Sign in to follow this  
Followers 0
Yannovitch

A midicontroller in my guitar !!!

38 posts in this topic

Hello !! I'm french so excuse me for the mistakes :)

I've a big project concerning a customization of my guitar, I want to have a total versatility of sound, and controlling all my effect and vsti  with knob and pot from my guitar. I think of a Midibox64E implemented in my guitar :)

All the subject concerning the pickup & cie is explained here : http://projectguitar.ibforums.com/index.php?showtopic=28305

I would know what do you think about this idea, and if it is realisable, and too if I can use a dsPIC or an AVR32 instead of a PIC for the MIOS ...

If you have some good ideas, help me please ;)

"Merci"

Yann

Share this post


Link to post
Share on other sites

Hi Yann,

MIOS only runs on PICs so if you want to go for an AVR32 then you have to start from scratch. For dsPIC, I'm not 100% sure but I would expect quite some rework will be necessary; if that is the case, for me that's clearly a showstopper.

If you want to stuff the whole parts in your guitar then I think you'll have to redesign the PCB layouts of the MB64E using probably SMD chips everywhere or almost everywhere and squeeze as much as possible everything you can.

I've not done MIOS programming yet but it's apparently very easy to get started so that should not be a big issue. I suppose that the power consumption was not a major constrain while developping the MIOS and applications so maybe small adaptations here and there would allow you to trade-off some functionality for some power reduction.

Concerning your thread on the other forum:

- If you want to do audio stuff, 12 bits won't be enough so dsPIC is does not really help there. You have to target for at least 16 bits otherwise the sound will be most probably crap. You also have to pay attention that a poor design, you can easily loose bits of resolution. I'm not a specialist in this domain but I know you can easily mess up things and end up with far less resolution than you thought.

- For the switching of the audio signals, TI has nice chips doing analog switching. They are available in small (maybe too small) packages and some are explicitly mentioned as made for audio. I however don't know if you can connect microphone ouputs to them. You can find all of them there: http://focus.ti.com/paramsearch/docs/parametricsearch.tsp?family=analog&familyId=520&uiTemplateId=NODE_STRY_PGE_T

- what's the exact reason for adding a DSP? If that's for the wireless transmission, is that really necessary? If that's for processing then you should go for more advanced DSPs directly I believe.

Last but not least, don't underestimate the importance of power supply. I mean you have to look at how much the thing will consume, how you will power it and then how you'll need to cool it. The problem here is not to design it really but more the space it may take in the end so I think it's wise to look at this point pretty early in your design as putting stuff in a guitar may not be so easy.

All in all, I think you'll have to have a pretty long thinking and design phase in your "crazy project"  ;) but if you finalize it that can really rock  :)

Best regards,

Lall

Share this post


Link to post
Share on other sites

Hello

Thank you for your quick answer :)

I have mentionned the DSP because since my last reply in the other forum, I have had other ideas, like controlling by a touchscreen a lot of parameters, like here in more small :) : http://hci.uniroma1.it/multimedialab/modules.php?name=Progetti&op=view&progetto=TouchBox&pulsante_locale=Overview

With this and 8 pot and knob, it can rock ! :)

And I need power to treat 10 input in 24bit/96khz.

I think a lot of the AVR32 because it's a good small DSP and big uC, and the developpment board is not expensive when you know that the developpment board of DSP from TI is 10x more expensive.

Concerning the ADC, I have done on paper the Nyquist and Shannon theorem applied to my problem, and effectively I must have better ADC. I think a lot of a multichannel audio ADC like this http://www.asahi-kasei.co.jp/akm/en/product/ak5365/ak5365_f01e.pdf or this http://focus.ti.com/docs/prod/folders/print/pcm4204.html, I need 10 input so I can use this.

Concerning the power, I've a lot thinked about this, and I've not found a solution.

All the guitarist I know would have a wireless system directly integrated, so I can't use for example PowerOverEthernet with an Ethernet cable to transmit all the signals. I need so battery. I've thinked of a laptop battery but it's not very easy to iimplement.

I would have digital switching but I ve seen the link you have given, it s interessant.

For the wireless transmission, I think to use a specialized chip to do this, or if I can find anyone who have do this with the AVR32, I will use the AVR32 ...

