270$ for a STM32F4 Discovery

[MyCo]

Hi!

I'm a regular reader in this forum and just haven't registered before. But now I stumbled upon a kickstarter project that really pisses me off and I thought it should be mentioned somewhere on the internet:

http://www.kickstarter.com/projects/marser/owl-programmable-effects-pedal

It's basically a 20$ STM32F4 Discovery Board without motion sensor, microphone and Audio DAC, wrapped in a really bad looking case. And they want 165 Pounds or 270$ (with shipping) for that!
Yes music equipment is quite often really expensive, but shouldn't a crowd funded project use realistic product prices? And there is absolutely no justification for this.
The basic unit has 12bit DAC/ADC... which means they use the µC internal DAC/ADC. So the board inside this box is basically a voltage regulator, a USB connector, some Audio jacks, user controls, the µC and then just passives. BOM: 30$ maximum.

This looks for me like someone wants to make big money.

Regards
Maik

[M. András]

lol 27000bucks for 1MB of ram :D:D

[Dave]

Not 1MB (megabyte), but 1Mb (megabit - 1/8 of a megabyte).

[AndyC_772]

Reality check here guys, STM Discovery boards are cheap because they're subsidised by the sale of STM32 microcontrollers. You can't judge the price of any product that happens to use an STM32 just because ST sell a cheap development board to people wanting to learn and evaluate their processors.

What you have here is an independently designed product, which has to be profitable in its own right. Sale of one of these units does not lead to a design-in and subsequent sale of thousands of chips!

For sure, the box contains a processor, power supply and audio I/O that's compatible with guitar signals - which, I assure you, are a law unto themselves in terms of signal level, impedance and noise floor. The board has to be laid out, manufactured, assembled, programmed, installed in a custom designed, rugged enclosure, tested, CE / FCC certified, packaged and retailed in very small numbers. Then it has to be supported, and a few of them will have to be repaired or replaced. Add in the office overheads too - heat, light, power, administration, taxes, and the fact that engineering know-how is a valuable thing it its own right: people who have it have worked very hard to get it and have every right to be rewarded for their efforts.

There's absolutely nothing wrong at all with wanting to make "big money" as you put it, and just in case you're in doubt, virtually every product on the market is priced at what the seller believes the market will pay. There's no rule anywhere that says a product "must" or "should" be sold at a price dictated by the cost of the parts that went into it.

I didn't see a retail price, but if it's around the £250 mark then that feels about right. If the product is successful then they'll have a chance at a decent living - but they won't make "big money". If you think they might, and that they've not really done enough to earn it, start a competing business.

[AndreasF]

Another thing to consider, I think, is that a large aspect of this project seems to be the software. Yes, you can probably build this up yourself for a lot less, but you're gonna have to put in a lot of time to develop the software. From that perspective it's actually quite surprising that they want to release it completely open source.

[MyCo]

It's ridiculous:
http://hoxtonowl.com/hardware/assembly/

The development time doesn't justify that price. They use open source components like JUCE and GCC, the only thing that they need to develop is a schematic, a C++ framework and a bootloader. The C++ framework is basically already done, because it's nothing else than the CMSIS and a Main-File.

And yes, I could compete against that. I've done several Audio-DSP projects and VST development as well. Even my Spartan6 Audio Dev-Board wouldn't cost half of this.

[FrankBuss]

Well, it's open source software and hardware, so anyone could build it cheaper themselves, or even create a competitive kickstarter project (with another name), which might be a good idea, if some of the stretch goals are not met, because they are really cool features for musicians.

[marshallh]

Lots of people offended at the idea of someone actually making a profit...

If you think it's too pricey, no one's forcing you to buy it.

There is much more to a product than just a BOM cost.

[rougeaux]

Lots of people offended at the idea of someone actually making a profit...

If you think it's too pricey, no one's forcing you to buy it.

There is much more to a product than just a BOM cost.

