BEBDigitalAudio

Do not worry, the STM32F4 DISCOVERY is not obsolete. ST is just replacing the chip on the board (with a VG version), but the rest of the board is exactly the same. By the way, some distributors are also confused because they already sell the new version with the old name (a friend of mine bought one and he got the new version but it's still the old name on the invoice)

Hi Zam, you are making a very confusion between endpoints and sessions (this is a common mistake in RTP-MIDI world. Some are even playing with that to try to fool people, isn't it Hugo? ). The confusion comes also from Apple's driver which mixes the names between endpoint and session). When you see that you create a session in Apple's driver, in fact you create a endpoint, not a session. Session will come later A endpoint represents a MIDI path between the RTP-MIDI network and the application. Basically, this is where you will see the MIDI IN and MIDI OUT streams. In RTP-MIDI, an endpoint is represented by a UDP port, most of the time it's port 5004. The KissBox OEM board supports one endpoint (while the MIDI2TR or CM-MIDI interfaces has 2, if you look to KissBox website). if you take the Behringer X-Control, it also has one RTP-MIDI endpoint. The sessions are also MIDI paths, but this time between two endpoints. For example, when you click on Connect button on the Apple's RTP-MIDI control panel, you tell the local endpoint to connect with a remote endpoint. The RTP-MIDI protocol uses the session mechanism for that. Let's make a practical comparison with phones : the phone in your house is an endpoint, with an input (in which you speak) and an output (from which you listen). The telephone in your neighbour's house is also an endpoint. When you call your neighbour, you dial his number and this establishes a session through the telephone network between the two phones. When you hang up the phone, you close the session. In most cases, a telephone can only manage one session (only one person can call you at a time). But you have certainly already use telephones for a group discussion. In that case, you can hear multiple people talking to you (even if you still have only one loudspeaker). When you speak in the microphone, all people in the call will hear you. You can not decide to speak to only one person. That's exactly the same thing with the KissBox RTP-MIDI OEM board. it has one endpoint, and this endpoint can accept up to 16 sessions in parallel. And each session is a MIDI stream, with the 16 possible MIDI channels (the sessions have nothing to see with the MIDI channels, it's just an accident that you have 16 of each). In other terms, the RTP-MIDI OEM can discuss with up to 16 remote devices. But when two devices send MIDI data on channel 1, you can not distinguish who sent what, because they all appear on MIDI Channel 1. This mechanism is called MIDI Merging, and you get the same result when you use a MIDI Merger device. In the other way, when you send something via a session, all remote participants connected to this endpoint will receive the same data, exactly like a MIDI THRU box. Technically, the KissBox OEM is able to report the session number on which it receives MIDI data from a remote partner. So you could identify the remote partner. However, this has not been implemented in MIOS nor in KissBox firmware for now. I discussed a little bit about that with Thorsten, but we never implemented it. Now, it may be the good time to implement it finally, if you really need this. But we have to check with Thorsten if the control software will be able to deal with that. Sorry for the long text, but I wanted to be as accurate as possible. if something is still not clear, don't hesitate to come back to me Benoit

Aaaaahh, I think I understand now... As a C programmer, array index starting at 0 is natural for me. The RTP-MIDI port is SPIM0 in MIOS32 (like first USB is USB0) in MIOS source code. But I think that NGC has a "human" approach, where first is "1", not "0". That may explain what you see. Any NGC expert who can confirm ?

Hi Zam, I am not sure of what you mean by "recompiling the code". I am not a specialist of MB_NG and the NGC, but I am a bit surprized that it does not take into account the SPI communication by default. The MIOS is fully transparent to RTP-MIDI communication, the MBHP can even be updated through the network (as MIOS Studio sees the RTP-MIDI interface as a normal one - which is the case by the way. RTP-MIDI is fully standard MIDI, after all ) Technically, the RTP-MIDI communication is available via SPIM0 handler. The code is exactly the same as with UART and USB MIDI ports, except that it shall be routed to SPIM0 rather than UART0, UART1, UART2 So, if the MB-NG does not use SPIM0, then yes, you should recompile it (I think there are many MB-NG specialists here who can help you with that and maybe recompile it for you) About the second question... I assume you are using the Mac version of the KissBox Editor, and then you see an endpoint located on port 5006. I have to say that it's a bug of the Mac version of the KissBox Editor (arrgh, yes, there is a bug ). The RTP-MIDI OEM has only one endpoint (which supports multiple sessions in parallel), but the Editor displays a non-existing one. That's why it does not work. For the last point, (the NG1.036), it's something for my friend Thorsten... Only him can tell if he will do it or not Benoit

