Cacophony Forums
Unofficial Mackie User Forums => DL1608/DL806/DL32R/ProDX Mixers => Topic started by: metacom on May 26, 2014, 02:28:10 AM
-
Spent 6 days in a darkened room in December 2012 hacking the DL1608; not had time until recently to try and finish implementation of my alternative software. I've made significant progress; Currenty able to control the DL1608 from a laptop only no iPad required! Rewriting my code for android atm. More recently I have also implemented (all beit via a grubby hardware hack, in which I almost destroyed the mixer!) recording. Requires significant modification to work, would not recommend in future! Working on an alternative, which I plan to release as an addon board. protype expected to be in place in 3 weeks. Also looking toward a general MIDI interface..
Watch this space! Who'i interested? What would you pay for each of these?
-
Welcome to the forum. Certainly would be interested in a lot more detail on your project. Mackie's com protocol etc..
-
As there may be a commercial aaplication, I'm being a little cagey with the details at the moment. Having read some of your posts, I think there are certain things you would be interested in.
it runs uCLinux.:
Image name: bf518-2.6.36-ADI-2011R1-pre-main
Created: Fri Jan 31 18:35:34 2014
Image type: AVR32 Linux Kernel (gzip compressed)
Data size: 1075595 Bytes
Load Address: 0x001000
Entry Point: 0x1d4f84
Header CRC: 0xd631187b ... OK
Data CRC: 0xa18e827a
There are several ways of getting custom code onto it, although I've developed a "one-click" solution for custom code.
Havent touched much on the DSP side, as it's not my field. Will be comparing versions of the PSoC3 cyacd stuff to see if I can see any pattern in the changes.
In terms of the protocol, I have developed a fairly complete abstraction layer. Currently reviewing the changes between the version I was working with vs 2.1.1.
Also, the root password is kinda funny...
-
metacom;
A big welcome! I think Mackie should hire you. 8) That is a positive complement! Looking forward to more interesting developments from you. :)
-
Interesting!!
-
Thanks for the info. If this is the case (need to be careful here) then Mackie is in violation of GNU GPL 2. Seems like the legal department is getting bored and needs some exercise. I dismissed it as a choice a while ago since they do have reference to GNU licensing for the multi-logon code. Ass-u-me strikes again. Could be why they've been so closed mouth about everything. The details of uCLinux for the hardware if used in the DL is here. http://blackfin.uclinux.org/doku.php
for those interested.
-
They aren't violating it. Their "modifications" dont deal with GPL stuff. They wrote their own application on a GPL platform. What's under the hood GPL wise is not relevant to the sound part of the device.
Quote from them a while ago
"We will provide the applicable GPL source code files via CD-ROM upon request for a nominal cost to cover media shipping and handling charges.
Please send me back your shipping address and I will reply with the total cost. If you are still interested, please allow 2-4 weeks for delivery."
The toolchain and build env are available for the board straight from AD. It is not required for Mackie to present them as an easy dl.
If they suddenly decided to open up the whole thing, they could.. but why would they do that....
I have a build rig that seems to work.. it is a fairly standard embedded build. The DSP code is. understandably, not open. That's what is tasking me right now!
-
Welcome. I'm interested in anything you're doing that has some level of success. And if you create a commercial product, I'd be likely to purchase a license.
-
As I understand it if the Blackfin is running uCLinux, no mater what the source of the code is, it's based on Linux. They need to comply with the GNU GPL. Receiving it from AD does not release them from that responsibility. As a mater of fact it's now back in the main Linux code base after a brief forking. Disclosure of that fact (using uClinux) is the minimum required, yes their add-ons can be proprietary. Linksys found that out the hard way. The statement in the manual refers to the log-on code not uCLinux unless they "forgot" to mention that. I'll ask them and see what they say. If on the other hand your implying that they used uCLinux as a development platform only then what your saying is correct but also lame on Mackie's part. In any case libraries still need to be considered.
Some reading required.
Licenses
Click any of these links for further information on Software Copyright, Software Licenses, Patent Licenses, Cryptographic Software, Support, External Links and your lack of Warranty.
This is off the AD site. http://blackfin.uclinux.org/doku.php
-
By "it" I believe you are referring to the DL hardware. Can you shed some light on the header info.
Image name: bf518-2.6.36-ADI-2011R1-pre-main
Created: Fri Jan 31 18:35:34 2014
Image type: AVR32 Linux Kernel (gzip compressed)
Data size: 1075595 Bytes
I interpret this as an AD BF518 Blackfin part using Linux Kernel V2.6.36 in the ADI environment of uClinux 2011 R1 a pitifully outdated release barely supported by ADI.
