iterative passes doesn't really fly. And I've actually gone the opposite direction and picked up Pi Zeros. But what's really interesting is that the GPU power of the Pi 2B I'm testing on and the Zero is pretty much identical. Better Pi models don't always guarantee more graphics processing. And just like with a video game console, there is a wealth of beautiful things that can be done with a "weaker" GPU. I'm excited to explore a lot of the things that we're currently using LZX hardware to do, but in a single thin module.

Somewhat related, there's that whole cryptocurrency gold rush nonsense going on right now, so the cost of GPUs has inflated well beyond reason. I'd really like to squeeze a lot of life out of a cheap chip first.

That said, there's got to be a beefy GPU out there that I could use instead of a whole Pi. I know nVidea started making prototyping single board computers. Because, damn, some of the shaders I've been trying out run better on my iPhone.

And yes, I've done all my homework. standard rca/rgb video input should be possible as well; i'm going to hold off on adding any more features or enhancements until the basic hardware and software are working.

One of my last frustrating hurdles is getting the application to start as soon as the operating system boots. You'd be amazed at how hard that is to do. Lousy computers... We should ditch all this nonsense and go back to the good old days before... oh wait.
Neat! :sb:

I can't wait to see what comes next.
My site with links to my work: Box of Tapes Sound Recordings

Twitter @BTS0

wrave
Learning to Wiggle
Posts: 42
Joined: Tue Mar 31, 2015 1:48 pm
Location: Raleigh, NC

Nearing the end of the prototyping stage!

Post by wrave » Tue May 15, 2018 10:00 am

Some fun progress videos



I'm working on the PCB layout, finalizing the BOM, getting a hand on finalizing power from a fellow DIYer, and plenty of twiddling of the software. It's stable enough that I'm comfortable enough saying that this will definitely be (self)produced, as early as this summer.

https://airtable.com/shrhn5OMtaoolS5Vo

Lento_Zoom
Common Wiggler
Posts: 111
Joined: Fri Nov 04, 2011 6:03 pm
Location: Portland, OR
Contact:

Re: Nearing the end of the prototyping stage!

Post by Lento_Zoom » Wed May 16, 2018 5:20 pm

wrave wrote:Some fun progress videos



I'm working on the PCB layout, finalizing the BOM, getting a hand on finalizing power from a fellow DIYer, and plenty of twiddling of the software. It's stable enough that I'm comfortable enough saying that this will definitely be (self)produced, as early as this summer.

https://airtable.com/shrhn5OMtaoolS5Vo
Keep going! This is looking more and more awesome!

moppe
1-Post Wiggler
Posts: 1
Joined: Fri Nov 30, 2018 9:25 am

Post by moppe » Tue Dec 04, 2018 6:45 am

Hi, how is the project going on? I just finished a small Vidboi synth for OpenGL ES- shaders. But it would be great, to use the device like a digital VST-like effekt with Video input and quality Output. Common USB- capturing devices are usually too slow, at least el cheapo devices that i tested. And also the composite out of the RPI is really shitty. How did you deal with this?
At least for the input there seams to be some ongoing effort to enable the CSI-2 Camera interface for this purpose, but there isn't a user-friendly board avaiable for this and the commercial Lintest- Boards are also just partially supported, closed source and way too expensive (at least for me). Maybe someone with more HF-Signal experience is interested in creating an 8ch- Input for 4:2:2 Composite, S-Video and YPbPr Signals to the Pi?
If you are looking at the ADV7281 Datasheet there are a lot of possibilities ;)

wrave
Learning to Wiggle
Posts: 42
Joined: Tue Mar 31, 2015 1:48 pm
Location: Raleigh, NC

Post by wrave » Thu Dec 06, 2018 2:19 pm

Hey gang. I'm still at it. In fact, so much has been going on that I launched a Patreon https://www.patreon.com/wrave/posts

To answer your questions, moppe: the module I'm making is aimed at being simple. If what you want is "video VST" that's "quality" you are likely better off saving up for the upcoming Erogenous Tones Structure.

I am thinking about video inputs through the camera connector, but that would come later.

How am I dealing with the "shitty" composite out? By using the not-shitty HDMI.

wrave
Learning to Wiggle
Posts: 42
Joined: Tue Mar 31, 2015 1:48 pm
Location: Raleigh, NC

The G2 Port

Post by wrave » Thu Jan 17, 2019 12:31 pm

Also posted here: https://www.patreon.com/posts/24043714

Hey folks. Small update this time—I've been way more focused on trying to get myself a new web design job. At the same time, figuring out the business-side of creating modules for you all. Probably sunk a few hundred hours into learning KiCAD as well. 2018 was full of video synth work. Trying to retroactively collect those journeys on my Patreon for those that want to read about them.

Anyhow, today, I have an update about analog video outs. My brother is much more an expert on the subject so I asked him to take the wheel. We've finalized that the MicroWrave Video will have a mini-DIN9 port. I still recommend you use the digital signal via HDMI, but for anyone wanting to take the signal further through analog, they'll have all of the options available: Interlaced or Progressive Composite, YPbPr Component (compatible with visual cortex)

You may have seen this connector before if you owned a Sega Genesis Model 2, and it's no coincidence. The Mini DIN9 connector, aka the "G2 port," was adopted by the retro gaming community as the new standard for analog video ports on small upscalers—the whole reason my brother knows so much about analog video. As a result, the A/V breakout cables are easy to get your hands on; and the jacks are being brought back into production. I really like this solution because it means I can keep the MicroWrave Video small and offer all the various kinds of AV outs without the need for an expander panel that I suspect not too many people would ask for anyway. And if you ever do need it, it'll just be there instead of using more HP.

This addition is going to slow me down in finishing the circuit, but thankfully most of the circuit designs are already out there in the retro gaming community, and I've got my brother to help me navigate.

Post Reply

Return to “Video Synthesis”