Definitely, particularly in the realm of instrument effects.  For example, the Dunlop Cry Baby wah pedal has an MSRP of over $150, even though the actual workings inside involve a handful of passive components, a couple of transistors and diodes, a pot, two switches, two jacks and a battery connector, all of which I probably have sitting around.  I could probably even piece together a reasonably sturdy case for it.  For my purposes (experimental electronic music) my effect would probably work just fine.  But if I was a professional guitarist, would I trust my own handiwork to function flawlessly, consistently, at every venue I play?  Probably not, and certainly not as much as I'd trust one made by people who specialize in them.

Keep in mind, I'm not saying these guys are experts, or that their product will come close to hitting the mark with regards to quality and consistency, only that the benchmark by which their product is judged should be how well it does what it's supposed to do, rather than part count or build complexity.

[MacAttak]

And yes, I could compete against that. I've done several Audio-DSP projects and VST development as well. Even my Spartan6 Audio Dev-Board wouldn't cost half of this.

Then stop crying about it and do it.

Despite the first impression you may have, this area of the forum isn't just a free-for-all place to bash crowd-funded projects. It is a place to discuss the technical aspects of those projects which include electronics engineering aspects.

Do you have a particular concern about the hardware?

Generally speaking, we don't really care how much they want to sell it for unless it seems far too low to cover a reasonable BOM (which is a sign of a project at high risk of failure).

[Corporate666]

It's ridiculous:
http://hoxtonowl.com/hardware/assembly/

The development time doesn't justify that price. They use open source components like JUCE and GCC, the only thing that they need to develop is a schematic, a C++ framework and a bootloader. The C++ framework is basically already done, because it's nothing else than the CMSIS and a Main-File.

And yes, I could compete against that. I've done several Audio-DSP projects and VST development as well. Even my Spartan6 Audio Dev-Board wouldn't cost half of this.

The price of any thing is justified by only a single metric... whether people are willing to spend the money for it, or not.

If someone will pay $1,000 for an LED and resistor connected to a battery, then it is worth $1,000 - at least to that person.  If you get enough people interested in your product at the price you offer it for, then you have a sustainable business. 

These guys were asking for 8000GBP.  Looks like they have gotten more than that, and therefore, it is unquestionably 'worth' 270GBP, at least to the backers. 


I am heartened that so many on this board see the business side of this and have no problem with someone making a profit.  As for your comments, if you believe the price is too high (even to the extent that it 'pisses you off'), and considering you say you could compete against that, then the only question in my mind is why aren't you competing?  If, as you say, there are $30 worth of parts in there, then you should make your own and sell it for the price you feel is fair - and if it's below these guys by a substantial amount, then surely you will make tons of money.

Are you going to jump into the market and make one?

[MyCo]

Why should I built that? It's a Discovery board. You can completely rebuild that yourself. Buy a Discovery board, connect an Audio Jack and 4 pots to it, and you have exactly the same result. It's even better, because it has a real Audio DAC on it.

The M4F is powerfull, no doubt, but it isn't powerfull enough to work as Audio DSP. Shure you can do some simple Amps and Filter, but that's all. It isn't powerfull enough to build a Reverb or even realistic Amp simulations (which such a device is aimed for). So this project is just a toy for developers... and those will be disappointed by the result very soon.

[Corporate666]

Why should I built that? It's a Discovery board. You can completely rebuild that yourself. Buy a Discovery board, connect an Audio Jack and 4 pots to it, and you have exactly the same result. It's even better, because it has a real Audio DAC on it.

The M4F is powerfull, no doubt, but it isn't powerfull enough to work as Audio DSP. Shure you can do some simple Amps and Filter, but that's all. It isn't powerfull enough to build a Reverb or even realistic Amp simulations (which such a device is aimed for). So this project is just a toy for developers... and those will be disappointed by the result very soon.

A discovery board won't do what that product does.  That product has a housing, dials, switches, and most importantly, software.