Anyone else have developped on AVR the midibox64 ?? I would do on this chip and not on a lot of chip because I would have more wood than electronic in my guitar ;) so if I can use a chip to do a lot of things it's better, and better too for the powersupply ...

PS : I've seen you use the AL3101/2 DSP from wavefront, what do you think about this DSP ?? Your miss parker is really interessant, I would be interessed to help to implement MIDI :)

Share this post


Link to post
Share on other sites
Anyone else have developped on AVR the midibox64 ??

A lot of nice ideas around your post. But I think you are at the wrong place here.

Midibox64 needs a PIC and is MIOS based.

I also don't see any reason why this should be changed. It works and ... never change a running system  8)

I'm sure you're welcome to re-program it to an AVR. But I'm also pretty sure, it won't work. I didn't remember the exact reasons (...and there were several good reasons!..) why TK chooses a PIC for his initial project. You may do a search in the forum. I remember some threads where the PIC / AVR discussion was discussed before. So have a look there.

greets

Doc

Share this post


Link to post
Share on other sites
And I need power to treat 10 input in 24bit/96khz.

I think a lot of the AVR32 because it's a good small DSP

So, I guess with "DSP" you mean you want to do audiosampling, then frequency detection and then Midi-Conversion? Is that right?

This is not part of MIOS and I guess even if you're an advanced programmer this is a challenging task. Don't forget that we're talking about tiny microprocessors here, so the problem starts only when you have to deal with multiplication and division. And it's not getting easier when you want to process floats, which you'd have to when dealing with frequencies! Doctor Float and Mr. Division kill your little processor's power faster than you can say boo.  :o

I wonder, if you're using a PC anyway (you're talking of some VST stuff), wouldn't it be possible to transmit the audio signal and do the frequency detection stuff on the PC? I guess there are some dedicated programs available for that purpose.

For the wireless transmission, I think to use a specialized chip to do this, or if I can find anyone who have do this with the AVR32, I will use the AVR32 ...

There is no DIY solution for wirless midi transmission to my knowledge up to now, because of various problems of possible signal loss. There is however a device from M-Audio that might interest you 139,- EUR.

I guess this task alone would cost about half a year development time at least (!) to get useable results.

http://de.m-audio.com/products/de_de/MidAir-main.html

So, have I been discouraging?

Sorry, didn't mean that ;D  ...at least I'd be happy for the DIY-community if you prove me wrong! ;) ;)

And if I totally misunderstood you and you just want a MIDI-controller for your VST software, then go ahead and build a Midibox64 (maybe you could do a small SMD board, I think I've seen someone doing that) and forget about AVRs, 'cause the PICs used for MBHP are better suited for MIDI-Tasks than AVRs. Besides there is no need to re-invent the wheel.

And you don't need no expensive Development Boards with MBHP. That's pretty straightforward with the Bootloader-/MIDI-Approach of MIOS.

Best regards,

Michael

Share this post


Link to post
Share on other sites

"Warning - while you were typing a new reply has been posted. You may wish to review your post." - Hi audiocommander :) ... Anyway:

I don´t really understand what it is you want to do to your guitar. First you talk about a Midi controller (MB64E), then about multichannel audio, DSPs, AVR32, DSPIC, PoE, transmission of audio over Ethernet or wireless (WLAN?). Perhaps you should first describe what this zombie guitar should be capable of?

some infos:

- MIOS has nothing to do with audio. It is purely MIDI. If you want to do hi res multichannel audio, you´ll have to develop a completely new system, in which the MIDI and interface stuff MIOS is capable of will be your smallest problem.

- the dsPIC is a quite different processor - it is a 16 bit architecture (the PIC18 series is 8 bit). The dsPIC is quite nice, but porting MIOS would be a lot of work and the DSP part is cool but not nearly as powerful as necessary.

- the "old" AVRs are cool. I like them better than the PICs. It would surely be possible to do something MIOS-like with them. But they are not that much faster or better than the PICs. So why bother?

- the AVR32 is a new series. A powerful processor, but it comes only in BGA package which can´t be soldered by hand and you need multilayer boards to connect all pins... the only currently available way to use it is the evaluation board from Atmel which is around 500$.