The Image type AVR32 refers to Atmel's AVR32 which has nothing to do with a Blackfin AFAIK.
Some contradictions here.
-
You are quite correct on that one! It's a typo in a uBoot unpacking script I used. ((c) 2012 by Ollopa / www.isysop.com) The array for CPU arhitecture is 0-indexed, but the author has an off by 1 error id-ing the arch type incorrectly as avr32 instead of blackfin. Mea Cupla.
I'm saying that they shipped hardware, and continue to do so, running ucLinux. Updated versions of firmware contain the same. I thought GPL wise, their ass was covered, although licensing issues can be a confusing quagmire after a while!
-
Mea Culpa those typos can really get you. :) If you followed some of my early posts you'll see I'm not convinced that the software wasn't outsourced. Could explain some of the lack of legal concern, the long release dates for minimal enhancements and the lack of bug fixes.
-
Please forgive a noobs question, but when I go to offline page under Master Fader, open More, you get the Master Fader version etc etc, and under that list, you see Mandolin version 1.2. Is Master Fader written in Mandolin, or is Mandolin the GUI running on top of uCLinux?
If I am totally way off base, please accept my apologies. :-[
-
Actually a very relevant question! Mandolin is the name of the protocol itself.
-
Ahhh, thank you for clarifying. 8) I'm not a programmer or anything like that, but am always interested in the inner workings. I suppose it's a good thing that they didn't base the O/S on Windows embedded. ::)
-
Very interesting. I've said in other threads that I'd like a way to control MF with midi via the camera kit or bluetooth or some such to allow me to use a hardware fader pack like the Behringer BCF2000, if you're interested look at the Master/My Fader V3 Wants and Speculation thread (sticky at the top) for my last post).
-
My e-mail to Mackie. Answer should be interesting.
I have been informed that uClinux is at the heart of your DL series of mixers.
Please acknowledge this or correct me if I have been misinformed. Also will you
ever publish the Mandolin communications protocol used in the DL series. What
does your Linux CD contain? Please direct this e-mail to the relevant party if
you cannot answer it.
-
I know you ain't holding your breath either........ but it'll be an interesting reply if it ever comes. Nice one WK
-
Well Sam were both surprised and of course the canned answer. Here it is from Ben.
Hello,
Thanks for the email. To comply with the GPL, LOUD Technologies Inc. will provide upon written request via email or traditional paper mail the applicable GPL source code files via CD-ROM for a nominal cost to cover media, shipping and handling charges as allowed under the GPL. The CD includes any modifications we have made to the Linux Kernel allowing it to run on our specific hardware. It does not include any user space applications.
If you are still interested, please send me back your shipping address and I will reply with the total cost. Please allow 2-4 weeks for delivery.
Ben Olswang
He doesn't answer any other questions, no multitasking here. I'll send him the request just to see what they will charge. I can just imagine mods to the kernel by Mackie. :(
-
Some interesting insight to the DL software foundation and possible future enhancements for the DL by it's product manager Ben Olswang.
Mandolin is our internal communications protocol. But Mandolin is not static. As
we add new features to the DL series, Mandolin gets significantly updated;
that's why you see the Mandolin version number listed in the Tools>About>More
popover. We need to change it arbitrarily at any time which would brake anything
using an old version. It is also so deeply ingrained in the product and app that
it is not suitable for generic control by a third party app. A third party
protocol you just want to be able to fire a few messages and then look for a few
simple responses. Mandolin is not architected that way. It has extensive flows
used throughout that require an extremely complex communication layer. It works
great for our needs but it just isn't designed for third party usage.
We understand that external control from a third party is an important feature
but something like MIDI over Ethernet, OSC or OCA are much more appropriate for
this purpose and adding one of these types of things to our product makes way
more sense than releasing Mandolin.
-
So a message for Ben Olswang... You mention "midi over ethernet", that would be a way to allow third party devices control over the DL. Is this in the pipeline?
I'm not a programmer but am I being naive suggesting that midi control over MF on an iPad via USB (Camera Connection Kit) may be a very simple way of gaining access to some physical controls?
-
Sam a wireless MIDI device would be the more direct and reliable way to go rather than go through MF. Like having another iPad controlling the DL but with faders and knobs and portable. There are lots of solutions in that method out there ie. http://createdigitalmusic.com/2011/01/new-solutions-for-wireless-midi-midiosc-developers-answer-questions/
-
Agreed but you'd need the iPad to deal with the functions not available on the control surface so I thought that way might be easier to implement. I realise you'd have to modify both MF and the DL firmware so it's probably academic whether it's less work to have the midi via the iPad or direct with the DL firmware. Either way it would be a major extra selling point for the DL.
