Hi Neil,
There doesn't seem to be a good man page or guide on this. Lots of info on the web, but it varies from "change this and that setting in config.txt" to "you need to recompile the kernel". The Pi 2 seems up to the task, and I think that multichannel audio via HDMI on that platform is getting a lot of interest, so it might be a matter of time until things work more smoothly out of the box.
I couldn't get it to work with the extractor. Your mileage may vary!
Hmmm ... here's a YouTube of the Kii in action - music starts halfway through, https://www.youtube.com/watch?v=Y1Nebe3ULcY&feature=youtu.be
First piece sounds impressive, but the second shows problems - early days, will be interesting to see how it develops ...
First piece sounds impressive, but the second shows problems - early days, will be interesting to see how it develops ...
Frank,
Though I think that the speaker looks okay, better than some, I only see one thing here that I haven't seen done in the past. That is the control system using a dsp. This configuration with the two side firing speakers has been done many times, that is not new at all and adding the rear firing speakers to act as dipoles has also been done, The combination with the steering supposedly by the dsp is a refined version of what I have seen just done acoustically with traditional cross overs. I am not saying they may not sound good, but that I don't think I see this as being all that unique in overall design for the acoustic outcome. A steerable polar plot is what you have, but who knows what the individual drivers would sound like themselves, that is always part of the question.
Though I think that the speaker looks okay, better than some, I only see one thing here that I haven't seen done in the past. That is the control system using a dsp. This configuration with the two side firing speakers has been done many times, that is not new at all and adding the rear firing speakers to act as dipoles has also been done, The combination with the steering supposedly by the dsp is a refined version of what I have seen just done acoustically with traditional cross overs. I am not saying they may not sound good, but that I don't think I see this as being all that unique in overall design for the acoustic outcome. A steerable polar plot is what you have, but who knows what the individual drivers would sound like themselves, that is always part of the question.
Last edited:
If we go back to the basic outline of what Overkill has outlined as the preferred implementation of electronics inside and outside the speaker enclosure now I have some questions. Let's concentrate on just the speaker implementation for a minute. If you place the amplifiers and the dsp and I assume at least a couple of dac conversions into that enclosure so you have power and crossover functions inside the enclosure what dsp would be the preferred device that could later be integrated with all the secondary functions that would later be outside the enclosures and which dac chips would you use? I know of one ESS dsp that would be relevant but what others could we or should we look at and how much does the dsp have to actually do in this simple part of the configuration? I am assuming that we need at most one a to d and one d to a dac chips with the dsp.
This just seems like a logical place to start with a powered speaker, we will need xo functions if we are going to have a bi-amp situation or greater and most I would assume would go with an active network and not a passive xo, that would seem to be going backwards if you did that. We can get to the software implementation later of what we use to set this up, let's stick with the hardware site for at least a bit of the conversation.
This just seems like a logical place to start with a powered speaker, we will need xo functions if we are going to have a bi-amp situation or greater and most I would assume would go with an active network and not a passive xo, that would seem to be going backwards if you did that. We can get to the software implementation later of what we use to set this up, let's stick with the hardware site for at least a bit of the conversation.
Last edited:
Steven, that is the $60,000 question.
We can find lots of botique options, or heck the RPi only costs 30 bucks and if it's DIY a lot of folks might think that's pretty nice to have that much scalability in a speaker, so then you are down to programming again. The problem is dealing with the unpredictability of an OS.
That said, if we have anyone on the hardware side who want to make an open standard board based on a chip solution like the TMS320VC5410A, that might be possible but then you are back to building a new untested wheel. I think that is the same chip that Overkill has on his board, so then you have a working design on the same chip that has some development behind it and a supply chain.
Other options?
We can find lots of botique options, or heck the RPi only costs 30 bucks and if it's DIY a lot of folks might think that's pretty nice to have that much scalability in a speaker, so then you are down to programming again. The problem is dealing with the unpredictability of an OS.
That said, if we have anyone on the hardware side who want to make an open standard board based on a chip solution like the TMS320VC5410A, that might be possible but then you are back to building a new untested wheel. I think that is the same chip that Overkill has on his board, so then you have a working design on the same chip that has some development behind it and a supply chain.
Other options?
Binely,
Since I am so new to this whole discussion I don't have anything relevant to add on a technical basis. I wonder if that chips 16bit architecture would have any limitation of dynamic range or if that is more than enough for the processing it would be required to do. Basically we are talking about filter functions and perhaps that is more than enough if other functions were off-loaded to any external electronics we would also use in the more complex systems. this I just don't know how to answer as I don't know how the configuration and divisions of functions are set up. I'm thinking things like room correction and such, I would imagine you would do this in the external electronics and not inside the speaker. Which functions go where and how much computing power we need at each level is still unknown to me.