Honestly, I don´t think you have any idea how much knowledge, work and money would be necessary to build something like the thing you´re dreaming of. Even if you know what you´re doing it would take years of work and cost way more than a PC with good audio hardware. One idea: You could put an USB audio/MIDI interface into your guitar together with some small (MB64?) MIDI controller. That way you´d have only one wire to the guitar and some way of remote-controlling some effects plugins on your computer.

Seppoman

P.S. Even if there´s no real advantage in porting MIOS to AVR, I still don´t know why people in this thread are claiming that it is less suitable for MIDI stuff. Has any of you made real experience with AVRs that leads you to this conclusion? For the last year I´ve been working on a quite powerful AVR-based project which also has MIDI capabilites, and I haven´t come over any problems that would be grounds for such a claim.

P.P.S. Word from the master ;) http://www.midibox.org/forum/index.php?topic=6446.msg40118#msg40118

Share this post


Link to post
Share on other sites
Hi audiocommander :)

hehe... Hi Seppoman :)

I still don´t know why people in this thread are claiming that it is less suitable for MIDI stuff. Has any of you made real experience with AVRs that leads you to this conclusion?

Honestly I haven't worked with any AVR. I just remember having read a longer post from TK about that issue and it sounded like there were some good reasons for that opintion.

But the truth is that I was tired of searching the forum. Of course you're right; I should watch my words when there's no real experience behind :-X :)

Cheers,

Michael

Share this post


Link to post
Share on other sites
But the truth is that I was tired of searching the forum. Of course you're right; I should watch my words when there's no real experience behind Lips Sealed Smiley

Me too ...  :-X ;)

greets

Doc

Share this post


Link to post
Share on other sites

Hi Yann,

For the wireless transmission, I think to use a specialized chip to do this, or if I can find anyone who have do this with the AVR32, I will use the AVR32 ...

I'm not sure I understand what you want to transfer over the air. I assume that you want to transfer the MIDI along with the audio otherwise I don't see the need of digitizing all your audio channels. If indeed, you want to transfer audio in a wireless manner, pay attention to the fact that a single channel of 24bits/96kHz audio will require ~2.2Mbps of bandwidth and I'm only counting pure raw data needs here. With 10 channels, you come at ~22Mbps. That's kind of enormous, even current WiFi stuff may have issues to sustain such a throughput rate.

PS : I've seen you use the AL3101/2 DSP from wavefront, what do you think about this DSP ?? Your miss parker is really interessant, I would be interessed to help to implement MIDI

Hehe, thanks. Addition of MIDI is on my todo list for a while along with too many other things  ;)

I quite like the AL3101 and AL3201. They are small and quite easy to use in practice and very much targeted towards audio effects which is our basic aim. You cannot do everything with them, for example FFTs are out of the way as the DSPs work sample based, huge reverbs /delays are not possible either due to too small RAM but for the rest they are quite nice beasts very well suited for audio effects.

Best regards,

Lall

Share this post


Link to post
Share on other sites

Yeaaaah it's interesting !!

So I explain more precisely my project :

I want a guitar with a total versatility of sound and with midi capabilities. That mean that I want to have a lot of pickups ( I think of an EMG 89; Dimarzio Tonzone ; Benedetti Rock ; and a piezo hexaphonic pickup) . I want to control for each pickup the volume and the tone, and maybe have the possibility to have a phase inverser and a low or high-pass filter for each pickup. But I don't want to have 20 pushpush or 30 normal pot + 30 knob on my guitar, it has to be simple to control !

For example that is not good for me :

IMG_0853.jpg

So I want to digitize the sound. For that, I need good ADC, for example the 4 stereo ( so 8 mono) 24bit/96 khz multiplex ADC from TI. And I need a good processor to support the 2,2 mbps by input ... I think so of the AVR32 which have a 150mhz/210mips and several io :). The price for the developpment board ( 500$) is the maximum for me, the DSP from TI or AD is min 2000 $ with specialized IDE...

Moreover, I want to have the capability to control the midi, without having 16 more pot and knob on my guitar.