WK, you have far more knowledge in this than me, in your opinion is this possible with the current hardware and how big a rewrite would it require?
BTW, wouldn't your solution require a bespoke surface? Mine should work with any midi control surface?
-
Current hardware can handle it, unfortunately I would have to have knowledge of the current code to answer the rewrite question and that's not to be.
-
quote BenO "We understand that external control from a third party is an important feature
but something like MIDI over Ethernet, OSC or OCA are much more appropriate for
this purpose and adding one of these types of things to our product makes way
more sense than releasing Mandolin."
We can but hope ;D
-
On that one I wouldn't hold my breath. ;D
-
I guess new features on the present DL aren't going to generate loads of extra sales, a new mk2 might shake thing up a bit, if it had a few essentials like multitrack recording, plug in fx etc but from a commercial POV I don't suppose we'll see any more major updates to MF.
-
Wasn't getting updates about the thread in my email for some reason.
An update on what will be possible:
Generic control surface: Yes
MIDI control: Yes (and OSC/whatever)
Recording:TBD -- I have an approach, but it may be a mode switch type thing. either mixing/effects/etc or recording. No PoC yet. Also may brick the mixer... Hardware addon possible.
Android: Yes!
PC Control: Yep! (already working)
Ben's point re: updates to the protocol will break things, entirely true. But if I'm selling stuff that relies on mandolin protocol integration, I'm gonna update as fast as humanely possible on any update...
Will post a video later showing PC control proof of concept once I get a working camera of some sort!
-
Kinda wondered what happened to you, glad you're back. Would the CD help in your efforts? Since I have no DL code to evaluate I can only speculate on the crumbs feed to us by Mackie. The presumption that Mandolin would not be useful to third parties is preposterous. There are much smarter people out there than the ones at Mackie. Could you imagine if iPad com or audio code would break apps when adding new features. That of course leads to one conclusion that modularity and good coding practices and backward compatibility are not the order of the day.
One another subject, the boot loader indicates that about 1 Meg is used for uClinux. Do you have a number for the whole DL app? They have 32 MB of flash and 32 MB of ram to play with. Would be useful to see how much is still available for new code.
-
https://vimeo.com/96917318
Will take about 45 mins before it's ready.
re: space..
Filesystem Size Used Available Use% Mounted on
/dev/root 18.0M 4.7M 13.3M 26% /
Plenty! The core app takes up about 560k Note in the above, I have a bunch of other stuff that's not stock. so nearer 3.9M used on stock.
EDIT: No, don't need the CD. Have a working toolchain and have successfully compiled and run a few things on in already.
-
So their head honcho in Europe former techie was right about the 25% usage of resources. That leaves a lot of space for new features. The DSP in my estimation is loafing and can do a lot more before running out of steam. Are you familiar with David Schumann's work with the Behringer x32? Need to check the spelling of his name. Name corrected.
-
http://dev-core.org/?p=mixingstation
Just looked him up.. Looks cool!
I'm not familiar at all with the DSP side; trying to familiarise myself with SigmaStudio /Crosscore embedded. Way outta my comfort zone!
EDIT: Video is live.
-
Great Job on the control without Mackie's help. The UI will be the key to this for commercial apps. David has features not even possible on the B programs such as IDCA so there is plenty of room for enhancements over the manufacturers. Lots of UI ideas presented in past posts here.
-
Update:
After a caffeined fueled weekend, I now have rudimentary OSC support. Cant figure out a way to take video of my phone and iPad at the same time, as they are my only camera sources! Tested with TouchOSC on a stock Galaxy S3. Latency is phenominal. Working on porting to run natively on the DL now, and building some patches. Before I get too far down that rabit hole, I figure it's best to engage the community now regarding proposed namespace etc. Trying to keep it easy for people to make their own customised interfaces.
-
So the plan is to make a OSC to Mandolin translator for the DL?
-
WK154
Yep thats the plan! I don't think I can outdo lemur/touchOSC in terms of ui.
Update: Ported the code to the DL, written the android loader app, in testing phase now. Coming soon to an App store near you :)
So to summarise:
An adroid app will be release that enables OSC on the DL1608. A layout file for touchOSC will be included. The namespace will be documented and released. This will allow you to:
Create your own layout in touchOSC
Create your own layout in Lemur
Connect external OSC hardware.
I will also be responsive to feature request at the protocol level i.e. you want a new type of message that links faders a & b at a ratio of c;
What is still to be done before relases is:
improve the code/polish ofr release
add a lot more message types.
get feedback working.
ui for the android app.
