dysentery
Exploring ways to participate in a Pioneer Pro DJ Link network.
Quick Start
To watch and analyze the packets being sent between your Pioneer gear,
download and run the latest dysentery.jar file from the
releases page. You
will need a
Java runtime envirnonment
Once you have a recent one installed, you can probably run dysentery
by just double-clicking the jar file. See the Status
section for more details, explanation, and a screen shot.
🔧 If you’re looking for a library to use in your own projects, that’s what beat-link was developed for, and @EvanPurkhiser is now also developing prolink-connect if you’d like a TypeScript version.
🌟 And if you want to synchronize shows without having to write your own software, check out beat-link-trigger.
Disclaimer
This is in no way a sanctioned implementation of the protocols. It should be obvious, but:
⚠️ Use at your own risk! For example, there are reports that the XDJ-RX crashes when dysentery starts, so don’t use it with one on your network. As Pioneer themselves explain, the XDJ-RX does not actually implement the protocol:“The LINK on the RX is ONLY for linking to rekordbox on your computer or a router with WiFi to connect rekordbox mobile. It can not exchange LINK data with other CDJs or DJMs.”
While these techniques appear to work for us so far, there are many gaps in our knowledge, and things could change at any time with new releases of hardware or even firmware updates from Pioneer.
That said, if you find anything wrong, or discover anything new, please open an Issue, contact us on the Zulip stream or submit a pull request so we can all improve our understanding together.
Analysis
A major goal of this project is the Packet Analysis, which is intended to be useful to anyone who wants to write code to interact with DJ Link networks. Check out what we have learned so far, and please help us figure out more if you can!
The packet captures used to create that document can be downloaded (Sections 1 and 2, Sections 3 and 4) so you can see if you notice anything we have not, even if you don’t have any Pioneer gear to try out.
Funding
Dysentery and its research products are, and will remain, completely free and open-source. If they have helped you, taught you something, or inspired you, please let us know and share some of your discoveries and code. If you’d like to financially support this ongoing research, you are welcome (but by no means obligated) to donate to offset the hundreds of hours of research, development, and writing that have already been invested. Or perhaps to facilitate future efforts, tools, toys, and time to explore.
If enough people jump on board, we may even be able to get a newer CDJ to experiment with, although that’s an unlikely stretch goal.
😀
Getting Help
Deep Symmetry’s projects are generously sponsored with hosting by Zulip, an open-source modern team chat app designed to keep both live and asynchronous conversations organized. Thanks to them, you can chat with our community, ask questions, get inspiration, and share your own ideas.
Status
Dysentery is currently being developed as a Clojure library, because I find that to be the most powerful development environment available to me at the moment. Once I figure things out well enough here, I implement them in beat-link, which is intended to be useful in other projects: it is a standard Java library available as a package from Maven Central. If you want to hack on the dysentery source, you’ll need to learn a little bit about Clojure. Finally, beat-link-trigger builds a friendly graphical interface on top of beat-link, making it easy to synchronize light shows, videos, and Ableton Live to tracks played on CDJs.
As mentioned above, other people are implementing projects with the help of this research (and in many cases, contributing to the research itself). Nice examples include:
You can run dysentery and look at what it finds on your network by just downloading and executing the jar, though, and we hope you will, to help us gather more information!
It is already able to watch for DJ Link traffic on all your network interfaces, and tell you what devices have been noticed, and the local and broadcast addresses you will want to use when creating a virtual CDJ device to participate in that network.
Here is an example of trying that out by running Dysentery as an executable jar on my network at home:
> java -jar dysentery.jar
Looking for DJ Link devices...
Found:
CDJ-2000nexus /172.16.42.4
DJM-2000nexus /172.16.42.5
CDJ-2000nexus /172.16.42.6
To communicate create a virtual CDJ with address /172.16.42.2,
MAC address 3c:15:c2:e7:08:6c, and use broadcast address /172.16.42.255
Close any player window to exit.
It also creates a virtual CDJ to ask those devices to send status updates, and opens windows tracking the packets it receives from them. When a packet changes the value of one of the bytes displayed, the background of that byte is drawn in blue, which gradually fades back to black when the value is not changing. This helps to identify what parts of the packet change when you do something on the device being analyzed.
To further focus analysis, if a byte has a value that we expect, it is colored green; if it has an unexpected value, it is colored red. Bytes that we don’t yet understand are colored white. If you see any white values changing, that is a puzzle that remains to be solved—see if you can identify any pattern, or figure out what they might convey. If you do, or if any byte value shows up in red, please open an Issue to let us know. Bytes which are expected to contain the device name and firmware version are rendered as text rather than hex, to make them more readable.
Underneath the raw byte values there is a timestamp which shows when the most recent packet was received. As with the byte values, its background will flash blue when the timestamp changes, and fade to black over the next second, until the next packet is received.
Beneath the timestamp is a an interpretation of the meaning of the packet, as best we can currently understand it, with italic field labels corresponding to the byte fields identified in the beats and status sections of the Packet Analysis.
If you have access to any Pioneer Nexus gear, please run Dysentery and see if the results it gives seem to make sense for your equipment. So far it has only been tested with a pair of CDJ-2000 nexus players and a DJM-2000 nexus mixer. Even better, if you can help us figure out more of the meanings of the packets, or identify things that we don’t yet have right, and thereby improve the analysis for everyone, please open an Issue!