I had thinked so of the Midibox16E (or why not have a sampler/sequencer on my guitar ;), with a MB SEQ ;))

And finally, I want to transmit all the channel and the midi separatly , or with an ethernet cable with PoE for the power supply of each chip, or by wireless but I've not found how load the battery without too big IC.

So my second big problem is the battery. And less chip is less power consumption ... So if I could use the AVR32 for the audio AND the midi AND the modulation for the wireless transmission, it will be groovy :D

P.S. Even if there´s no real advantage in porting MIOS to AVR, I still don´t know why people in this thread are claiming that it is less suitable for MIDI stuff. Has any of you made real experience with AVRs that leads you to this conclusion? For the last year I´ve been working on a quite powerful AVR-based project which also has MIDI capabilites, and I haven´t come over any problems that would be grounds for such a claim.

P.P.S. Word from the master Wink http://www.midibox.org/forum/index.php?topic=6446.msg40118#msg40118

- the "old" AVRs are cool. I like them better than the PICs. It would surely be possible to do something MIOS-like with them. But they are not that much faster or better than the PICs. So why bother?

If we read the "Word from the Master Wink":D we see that he think about the ATMEGA644 which is old, but a uC like the ATMEGA128 or ATMEGA256 is really more powerful than a PIC18f52. It's a big changement,sure, because even if it's based on MIPS architecture, an ATMEL have around 80 instructions basicly :), but it's a good changement :)

And if you want absolutely a microchip product, the PIC24f is a 16 bit architecture too :).

Honestly, I don´t think you have any idea how much knowledge, work and money would be necessary to build something like the thing you´re dreaming of. Even if you know what you´re doing it would take years of work and cost way more than a PC with good audio hardware

Honestly, I think I have :D ... I think it will take around 1 year of developpment, and without the price of pickup and developpment board, around 100 €.

- the AVR32 is a new series. A powerful processor, but it comes only in BGA package which can´t be soldered by hand and you need multilayer boards to connect all pins... the only currently available way to use it is the evaluation board from Atmel which is around 500$
You can find it http://www.digikey.com/scripts/DkSearch/dksus.dll?Criteria?Ref=8324&Site=US&Cat=33424297 . Sure it's expensive, but an AVR32 is like an ARM and a little DSP in one chip !!

MIOS has nothing to do with audio. It is purely MIDI. If you want to do hi res multichannel audio, you´ll have to develop a completely new system, in which the MIDI and interface stuff MIOS is capable of will be your smallest problem.

Yep, I know than MIOS do only audio, my question is that if I can have MIOS in my AVR32 and if you think it is realisable in a guitar :)

There is no DIY solution for wirless midi transmission to my knowledge up to now, because of various problems of possible signal loss. There is however a device from M-Audio that might interest you 139,- EUR.

I guess this task alone would cost about half a year development time at least (!) to get useable results.

http://de.m-audio.com/products/de_de/MidAir-main.html

We are not forced to have a specialized wireless transmitter for midi, I want to transmit all kind of data, the audio AND the midi if I can, and here, as it's digital 0 and 1 which is transmit and not analog data, I have just to build the adaptater for the recepter so the recepter can understand the modulation of data that I use in the transmitter and rock'n rooooll :)

So, I guess with "DSP" you mean you want to do audiosampling, then frequency detection and then Midi-Conversion? Is that right?

That will be here perfect, but it begin to be very complicated :D. If I can do an onboard system to do an Audio-to-Midi conversion for each string ( with my piezo hexaphonic pikcup) like for the Roland VG system, it will be amazing possibilities in my guitar :D.

But converting 10 input in 24 bit 96 khz, treat them, and transmit them, is a big work, control with a touchscreen and 8 pot the audio and midi is a real big work, so I don't know if I will have this on my todo list :D

Share this post


Link to post
Share on other sites

MIOS has nothing to do with audio. It is purely MIDI.

Yep' date=' I know than MIOS do only audio

[/quote']

MIOS is NOT audio, it's only MIDI.

Have you seen what Lall posted?

Share this post


Link to post
Share on other sites

I want a guitar with a total versatility of sound and with midi capabilities. That mean that I want to have a lot of pickups ( I think of an EMG 89; Dimarzio Tonzone ; Benedetti Rock ; and a piezo hexaphonic pickup) . I want to control for each pickup the volume and the tone, and maybe have the possibility to have a phase inverser and a low or high-pass filter for each pickup. But I don't want to have 20 pushpush or 30 normal pot + 30 knob on my guitar, it has to be simple to control !