Hi Zam, the choice between USB, DIN MIDI and RTP-MIDI (via SPI) is just by selecting another index (it's something like SPIM0, I can't remember exactly without some source code in front of me) Thorsten has provided also a kind of basic demo program in the source code tree, where you see the RTP-MIDI OEM acting as MIDI THRU. The rest of the code is exactly the same as for other interfaces. The only trick you have to know is that you must activate the RTP-MIDI in MIOS when you install it for the first time (using the monitor - but I think you have done it) Tell me if you can't manage to do it with the previous advice, I will look in the source code I use when I will be back home Benoit

Will be a pleasure to meet another french fanatic if you come (By the way, I live in Valenciennes...)

That's very, very strange... and not logical at all. Maybe your Windows installation is corrupted or something like that (or typical in Windows world : there are many DLL missing if you made a "light" Windows machine) A long time ago (2008), I made a VST host similar to that, called the Expandium. There a few running over the world (it was not a big success). It was a PC running a true XP, and I got a few problem reports similar to that. The reason I found was that many plugins are coded with a given Windows set of files (like the VC runtime). If you are lucky and the files have been installed by another application, the plugin start. If not, the plugin fail. And since the Expandium was not intended only to run VST host and not external software, the probablity to install such libraries "by accident" was low. The Axoloti is a very nice design, I really like it. And the programming environment is wonderful. Johannes (the designed) made an excellent work. I am working on my side on something similar but based on MIDIBox and Arduino "hosts" (with an external true DSP), I hope to be able to present something very soon (my plans are to present the prototype during the SynthFest in Nantes this year)

Take care, you have only four MIDI outputs on the MBHP. You will run out of ressource if you continue (just joking) Probably because of a missing DLL or something like that. Or the emulated DLL by Wine is not compatible with original one coming from Microsoft

(Got notification this time, but too busy to answer immediately ) Interesting links, Colm. I stopped to look for the plugin packages since a long time, convinced that nobody were selling them anymore. Clearly not the case... 85 euros is a very good price for the whole package. I paid twice the price when I bought my first ASX a few years ago... If I would need to give a personal evaluation to these plugins: - Prodyssey : 9/10 - Pro12 : 9/10 - FMAgia : 7/10 (sounds good, but I prefer myTX81Z for this kind of sounds). Deserves experimentation and can give you good surprizes (the factory sounds are not good) - Drums&Bass : 3/10 (clearly something they delivered in emergency, because they promised it). Easily beaten by a PC running ReBirth... (and I don't like ReBirth ) Now, I would recommend to be careful about these "packages". When you buy them, you just get a voucher code, not a file or something like that. You then need to connect to UseAudio website, create an account, provide them the voucher code and the serial number of the ASX or the Plugiator. Their server then sends the file to the plugin manager software. So first, you should check that the plugin manager works on your machine and is able to see the server. The plugin manager is based on a very, very old version of Flash, and it simply does not work anymore on recent versions of Mac OS or Windows (ok... it never worked on Mac...) The last time I played with the plugin manager, I had to remove the last version of Flash and force an obsolete version to install... All of this manually... So rather funny to do... (but very cool to be busy for a whole rainy weekend) I don't like this machine. It's a Linux computer with a Windows emulator (Wine) to run VST plugins. It fails with almost all commercial plugins (because it does not support the copy protection mechanisms) and many freeware plugins act buggy on it. The Receptor from Muse is much better... but the price is also much higher. And as you said, the CPU is rather on "low end" side... giving high latency (15ms is common) Yes, it can work with 5V too. The local switching regulator on the ASX turns it into 3.3V and 1.0V (needed by the DSP). As far as I know, there is no 5V component on the board And, no, the USB and the MIDI OUT are completely independant. They did that clearly on purpose, to avoid "sniffing" the System Exclusive when you load a new firmware in the Flash

Excellent I was pretty sure that it would work with 9V, but I did not make the test. But I think 9V is really the maximum. When I look to the datasheets of the Intersil regulators used on the board, they say 10V max on input. The CME keyboards have a 7809 regulator, so there is no risk, but I would be very cautious with chinese "regulated power supplies"... (I have one here which is a killer : there is a slider on it to choose 6V, 9V, 12V output... but it's delivering 15V on the 9V position )

You really want me to reveal the secrets of the DSP loading or what... Got it too this time