So far, they have gotten around $20,000USD with just under 30 days to go - so most likely they will end up around $50,000, or more.

If it is as simple as you say, then you could easily build it, undercut them on price, out-compete them on features and function, and do even better than they did.

Competition is the foundation of capitalism, and how people get rich.  If their product is such a gross infraction upon the wallets of their customers, then you have an easy business opportunity.  There's no reason you wouldn't take it and make a ton of money.  Unless.....

[MacAttak]

If you could produce a superior product for as cheaply as you claim - then I don't understand why this product would piss you off instead of making you very happy?

Someone else has essentially proven the market for the product. They have taken that risk and done the marketing research for you. And at least 200 people are willing to pay multiple times what you feel is a fair selling price point. A person willing to pay for that product would certainly be willing to spend 1/5 of it for a superior product. This is just common sense. Plus there are a range of people who are unable to afford the higher-priced product but who would be able to afford your cheaper alternative. And that's just the limited Kickstarter audience.

You should be ecstatic about it. This is a veritable gift horse for you, don't look in it's mouth!

[FrankBuss]

The M4F is powerfull, no doubt, but it isn't powerfull enough to work as Audio DSP. Shure you can do some simple Amps and Filter, but that's all. It isn't powerfull enough to build a Reverb or even realistic Amp simulations (which such a device is aimed for). So this project is just a toy for developers... and those will be disappointed by the result very soon.

You should try it. The CPU has at least some kind of floating point acceleration (but not as fast as on a desktop CPU). And it is easier to program it than a DSP, and one of their goals is to provide a system for programmers to implement their own effects.

But building such a device with a DSP would be a nice new Kickstarter project, for a different target audience (programmer side), or you might need to implement some cool effects yourself to get some backers. Anyone?

[FrankBuss]

I've looked up the numbers in the ARM System Developer's Guide book: There is a reference implemenation for ARM7 devices, which uses fixpoint calculation for FIR filters. It needs 9 cycles per tap. So for a very good FIR with maybe 40 taps, it needs 360 cycles. This means you can calculate 460,000 samples per second. If you sample with 48 kHz, you can use 10 filters for your effect. And if you use IIR filters, it would need even less cycles and the reference implementation didn't use the new floating point operations of the CPU (my exemplar of the book is from 2004). So looks quite powerful for me.

Of course, I guess with a dedicated DSP running at the same speed, you could easily archive a multiple of the performance, especially with one of those shiny new multicore DSPs. Or just use a bunch of some cheap DSPs, you can get them for single digit dollar prices.

[MyCo]

If it is as simple as you say, then you could easily build it, undercut them on price, out-compete them on features and function, and do even better than they did.

Jesus, why do I have to proof it. Yes I can, I've done it already. I don't want to make money with my hobbies, I release all of my projects for free. Here is my competitor product with all the features that they want to provide with their extended goals and even more:


Yes there is actually an Arduino under it... I thought it would be nice to be "Arduino compatible"... stupid idea!

[Corporate666]

If it is as simple as you say, then you could easily build it, undercut them on price, out-compete them on features and function, and do even better than they did.

Jesus, why do I have to proof it. Yes I can, I've done it already. I don't want to make money with my hobbies, I release all of my projects for free. Here is my competitor product with all the features that they want to provide with their extended goals and even more:


Yes there is actually an Arduino under it... I thought it would be nice to be "Arduino compatible"... stupid idea!

So if you could do it, but you don't want to do it because you don't want to make money with your hobbies, what are you "really pissed off" about?  That someone else is making money?  That people are spending their money on devices they want to spend their money on?

I really don't understand.  It sounds like you have an idea of what prices should be for things and you're angry that people pay prices other than what you think they should.  I can't understand why anyone would be angry about someone making a product people want and getting paid for it.  I'm sure your house or apartment wasn't free, and I'm sure your car isn't free or your electricity or petrol, etc, etc.  Capitalism makes the world turn.