Any questions?
:)
-
Lots of questions but for now just one. How will your code live with Mackie's in the DL. Both transparently available for differing hardware (iPad/Android) or not.
-
It will live happily side by side; Yes to transparently available for different hardware. i.e. you can simulatenously run the same mix from an osc source and mackies original app.
The app to load up the code will initially only be able for android as I don't have any means to develop for iOS right now; A coleague has promised to let my use his dev account when he gets back from vacation.
-
Sounds quite promising. Now have you approached the paranoid's at Mackie for a possible ND on the mandolin spec? I trust you are familiar with the OSC resources but for others, here is most of the non-proprietary OSC world. http://opensoundcontrol.org/implementations
-
On an aside, the real blow to Mackie would be to control the DL with the Behringer app. >:D Of course the DL is missing a few hardware functions for that. I wonder if David Schumann is up for supporting the DL? Metacom you might want to ask him.
-
Lol.. It is doable, using the same techniques, but its a giant pain in the ass...
Looking for ideas re: control intrarelationships. i.e linking fader x to fader y by function z. Comes from the mathematician in me. When I'm mixing, i would really like to define the snare level as a function of the kick, or even a function of the sum of the sum of the drum levels.. I want ideas re: what sort of funtions are desired.
-
Metacom check your PM ( Forum private messages) box.
-
Lol.. It is doable, using the same techniques, but its a giant pain in the ass...
Looking for ideas re: control intrarelationships. i.e linking fader x to fader y by function z. Comes from the mathematician in me. When I'm mixing, i would really like to define the snare level as a function of the kick, or even a function of the sum of the sum of the drum levels.. I want ideas re: what sort of funtions are desired.
In that vein take a look at Dave Rat's concept of mixing and turn the math ratios idea loose. I think your on the same path but different ways to implement control.
http://www.youtube.com/watch?v=MMMmR1u0CFk
Hope that PM contact will turn out with positive results.
-
Any news you care to share? On the live mixing for larger groups or at least a drummer in the band the mixing concept I was referring to is known as stems (sub-channel) mixing. Here is a detailed explanation used in mix-down and mastering.
http://www.digido.com/articles-and-demos12/13-bob-katz/19-mixing-tips-and-tricks.html
-
Lots of developments.. release v0.01 will be up soon. A sneak preview here:
https://vimeo.com/97994123
-
I'm not sure if you have knowledge of the following sites and the work being done.
http://puredata.info/
http://vvvv.org/propaganda
http://cycling74.com/products/max/
http://www.youtube.com/watch?v=bHuIL9wYjxA
Any news on the contact you can share?
-
Update: Ported the code to the DL, written the android loader app, in testing phase now. Coming soon to an App store near you :)
Do I understand you to be saying that the Android app is just needed to load code into the DL, and then we can go directly there with OSC? If so, is the code in the DL permament, or does it need to be loaded each boot?
If all you accomplished was to give me access to a hardware MIDI fader bank, with reasonable latency, I would be delighted. MF has all the functionality that I need. At least until it loses sync.
-
I'm not sure if you have knowledge of the following sites and the work being done.
http://puredata.info/
http://vvvv.org/propaganda
http://cycling74.com/products/max/
http://www.youtube.com/watch?v=bHuIL9wYjxA
Any news on the contact you can share?
I'm familiar with a few of these but will dig deeper.
RE: contact... had a good conversation, now under NDA, so.. umm... cant really say! Thanks for the hook up!
Update: Ported the code to the DL, written the android loader app, in testing phase now. Coming soon to an App store near you :)
Do I understand you to be saying that the Android app is just needed to load code into the DL, and then we can go directly there with OSC? If so, is the code in the DL permament, or does it need to be loaded each boot?
If all you accomplished was to give me access to a hardware MIDI fader bank, with reasonable latency, I would be delighted. MF has all the functionality that I need. At least until it loses sync.
What I'm saving is in a couple of hours I will be releasing the beta version :)
The major change in the release plan is that it will no longer be an android app. A more cross platform solution is in place as a jar file. It's freely downloadable, but actually getting the patch/my app requires $$.. Also means avoiding Google/Amazon's 30%.
Right now I've tested interfacing with TouchOSC and python on the rasPi with a hardware fader, some wires and a single capacitor. Also tested using perl. My iTunes account is on the fritz so I cant seem to grab a copy of Lemur right now, but I'm trying to work around that now. Cant post a video as I only have 2 camera devices, which both need to be filmed.. short of an intricate series of mirrors I cant see how to achieve that!
Untested, by should probably work out of the box:
Lemur
C++/C/Python/Perl/node.js/javascript/MaxMSP/puredata etc etc etc...