I just checked the UseAudio website... it's even older than what I wrote : the last news is from 2009 About getting the synth, it still appears as an active product within iCon catalog (http://icon-global.com/product/x-synth/). I can still find some on Amazon and from some vendors in France. Nnow, they are probably selling all they have in stock, nothing else (but the price is incredibly high : 185 dollars each on Amazon, 80 euros for the French vendor - I got mine at 30 euros ) And iCon never, never provided any link to download the optional plugins (I think they never got the license from inDSP or something like that), so the X-Synth will live and die with the three base plugins. Of course, if one day, all of them disappear (or as Shuriken said, if their software does not work anymore on the last OS... which basically the case already : just look to the OS compatibility list on the iCon webpage... it's scary : only supported by XP and Vista 32 bits...and they forget to say that it also requires an obsolete version of the Flash player which is not supported at all) , then I will consider these boards as abandonware... and do what is necessary to share what I have

Hi Colm, my best wishes for 2016 I am almost happy to know that I am not the only one experiencing issues with the notifications Yes, they are using the AT in parallel with CC (and not only for the effects), because they went over the limit of 128 parameters... More precisely, they split the parameters betwen CC and AT even when there are still empty slots in CC (look for the B4000 parameter table, it's a typical example) Officially, UseAudio still exists, but their website did not change since 2011 or 2012. I contacted them one time two years ago, and they forwarded me to inDSP (the designer of the card in India), because they simply sell the boards. They did nothing in terms of software development (as far as I understand). And at this time, the contact I had at inDSP explained me that there was only one guy remaining who know the details about the ASX/Plugiator/X-Synth (which are exactly the same devices). I think that now, there is no more people at all working on that. I submitted them some work I did (like a VST plugin to control the ASX in place of their ugly and buggy Flash applications...), they thanked me and told me that all development were stopped. Apparently, it is still possible to buy vouchers and download new synths from the UseAudio website, but simply said, nobody sells the vouchers anymore. I can tell it still works, because I have been able to use one of the vouchers I have here to download a synth I lost. Technically speaking, the ASX is completely running over the 21364 DSP. The microntroller (ATMega32) does nothing at all in the sound synthesis control (compared to many other synths where envelopes or LFO are typically computed on an external processor while the DSP deals with the audio). On the ASX, the ATMega32 only receives MIDI from the expansion connector (true MIDI current loop at 31250 bauds) and the MIDI from the USB chip. It also reads the SPI Flash sector per sector and uploads the binary data for the program into the DSP via SPI. Period. The pain is that inDSP has encrypted everything within the Flash and also on the MIDI USB, probably to protect their intellectual property (however, this protection is incredibly easy to defeat with a simple logic analyzer which costs 30 euros... ). Personnally, I do not want to steal their software, I am also a programmer who sells programs and I see no reason to give for free somebody's else work (that's why I said earlier that I will not transmit the binary images I have made). I use this as a backup (I already lost two ASX in the past because of the microcontroller, becoming suddenly unable to reload the DSP). As I said, I started also to tweak one of my ASX, to load DSP images over RTP-MIDI (it takes 2 seconds to load the DSP on RTP-MIDI while the USB takes more than 3 minutes ) It's not an easy task to do (and you need a Visual DSP license to create the SHARC program), but if somebody is interested, I can give the details. Benoit

Hello Colm, sorry, I did not come on the forum since some time (extremely busy with some new designs for my company... and I did not get any notification even with the "notification = ON"...) Personnally, I use a SD card attached to the SPI to store and retrieve the patches. It's not complex to do: the patches are just a list of Control Changes and Aftertouch parameters (yes, they use Aftertouch to pass over the 128 parameters limit.. rather than using the standard MIDI method of NRPN, which would allow to have 16384 parameters...) Tell me if you need help for that Benoit

Salut Mickael, désolé, je n'avais pas encore vu ton message (et il faut dire que je suis tellement occupé que je ne passe plus trop sur MIDIBox) La principale question à se poser avant tout est : qu'est ce que tu veux automatiser sur ta console? La réponse à cette question va définir le nombre d'E/S dont tu vas avoir besoin. Pour répondre ensuite à ta question: non, la commande d'une matrice de LEDs ne permet pas de commander des relais. La matrice de LED utilise un multiplexage, et les relais ne vont pas réagir correctement, car il leur faut une commande "stable" (le problème étant que l'alimentation des LEDs se fait avec des impulsions dans une matrice, et elles ne sont pas assez longues pour coller un relais) Pour augmenter le nombre d'E/S, tu pourrais effectivement utiliser plusieurs cartes STM32, mais ce serait un peu idiot (plusieurs programmes à écrire en parallèle). Je te conseille plutôt de regarder du côté des "I2C expanders" qui sont des composants spécialement créés pour ça. Tu les branches sur le port I2C du MBHP, et tu peux facilement augmenter le nombre d'E/S. N'hésite pas si tu veux plus d'infos, ce sera avec plaisir Benoit