[cloudscapes]

I'm normally excited for new open DSP platforms, but the specs are a bit underwhelming for audio DSP. 12bit isn't great, especially if they're using the micro's on-board ADC. Even an external MCP3201 would be better. What are those, two bucks each and an afternoon of work to interface? I use those all the time. They're incredibly easy to use and don't require a 256/128x system clock like other (and better) ADCs.

I designed built an audio board (with a bit of help from this forum for a few layout details) all by myself, in little time, while learning C and high-speed layout rules, and it's already better than what they're offering. It's only 80MHz instead of 168, but that's little more than a micro swap. It also has a dedicated 12bit ADC and 16bit DAC, and analog buffer/gain/blend controls. I could add way more than 1Mb RAM in about a weekend's worth of work. Granted it doesn't have the grunt to do hardcore audio DSP, but neither can the micro they're using.



I'm sure they have the talent and brains to do something much better than what I made, but currently I don't see it. I made mine using basically pocket money. They're four people and asking for a lot more. That's the confusing part. It's not that difficult to build something with the basic specs they're offering, so where's the money going?

I REALLY want to see more DSP hardware like this! But I think they have to step up with the ADC/DAC and offer a better RAM upgrade than a second and a half worth of audio. Then it starts being interesting. A true DSP chip would actually make it exciting!

Skip those lower kickstarter tiers and it might be interesting! 

[FrankBuss]

Granted it doesn't have the grunt to do hardcore audio DSP, but neither can the micro they're using.

Their micro is quite powerful. Unlike your PIC it runs every command in 1 clock, not 4 and it has some powerful DSP like multiply-add commands. As written before, implementing 10 filters with 40 taps each should be possible at 48 kHz sample rate. Now I feel like I should try this

so where's the money going?

At least 4 people, working full time 1-2 months for it. For their money goal at Kickstarter, they will get paid very low compared to the usual hourly rate paid for engineers and programmers.

But I think they have to step up with the ADC/DAC and offer a better RAM upgrade than a second and a half worth of audio. Then it starts being interesting. A true DSP chip would actually make it exciting!

It's one of their stretch goals and right, they should really do this.

[cloudscapes]

Granted it doesn't have the grunt to do hardcore audio DSP, but neither can the micro they're using.

Their micro is quite powerful. Unlike your PIC it runs every command in 1 clock, not 4 and it has some powerful DSP like multiply-add commands. As written before, implementing 10 filters with 40 taps each should be possible at 48 kHz sample rate. Now I feel like I should try this

so where's the money going?

At least 4 people, working full time 1-2 months for it. For their money goal at Kickstarter, they will get paid very low compared to the usual hourly rate paid for engineers and programmers.

But I think they have to step up with the ADC/DAC and offer a better RAM upgrade than a second and a half worth of audio. Then it starts being interesting. A true DSP chip would actually make it exciting!

It's one of their stretch goals and right, they should really do this.

Mine is a PIC32 though, not a PIC, and though I'm far from a micro expert, I'm pretty sure they also run one instruction per clock (according to microchip's website). There is the MHz speed difference, though.

Still, it really feels to me that the first thing they need to do is at the very least, if they MUST use 12bit converters, use external ones. You can't properly decouple and seperate the internal ADC from the digital, obviously, and unless you're doing fuzz, distortion or bitcrushers (where noise isn't the greatest of concerns), should be at the absolute top of their priorities for even the most basic kickstarter milestone. it shouldn't be something people need to "buy" at a higher price, even if it looks like it'll probably get there. As I've said, it literally takes a weekend for one person to hook up cheap and easy to use external 12bit converters. Microchip's especially, they're really well documented, take almost no layout space, and will have a lower noise floor than the micro's on-chip stuff.

My main point is that only at the higher price tiers does the funded project make sense and becomes interesting. At the bottom tiers, without external converters, it makes no sense considering how easy it is to hook up the most basic external 12bit ADC and DAC. Literally one person and a couple days. Absolutely should be in the base package.