What I'm saving is that any OSC implementation can contorl the DL1608 now. Android, arduino, iPad, iPhone, Mac, Windows, Linux, RasPi. Could also do serial, but requires dissasembly. Heck, bluettoth could even work! Have some way outta left field ideas for a way to record, ultimately, unlimited channels in a novel way. Just need to find someone with a hell of a lot steadier soldering hand than me!
As I hope you can understand, there is a charge. I've decided to set this at $49.99. This will allow me to stop freelancing for a while to solely focus on the development of this. I can see some really cool things happening with this. Plus, it's a $1000 mixer.. what's another 5%? :)
Case in point, it took 20 minutes from idea to implementation for the raspi fader demo. (excluding the time it took to find the famn fader, which would make it 2 hours...)
The major thing NOT working in the v0.01b release is feedback/2 way comms. i,e, there are no signals back from the 1608 yet, so if you change something on the iPad, it will not be reflected in touchOSC etc. This is the next major feature, and will be released in v0.02 . This means it is not production ready, however I'm lloking for feedback and bug reports as at this point the code seems stable enough.
Support wise I will be watching these forums like a hawk.
Time to put on the coffee.. watch this space!
-
Check your PM.
-
Hmm... asking alpha testers to pay you 50 bucks for the privilege isn't likely to net you many...
-
Ok the feedback so far on the price isn't good.... How does this sound?
Alpha price of $15, will full beta and final access.
Beta price of $25 with final acces.
Final release at $35.
P.S Confirmed Lemur and node.js work fine. Wrangling with pd install... waiting for another arduino in the mail, will see what I can get it to do..
-
Release time!
http://app.metacom.gd/
Feedback and support in this thread. WKD check your PM!
-
TBH, I'd love this if it'll work with, say, a Behringer BCF2000 but I'd be seriously reluctant to pay a fee for a beta release, never mind an alpha version (I've been burned on so called "final" releases before now). And I would expect it to have templates for BCF2000, X-Touch and others so I didn't have to do any setting up. Given the above $30 is would be acceptable for a full working and proven version but only if I could try first and be sure it does what it claims.
-
As per the video, a hardware fader can be used with a raspi and a simple circuit. a single raspi could support up to 16 faders with this method, although there is some jitter. A proper ADC would do a lot better job.
In this case, I used the circuit from here
http://www.raspberrypi-spy.co.uk/2012/08/reading-analogue-sensors-with-one-gpio-pin/
(swicthing out the LDR for the fader, the 2.2k is optional);
and adapted the code to suit my fader/cap combo. I used a 1uf cap and a 10k log fader (ALPS 10KA). The code is here:
http://app.metacom.gd/fader.py
You may need to adjust the coefficents according to your specific fader. Essentially you want a value between 0 and 1.
Requires python 3 on the raspi plus a pythonosc and raspi.gpio.
Could do up to 16 on a raspi, more with multiplexing.
-
TBH, I'd love this if it'll work with, say, a Behringer BCF2000 but I'd be seriously reluctant to pay a fee for a beta release, never mind an alpha version (I've been burned on so called "final" releases before now). And I would expect it to have templates for BCF2000, X-Touch and others so I didn't have to do any setting up. Given the above $30 is would be acceptable for a full working and proven version but only if I could try first and be sure it does what it claims.
That is certainely an option in the future. One of the major feature milestones is bidirectional midi<-> osc support, as this really opens up what is possible. I dont have access to either model right now, but I believe it's easy enough to support a wide range of things once the midi milestone is passed.
EDIT: Put up an example layout for touchOSC. Sized for Galaxy S3, but should work fine elsewhere.
http://app.metacom.gd/lay1.touchosc
-
Please remember that the DL target market is for 99%+ a non tech market. Check their advertisements. What you have here is the other 0.001% of mostly tech savvy people. See David Schumann's approach and at present with a usable app is still not charging although some of his users are in fact encouraging him to do so. I was trying to get a handle on the # of DL's in the field and to date have not seen a serial # over 0500 with an alpha of A-Z that adds up to 12,000 units (24x500) in the field. If someone has better verifiable figures please speak up.
Update: they are using two alpha fields for their serial number bringing the above total to 288,000 which I doubt they are even close to.
-
EDIT: Put up an example layout for touchOSC. Sized for Galaxy S3, but should work fine elsewhere.
http://app.metacom.gd/lay1.touchosc
That appears to be a dead link
-
Metacom any updates you can talk about or did you vanish into that NDA rabbit hole?
-
Website (still) under construction:
http://www.metacom.gd/