The ASX works between 5V and 9V (the CME keyboard provides 9V). Take an extreme care about this voltage: the integrated circuits for the DSP power supplies (3.3V and 1V) do not accept more than 12V as MAXIMUM LIMIT. So use a regulated power supply, or you may destroy quickly everything. The X-Synth is powered on 5V by the iCON keyboards, but since it's 99% the same board as the ASX (especially voltage regulators), I think it would support 9V too (but I did not make the test, all my boards are powered on 5V) A last remark for Pasbel : take care about the power consumption. The ASX and X-Synth take around 250mA on 5V. You may reach quickly the capabilities of a 5V regulator on an Arduino board

Hi Pasbel, the ASX and the X-Synth are using MIDI signals (5mA current loop), not TTL. So you need to use signals from a MIDI IN connector. Benoit

Yes, that's exactly the first reason I wanted to build a synth around this board. It was mainly designed to be used in a master keyboard but connected to a PC via USB. Honnestly, this board has incredible sound (the Minimax and the optional P12 being the best imho), but the way they designed the integration within DAW is... hummm... And their VST plugins were absolutely horrible to use (just imagine that the first version of the plugin manager was based on Flash :wacko: ) Oh, there is a misunderstanding here. The way I am using the board is still involving the normal firmware (the ATMega on the board loads the DSP locally, the DSP is not loaded by the MBHP). The MBHP is used for controlling the board via MIDI/RTP-MIDI and help edition. There is no copyrighted code involved anywhere What I explained earlier in the post is another topic on which I am working since a long time. The idea is to disconnect the ATMega from the SHARC DSP and load custom DSP code in place of the microcontroller. But this is pretty hard to do (there is a SMD quad resistor to unsolder, and this is not something for beginners) And I would not distribute any original binary file for the DSP (it is not needed to reload the DSP anyway, since the edition tool provided by UseAudio or iCON allows to reflash the board if needed), and I will not provide the binary files of the optional firmware, sorry :zorro: Seriously, since I have a licensed VisualDSP toolchain (needed to make custom code for the DSP), my idea is to use the board as a "ready to use" DSP with any firmware written by me or other VDSP users

As you can see on the display, you can change the synth model (so yes, you can switch between Minimax, B4000, Lightwave, etc...). And I can edit parameters too (even if this functionality is still very basic for now, because of the multiple differences between the different synths. Moreover, my ASX is fully loaded, I have all the synths in it : Prodyssey, P12, FMagia and Drums&Bass, each of them having a completely different control matrix...) To give you a little bit more detailed explanation about my approach (since you can't see the push buttons. In all of the synths I design, I have 3 modes : PERFORMANCE, EDIT and UTILITY. Here, with the performance mode, you can choose the synth model (using the PARAMETER+ and PARAMETER- push buttons, then validate with ENTER button). The DATA+/DATA- buttons allow you to load a program. In EDIT mode, you have access to "quick edit" in which you can change directly a parameter value using the 8 potentiometers. You have a set of pages for direct access to the most common parameters (cutoff, resonance, EG, etc...). All "specific" parameters are reached using a menu on LC display. I am preparing a small video to show how this is working (and to hear some sounds from the ASX board), it should be easier to understand

And as promised, here are the pictures of the quick and dirty prototype I made to show how to use the ASX and the X-Synth boards along with the MBHP. As you can see, the only difference between the two boards is the connector used to attach it to the MBHP. The prototype is not designed to look pretty, but just to help me to write and test the MIOS application to control the boards. The connector you see on the X-Synth is not the correct one. It should a HE10 connector, but I ran out of them, so I made a cheap adapter. The RTP-MIDI board is installed under the display, so you can control the synthesizer from DIN, USB and RTP-MIDI too. Note that there is a small control panel with 8 push buttons to navigate in the menus and 8 potentiometers, but I did not connect it when I took the pictures, to avoid a mess of cables in the middle of the photo. I will check with Thorsten how we publish the technical description and the firmware, maybe I will start a dedicated topic for it.