I think one knob per function is the most simple way to control things. All the functions mentioned above could be done easily in the analog world. Even if you want to have e.g. only pickup select buttons, a volume and tone knob and a phase button, it would be easier to process/mix everything with analog cirquits that are only controlled by a MCU.

So I want to digitize the sound. For that, I need good ADC, for example the 4 stereo ( so 8 mono) 24bit/96 khz multiplex ADC from TI. And I need a good processor to support the 2,2 mbps by input ... I think so of the AVR32 which have a 150mhz/210mips and several io :).

that means with 8 channels in/2 channels out you have around one million samples per second. so the processor would have 210 cycles time per sample. With even the processed channels seperately transmitted, some cycles to input/output these, handle control surface etc, if you´re lucky you have 100 cycles to do something with the samples. So any processing/effects can´t be very sophisticated.

Moreover, I want to have the capability to control the midi, without having 16 more pot and knob on my guitar.

I had thinked so of the Midibox16E (or why not have a sampler/sequencer on my guitar ;), with a MB SEQ ;))

I´m still not sure what you mean by MIDI. Do you want a Guitar-to-MIDI interface? Midibox 16 or Seq have nothing to do with that. If you don´t want many knobs, for what do you need the Midibox at all? Its central purpose is to connect many knobs to it (and the MB SEQ needs dozens of knobs and buttons and two displays to controll it ???

And finally, I want to transmit all the channel and the midi separatly , or with an ethernet cable with PoE for the power supply of each chip, or by wireless but I've not found how load the battery without too big IC.

Again: what MIDI? Notes or CCs? Transmit to what application? Why would you want to transmit all audio channels seperately if you´ve already done EQ/Volume etc. processing in the guitar? What protocol would you use to transmit audio? M-LAN? Do you have an application on the PC that can receive this?

If we read the "Word from the Master Wink":D we see that he think about the ATMEGA644 which is old, but a uC like the ATMEGA128 or ATMEGA256 is really more powerful than a PIC18f52. It's a big changement,sure, because even if it's based on MIPS architecture, an ATMEL have around 80 instructions basicly :), but it's a good changement :)

The 644 is not old. It´s actually newer than the 128 and 256. The others only have more RAM/FLASH, but the 644 runs 20 MHZ instead of 16. But what´s the reason to change to Atmel at all? Ok, it´s 20 MIPS and not 10 like the PICs, but it is still in the same performance league. And MIDI is a really slow communication. MIOS + PIC is able to process everything without adding noticeable latency to the MIDI out. The applications exist and do work well, so what´s the real advantage?

Honestly, I think I have :D ... I think it will take around 1 year of developpment, and without the price of pickup and developpment board, around 100 €.

Good luck, you´ll need it  :o

You can find it http://www.digikey.com/scripts/DkSearch/dksus.dll?Criteria?Ref=8324&Site=US&Cat=33424297 . Sure it's expensive, but an AVR32 is like an ARM and a little DSP in one chip !!

Sure, but when you have it, how would you solder it and on what board? Even if you´d make an 8 layer board, you need a lot of experience and time - design for 150 MHz is really different than for small PIC/AVR toy stuff.

my question is that if I can have MIOS in my AVR32

No. But you could port it or program something like MIOS on it. Both needs time and knowledge.

I know we´re at an electronic forum, but anyway here´s my philosophical part:

Most world class guitarists play really conventional guitars. If you´re a great guitarist, you can do magic on a Squire Tele with standard single coils. If you are not, 10 different expensive guitars/pickups and FX units won´t change it. 90% of a good guitar sound is made with your fingers and not by using the best 24/96 DSPs. I know a lot of guitarists who didn´t realize this and spent thousands of Euros on equipment while still playing mediocre shit.

Anyway, I don´t want to stop you from dreaming :)

Seppoman

Share this post


Link to post
Share on other sites
MIOS is NOT audio, it's only MIDI.