As perspective, this is how I'm seeing it, just as example: fictional Company X offers a micro dev board at the $10,000 fund milestone, but with only a functional minimum of decoupling caps. One or two for the micro, and none whatsoever near the power and for filtering. It works, but is unstable. But then Company X says that if they reach $12,000, they'll add half a dozen decoupling caps on all the micro's power pins and a few extra around the board. To me it absolutely makes no sense to offer such a deal when all it takes is a few minutes work and pennies per board, for something that should have been there in the first place. That's how I'm seeing it.

[FrankBuss]

Right, the PIC32 series has a MIPS core and executes commands in one clock cycle, only the low cost PIC18F and PIC16F series need 4 clock cycles. And more important: there are multiply-add commands integrated for a fast FIR filter implementation. And yes, they should add external ADCs and DACs.

[Hideki]

While STM32F4 isn't awesomely powerful, it is fast enough that I made this synth board using it: http://sybo.co.nf/

It has a 24 bit stereo codec, and it does handle quite a few simple voices even with reverb added. You can even play drum samples from an SD card
I would probably have redesigned it a bit if I did it again, but it's still a cute little board.

12 bit mono on their kickstarter pedal (unless they get a lot of $$$) sounds like a joke, but apparently many people like the joke so much that they want to pay for it.

[MyCo]

So if you could do it, but you don't want to do it because you don't want to make money with your hobbies, what are you "really pissed off" about?  That someone else is making money?

No, they want to make big money! There are a huge amount of open source projects and examples for what their project does. There is SYBO (thanks Hideki) or this one:
http://ebrombaugh.studionebula.com/synth/stm32f4_codec/index.html

And those projects are free and already outperform their 270$ joke project. And I bet, they know that and use that as reference!
And also when I read something like this:

12 bits represents about 70dB of dynamics, which is pretty good for most effects. Having a simpler design helps keep the costs and risks down.

How hard is it to add a Codec on a board? In their video presentation the people have job descriptions like "DSP engineer" or "hardware/software designer" and a codec is a risk? Are you kidding? The parts cost 10$ including all passives.

Their project kicks everyone in the butt who has already done DSP projects with the same, similar or better hardware. And that's what pisses me off: They want too much money for nothing. Everything that this project can do, has been done before and is freely available. To clarify that: I've nothing against offering such a project for crowdfunding, but then they have to use a realistic price point.

@FrankBuss: Yes, an ARM can perform instructions in single clock cycles, but often it doesn't do that. You have memory and state change latencies like in every RISC architecture.

[millerb]

I don't see any big money here. Even if 100% of what's been raised was pure profit, it's not enough to cover one engineer's salary, let alone 4 guys.

Hopefully there's a good amount of profit in there. That will make for some essential breathing room if they underestimated anything.

I think when some people look at these projects, they forget the expense of having the product in a housing. The stupid box you plop your PCB in can easily cost more than the all of the other parts combined. And this box has to be rugged enough to stand up to the kind of abuse that's expected in a stage environment. This isn't just a board that can just be sent out in a plastic bag.

Another thing to consider is that their support overhead is going to be larger than a lot of products because the community aspect is one of the main features. They'll have a monthly expense just in keeping up a server to host their community forum and serve the device profiles they and their members create. They'll also have to keep putting out new profiles and keep refining their SDK and firmware to keep it fresh and interesting, again time and money.

Given that they're asking for more than enough money to actually have a good chance of delivering the product, I'd say the main concern here is that they could fail to garner any more interest after the initial funding campaign. Who knows how many of the backers have the skill set to create interesting profiles for the thing, I'd guess not most of them. That means that if they can't keep sales up then the software guys are going to move on to other things and leave the users with a stagnated device. An open C++ library is cool but that meas dick to your average guitar guy looking for an fx pedal. And the fx pedal market has been saturated for a long time.