Great news : it's possible to get an iCON X-SYNTH for 19 EUROS !!!!! Look here : http://www.woodbrass.com/synthe-option-tiseur-icon-x-synth-dsp-p103640.html For this price, I bought two, and I have update the schematics to show what is the exact difference between ASX and X-SYNTH... Finally, it's the model of a connector which change. A JST-XH8 is replaced by a HE10 with 16 pins. No more, no less. I just needed to make an adaptor to be able to switch from ASX to iCON, the MIDIBOX firmware I wrote for my ASX did not need to be changed in any way. Be patient, I will publish tomorrow the photos of the prototype and make you listen to the sound of this marvelous little board. The firmware still needs some work, but I will publish soon a complete description of the project based on the MBHP RTP-MIDI (so fun to drive the Minimax over a RTP-MIDI network :rolleyes: ). This is probably one of the first widely available RTP-MIDI synthesizer available by the way By the way, I have updated the schematics for all members being interested. These schematics show the difference between the ASX and the X-SYNTH (first page of schematics). For those being interested, I still have a few MBHP PCB for sale, if you want to build this synthesizer. Benoit ASX_Schematics_V1_R2.zip

Hello Roel, you should first check that the MIDI ports are working at electrical leve, even if the probability to have a failure here is very low (it's just a buffer connected to UART 2 and 3, nothing magic). At least , this will remove one cause of problems (and avoid to concentrate first on complex tests with SPI) The easiest thing to do is to write a simple piece of code which sends the same MIDI message to all ports (without involving the communication with SPI to RTP-MIDI) Maybe just write something which takes data from MIDI 1 input which sends the incoming data to MIDI 1, 2, 3, 4 outputs. Connect a master keyboard or anything else and check that you get MIDI on the four MIDI outputs. To be honnest, I remember that I had to fight at the beginning with MIDI OUT 3 & 4, because I could not find the correct configuration to declare (on my MBHP, I had OUT 1 & 2 working immediately, but 3&4 remained inactive). I remembet that I had to change something in my program to let MIDI OUT 3&4 working, but I can't remember what. I will check in my notes. Benoit

Hello, I am not sure to understand what you want to do, so tell me if I am wrong. You would like to see the four MIDI ports (IN and OUT) over RTP-MIDI (like getting a kind of RTP-MIDI interface with four ports?) directly over RTP-MIDI? If it's the case, the current firmware is not intended to work directly in this way. The RTP-MIDI interface is an "autonomous" MIDI interface, like the USB one and like the four DIN ports. In other terms, you have six MIDI ports, without correlation between them (if you prefer, what happens on one port has no direct effect on any other port) If you want to get data from RTP-MIDI and forward them to a DIN port, you have to tell it to MIOS. In other terms, you have to write a "routing" software. MIOS will not route directly MIDI data from one port to another one (whatever the port type by the way) As far as I remember, the demo code written by Thorsten was doing that on one port (I do not have the source code in front of me, I can check later if you want). I think it was getting everything from RTP-MIDI and send it to DIN port 1, and vice versa. Benoit

A detail about the iCON and ASX : I just found that they do not look exactly the same (I have only ASX boards, so I made comparison with pictures of the iCON) The PCB is slightly different around the power supply area. On the ASX, there 3 connectors JST XH connectors (two XH-4 and one XH-8), while the X-Synth has a HE10 10 pins male connector The J1 connector is also not installed on the ASX, while the X-Synth has a XH-8 installed. I noticed that I made some slight errors on the schematics I published here (wrong number on connectors), I will upload a corrected version as soon as I have finishe the prototype

Wow, just discovered this : the XSynth is sold for 141 euros by Woodbrass : http://www.woodbrass.com/product_info.php?products_id=103640&af=2013 Amazon sells the same product for.... 300 euros!!! :mad: http://www.amazon.fr/Icon-208167-X-synth/dp/B003YULO76 (By the way, I discovered that Amazon performs some absolutely funny translation from iCON datasheet. For those who read French, take a look at Amazon webpage, it's absolutely crazy to read... and it's completely false!!! The XSynth is not General MIDI, and it's far, far away from 126 layers :geek: ) I also discovered this other board from iCON : the G-Synth. It seems to be a Grand Piano module, not sure it's the same PCB, but it seems to have many similarities (same DSP and RAM chip) For those interested, I am building right now a crappy demo around the MBHP NG and one of my ASX... Plugiator is coming back :devil: Benoit