Excuse me it was a lapse I would say that it's only MIDI :)

that means with 8 channels in/2 channels out you have around one million samples per second. so the processor would have 210 cycles time per sample. With even the processed channels seperately transmitted, some cycles to input/output these, handle control surface etc, if you´re lucky you have 100 cycles to do something with the samples. So any processing/effects can´t be very sophisticated.

Yep, it's true if I have separate ADC, not with a MUX 10:1 ADC.

I´m still not sure what you mean by MIDI. Do you want a Guitar-to-MIDI interface? Midibox 16 or Seq have nothing to do with that. If you don´t want many knobs, for what do you need the Midibox at all? Its central purpose is to connect many knobs to it (and the MB SEQ needs dozens of knobs and buttons and two displays to controll it Huh

Again: what MIDI? Notes or CCs? Transmit to what application? Why would you want to transmit all audio channels seperately if you´ve already done EQ/Volume etc. processing in the guitar? What protocol would you use to transmit audio? M-LAN? Do you have an application on the PC that can receive this

I've wrotten somewhere this "control with a touchscreen and 8 pot the audio and midi (is a real big work)" so I don't want to have only 2 knob and pot, but I don't want to have 30 pot, I know the best is to have one-pot-per-function, but it's not realist, so I try to find the best compromise.

Here I speak of transmit CC to control at distance parameters of FX board and FX software.

I want to have separate audio channel because I can have more possibilities to morph the sound by this, imagine applying a reverb on E string, flanger on B, a fuzz on A, a metal overdrive on an other string, ...

But as I said here, if I can make a conversion of audio to midi to transmit notes by midi, that will be the ultimate guitar :)

That will be here perfect, but it begin to be very complicated. If I can do an onboard system to do an Audio-to-Midi conversion for each string ( with my piezo hexaphonic pikcup) like for the Roland VG system, it will be amazing possibilities in my guitar Cheesy.

Most world class guitarists play really conventional guitars. If you´re a great guitarist, you can do magic on a Squire Tele with standard single coils. If you are not, 10 different expensive guitars/pickups and FX units won´t change it. 90% of a good guitar sound is made with your fingers and not by using the best 24/96 DSPs. I know a lot of guitarists who didn´t realize this and spent thousands of Euros on equipment while still playing mediocre shit.

You're totally right, and I'm not a world class guitarist, but my mind is not in peace :) since I've had this idea, and I think more of the ability of controlling everything, to discover new sound, like a new instrument, than optimising guitar for a traditionnal play of guitar. I think a lot of the guitars and set up of Bellamy, for example his new guitar :

M1D1_Manson.jpg

MUSE_2L.jpg

You can find other images here http://fr.forums.audiofanzine.com/apprendre/mailing_forums/index,idtopic,12091,idpost,,highlight,Gremlins_4u.html

Have you seen what Lall posted?
I'm not sure I understand what you want to transfer over the air. I assume that you want to transfer the MIDI along with the audio otherwise I don't see the need of digitizing all your audio channels. If indeed, you want to transfer audio in a wireless manner, pay attention to the fact that a single channel of 24bits/96kHz audio will require ~2.2Mbps of bandwidth and I'm only counting pure raw data needs here. With 10 channels, you come at ~22Mbps. That's kind of enormous, even current WiFi stuff may have issues to sustain such a throughput rate.

I've seen what Lall posted, but I've said, I don't know where, that it's multiplexed data, not simultaneous data ... And if I can't do that by wireless, I will transmit by Ethernet ...

Share this post


Link to post
Share on other sites

Hi Yann,

If you want to simplify a bit the design at first you could imagine of using an AL3101 as DSP to mix, EQ, filter the 8 mono inputs it can support and then have one mono output. With the available 1024 cycles the chip has, you would have 128 cycles per input channel and that is already quite comfortable.

The system is running at 48kHz/24bits. It would be an A-to-D and D-to-A conversion but at least you could use at first a conventional wireless system made for guitarist. Then you would have a certain flexibility but also robustness and more rapid results than developping everything.

If you plan to use ethernet or any other means to transport the ~22Mbps of data you need for all the channels, then the network should be pretty empty (I mean you should not connect that to your home network with media center and the likes on) otherwise you may end up with issues due to the packet collisions, the fact that every layer adds its overhead i.e. a 10Mbps network at the physical layer does not mean that you have those 10Mbps available at the applicative level, ... Anyway that's another story.