To try this, download the latest dysentery.jar from the
releases page, make
sure you have a recent Java environment installed, and run it as shown
above.
To build it yourself, and play with it interactively, you will need to
clone this repository and install Leiningen.
Then, within the directory into which you cloned the repo, you can
type lein repl to enter a Clojure Read-Eval-Print-Loop with the
project loaded:
> lein repl
nREPL server started on port 53806 on host 127.0.0.1 - nrepl://127.0.0.1:53806
REPL-y 0.3.7, nREPL 0.2.12
Clojure 1.8.0
Java HotSpot(TM) 64-Bit Server VM 1.8.0_77-b03
dysentery loaded.
dysentery.core=>
At that point, you can evaluate Clojure expressions:
(view/find-devices)
;; => Looking for DJ Link devices...
;; => Found:
;; => CDJ-2000nexus /172.16.42.5
;; => DJM-2000nexus /172.16.42.3
;; => CDJ-2000nexus /172.16.42.4
;; =>
;; => To communicate create a virtual CDJ with address /172.16.42.2,
;; => MAC address 3c:15:c2:e7:08:6b, and use broadcast address /172.16.42.255
nilTo log details about beat packets from a particular player (this was
built to help get the details of Beat Link’s BeatSender
implementation correct), bring up the device windows using
(view/find-devices) from the REPL as shown above, then evaluate an
expression like:
(view/log-beats 3 "/Users/james/Desktop/beats.txt")This causes all beats from the player 3 to be logged to the specified file, producing output like this:
Starting beat log for device 3 at Sat Sep 01 15:17:23 CDT 2018
Beat at 0.444, skew: n/a, B_b: 2 [1 @status +117, beat: 285], BPM: 129.0, pitch: +0.00%
Beat at 0.910, skew: 1ms, B_b: 3 [2 @status + 76, beat: 286], BPM: 129.0, pitch: +0.00%
Beat at 1.375, skew: 0ms, B_b: 4 [3 @status + 53, beat: 287], BPM: 129.0, pitch: +0.00%
Beat at 1.841, skew: 0ms, B_b: 1 [4 @status + 55, beat: 288], BPM: 129.0, pitch: +0.00%
To stop the beat logger (without having to exit dysentery):
(view/log-beats)To build the executable jar:
> lein uberjar
Compiling dysentery.core
Compiling dysentery.finder
Compiling dysentery.util
Compiling dysentery.vcdj
Compiling dysentery.view
Created /Users/james/git/dysentery/target/dysentery-0.1.0-SNAPSHOT.jar
Created /Users/james/git/dysentery/target/dysentery.jar
History
This research began in the summer of 2015 as I was trying to figure out a reliable way to synchronize Afterglow light shows with performances on my CDJs. I broke out Wireshark and after staring at packet captures over a weekend, I was able to identify how to track the current BPM and beat locations by passively watching broadcast traffic, which was my main goal. I still could not get a lock on where the down beat fell, because I could not tell which player was the Master.
Virtual CDJ
In the spring of 2016 I saw a posting on the original VJ Forums thread where we had been discussing this, announcing that Diogo Santos had made an important breakthrough. By broadcasting packets that pretended to be a CDJ, his software was able to get the other players to start sending it more details, including information I had not been able to find in other ways. He was kind enough to share his code, and that was the impetus behind starting this project, to consolidate what people are learning about this protocol, and make it available for other projects to benefit from.
Initial metadata breakthrough
In December 2016 I heard from @EvanPurkhiser, who had found this project, and went on to make important breakthroughs in obtaining track metadata.
Robust metadata understanding
In May 2017 Austin Wright contacted me on the (retired) Afterglow Gitter channel and told me about some really cool work he was doing. He was even gracious enough to publish a bunch of source code that I’ve been able to use to get a much deeper understanding of how metadata queries work, and to gain access to things like beat grid information (and eventually track waveform images). This is the current area of active research.
Sync control and tempo mastery
In the summer of 2018 I dug into implementing more of the protocol, so that the virtual CDJ could send its own status updates and beat packets, become tempo master and control the tempo and beat grid, as well as telling other devices to turn sync on or off, or become tempo master. Also figured out how to respond correctly when the nexus mixer told the virtual CDJ to do those things.
nxs2 and beyond
Throughout the succeeding years we continued to expand our knowledge and ability to use more elements of the protocols and data files, including new nxs2 features like colored and named cues, phrase analysis (thanks to Michael Ganss), and towards the end of 2020 we figured out how to support new CDJ-3000 features, including supporting six channels and exciting new packets with very valuable information contributed by David Ng.
Why Dysentery?
The name of this project is a reference to one of the infamous hazards faced in The Oregon Trail, a game which helped many students in the eighties and nineties understand what life was like for pioneers exploring the American West. Since we are exploring the protocol used by Pioneer gear, it seemed at least slightly appropriate. And, ok, I have a hard time resisting forced puns. Let’s hope none of us see:
License
Copyright © 2016–2023 Deep Symmetry, LLC
Distributed under the Eclipse Public License 1.0, the same as Clojure. By using this software in any fashion, you are agreeing to be bound by the terms of this license. You must not remove this notice, or any other, from this software. A copy of the license can be found in LICENSE within this project.