Last but not least on that item, you will then most probably have to write your own driver for your OS which is far from a piece of cake. From that point of view, using a known USB class like audio could for example allow you to avoid that driver work.

In any case, as Seppoman said, I would certainly not do the DSP stuff and the controling stuff on the same chip unless you have really a big one. The problem of having everything on one chip is that you'll have to make compromises e.g. latency vs stability, ...

Also the BGA stuff, again as Seppoman said, you should really forget about it.  It's absolutely no DIY packages, you have to make them solder by specialized companies who have the right tools to do it (and to check that the soldering has been done correctly). If you don't or if they are not equipped to check, that's extremely hard to know where a bug is coming from when something goes wrong.

Last thing, if you make all that for ~100 euros (excl pickup and dev board), you're a champion  ;)

Best regards,

Lall

Share this post


Link to post
Share on other sites
If you want to simplify a bit the design at first you could imagine of using an AL3101 as DSP to mix, EQ, filter the 8 mono inputs it can support and then have one mono output. With the available 1024 cycles the chip has, you would have 128 cycles per input channel and that is already quite comfortable.

The system is running at 48kHz/24bits. It would be an A-to-D and D-to-A conversion but at least you could use at first a conventional wireless system made for guitarist. Then you would have a certain flexibility but also robustness and more rapid results than developping everything.

Okay, I will see in more details what this DSP could exactly do :D

If you plan to use ethernet or any other means to transport the ~22Mbps of data you need for all the channels, then the network should be pretty empty (I mean you should not connect that to your home network with media center and the likes on) otherwise you may end up with issues due to the packet collisions, the fact that every layer adds its overhead i.e. a 10Mbps network at the physical layer does not mean that you have those 10Mbps available at the applicative level, ... Anyway that's another story.

I think you don't see exactly what I mean. Imagine more a guitar like the Gibson Digital to imagine what I think when I speak of Ethernet transmission ( http://www.gibson.com/DigitalGuitarNew/gibsonDigital.html ). Generally a 100 mbits ethernet link correspond  to 40 mbits pratically.

Last but not least on that item, you will then most probably have to write your own driver for your OS which is far from a piece of cake. From that point of view, using a known USB class like audio could for example allow you to avoid that driver work.

Again, I don't want to plug directly to a computer, or a thing like that, but more a separate box like the box of Gibson digital.

If I want to plug directly into my computer, it's more reasonnable to have a ADAT link produced for example by the ADAT chip of Wavefront Semiconductor.

Also the BGA stuff, again as Seppoman said, you should really forget about it.  It's absolutely no DIY packages, you have to make them solder by specialized companies who have the right tools to do it (and to check that the soldering has been done correctly). If you don't or if they are not equipped to check, that's extremely hard to know where a bug is coming from when something goes wrong.

I study electronic and computer science in France, that mean we have BGA equipment ;)

Last thing, if you make all that for ~100 euros (excl pickup and dev board), you're a champion  Wink

MMhh, and for a little bit more, 100 to 200 €, too ?? ;)

So to resume, finally I will maybe not try to make a midicontroller from scratch on AVR32 , or if I do, I will do all the treatment and audio processing on a separate chip, unless AVR32 is enough fast, and I will use only the MCU to control, for example, 8 pot and a touchscreen ;) .

I'm here too to ask, if someones are ready to help me a lot in this project, developping together a really cool system :) ... Mmmmhh ??

Share this post


Link to post
Share on other sites

OK, I think I see more what you intend to do with the transmission of the data. That's a quite nice product from Gibson in the link. A friend of mine has bought such a crazy thingie from Line6, I think it's that one http://line6.com/variax/overview.html but I haven't got the chance to listen to it yet...

I study electronic and computer science in France, that mean we have BGA equipment

I understand now a bit more that you were not more afraid of BGA  ;)

Good luck in your project,

Best regards,

Lall

Share this post


Link to post
Share on other sites

hi

i worked with this idea of the MB in a guitar a while ago and you can see what i ve achieved here

www.io-sound.org

i have another project almost ready as an external box for guitar/bass

but i have to admit that your ideas go quite far from my project, keep us posted i d love to participate somehow on this project, very interesting..

.. but be aware: since i ve started playing with MB in guitars i don t play the guitar anymore!  ;)

simone

Share this post


Link to post
Share on other sites

Very nice site cimo :)

since i ve started playing with MB in guitars i don t play the guitar anymore!

Yep that's the risk, but my mind is focused on this electronic challenge more that I play like a noob :) ...

Can you explain in details what you used for the XY controller for the touchpad ?

And what do you use for the USB controller to be completly plug and play ? I don't want to have USB impemented into my guitar but it can be an idea to have an USB2 into the wireless demodulator :) if I can have wireless on my guitar :) . And you speak about 64 mb on inside memory, but I haven't see a CF adaptor, so what do you use??

Thanks :)

Good luck in your project,

Thank you, I will need it :D

Anyone can just say me if I can go with the AVR32, or if it's a bad idea, and that I must absolutely do my project with more powerful DSP/MCU ?

Share this post


Link to post
Share on other sites
Can you explain in details what you used for the XY controller for the touchpad ?

it is connected to a usb hub so it is "just" as a normal mouse (you can keep your hands on the controller without having to grab the mouse) and it has a special latch rocker switch that will act on the left button (if you use ableton you will clearly understand the function of it, anyway also in an another software it can be used to move 1 knob/fader (reminder for me = adding 2 ledbars to monitor the events)

And what do you use for the USB controller to be completly plug and play ?

audiotrack MidiMate http://www.audiotrak.net/xpmidi.htm - I also used a cheaper kind of USBtoMIDI cable which used the same driver (romio)

And you speak about 64 mb on inside memory, but I haven't see a CF adaptor, so what do you use??

also the memory is attached to the usb hub so it can only be reached via the computer, well all the guitar can only used with a computer..

let s keep in touch

simone

Share this post


Link to post
Share on other sites

Okay,

I understand better what's the concept.

However, I don't want to need absolutely a computer with my guitar ...

A friend has told me of the product of Zilog, has anyone heard about this product ???

He said me too that it's utopic actually to want transmit multichannel audio by wireless, I will look at powerful RF chip, and if I'm not lucky, I will concentrate on the transmission by ethernet or ADAT ...

Share this post


Link to post
Share on other sites

Hi Cimo,

Just wanted to say that I really like the music you're doing. I think my preferred one is Sounds for Zara's dreams but I really like them all.

Also very nice job on your guitars  :o

Yann,

If you don't find an RF solution, I think ADAT would be a good choice and surely would be far less work than ethernet.

Best regards,

Lall

Share this post


Link to post
Share on other sites

Yep, I like too the Sounds from Zara's dream :)

I haven't found solution for the RF, but I've found a practical question for the midibox expert:

I will use the normal solution for the MIOS & cie. But I want to interface a touchscreen controlled by the AVR32 ( it's too much for the PIC) to the Midibox. Does the Midibox allow this ? Controlling by a XYtouchpad instead of a sensor or a pot/fader ??

Share this post


Link to post
Share on other sites

hi

txx guys for the comment .. i had forgotten that i used to play music before getting stuck here  ;) maybe it s time to get back in track.. properly said..

about Sounds for zara s dreams it can be interesting to know that it was not overdubbed (except for 1 single Subtraktor synth s track) and that all layers (3) (yes of course except the reason s) are played simultaneously (waves tap delay with full feedback)

hopefully i will soon post pictures of my latest project: it s a simple idea but quite innovative .. i am still having some troubles with ledrings but i am posting already here http://www.midibox.org/forum/index.php?topic=8551.0 so if you think you can help ....

simone

Share this post


Link to post
Share on other sites

Hello

Concerning my project, for the wireless transmission, the only thing which allow me sufficient datarate is a wifi chip, or an other chip bsed on the OFDM modulation (DAB,..)... Do you think it is a realistic idea to implement this ??

Share this post


Link to post
Share on other sites

hi

i have a question>> how are you going to supply the whole thing? we don t wan t a wifi system if we have to go around with a supply cable right?

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now
Sign in to follow this  
Followers 0