Since I am so new to this whole discussion I don't have anything relevant to add on a technical basis. I wonder if that chips 16bit architecture would have any limitation of dynamic range or if that is more than enough for the processing it would be required to do. Basically we are talking about filter functions and perhaps that is more than enough if other functions were off-loaded to any external electronics we would also use in the more complex systems. this I just don't know how to answer as I don't know how the configuration and divisions of functions are set up. I'm thinking things like room correction and such, I would imagine you would do this in the external electronics and not inside the speaker. Which functions go where and how much computing power we need at each level is still unknown to me.
Steven, coming from a computing background I don't see the DSP side of things as a big deal - just use whichever chip is best value for money at the time, and just do enough processing to get the job done - whatever you want in terms of the signals been fed to the DACs would be possible.
My biggest concerns would be getting the power supplies, and isolation between digital and analogue right, and making sure vibrations from the drivers operating didn't affect the electronics - this could make or break the subjective quality of the end unit.
Thanks Frank,
Bottom line is someone else is going to have to design the circuit and walk me through how to do all the software manipulations to do all of this. I am a computer user and not really a programmer. My daughter would be light years ahead of me on that side of things, she is learning to write code, I have no interest in that. I can build a computer from parts and install the software and make it run but that is where I run out of steam.
Bottom line is someone else is going to have to design the circuit and walk me through how to do all the software manipulations to do all of this. I am a computer user and not really a programmer. My daughter would be light years ahead of me on that side of things, she is learning to write code, I have no interest in that. I can build a computer from parts and install the software and make it run but that is where I run out of steam.
I really don't mean to be against the grain here, but aren't you going for too large a first step then? If I'm reading you right, most of what you are trying to do could also be accomplished with a computer running Acourate and a multichannel ADC/DAC, and just run analog signal lines to your speakers housing the amps and drivers. One signal line and one amp per driver per side. You can do DSP (driver linearization, room correction, time alignment) and have active multi-way speakers.
Once you get that right, which might take a good while, you could take a further step and do the digital XO and DSP inside the speaker box. By that time, BTW, hardware might have evolved to a point this last step might be even easier.
Don't know. I'm from the camp of taking incremental steps. Keeping the audacious goal in mind, though.
my half cent
Lewinski,
I've done the speaker side of things, and have done passive networks and that is old hat at this point. Taking things to the next level which is what I am trying to do can be in steps. The first step is to move away from passive networks and I was thinking originally of using active analog networks and some analog eq. But why do that today, it just seem the right time to move into the digital age of control. I may have to learn some things, but there are enough others who have been doing this long enough it shouldn't be rocket science anymore. I have some local gurus I can ask and they would probably find what I am trying to do rather mundane to implement.
I've done the speaker side of things, and have done passive networks and that is old hat at this point. Taking things to the next level which is what I am trying to do can be in steps. The first step is to move away from passive networks and I was thinking originally of using active analog networks and some analog eq. But why do that today, it just seem the right time to move into the digital age of control. I may have to learn some things, but there are enough others who have been doing this long enough it shouldn't be rocket science anymore. I have some local gurus I can ask and they would probably find what I am trying to do rather mundane to implement.
Agree. The proliferation and technology to bring it to a discrete level inside the monitor is here. To a point, there was a time when HTPC's were considered "out there". Now, what has happened is they are surpassed by technology that has decentralized the processing to a discrete level. For example, instead of having an "all in one" box that stores your programming, tunes and keeps up the programming guide, all those functions are broken down amongst pieces that talk just in time to each other.
For example, most people have TV's with a "roku" on it. Well, it wasn't long ago that you had to buy a discrete one.
And when Netflix started streaming services, everyone thought "oh that won't work". But here we are, the cord cutting is getting ever so prevalent, and people are integrating technology with their own preferences at an increasing rate.
As far as I can tell, the only real mainstream integration of any of this technology is in the SONOS product, which while probably acceptable, I do not expect it to be cutting edge. In fact they are pretty set on the idea of controlling everything about their experience, they have their own little walled garden there. Keep in mind a full 5.1 Sonos system will set you back 1800, and they aren't even willing to give you any meaningful metrics on the specs of their system. Pretty uneven trade, IMHO.
Also how many people do you know who are honest-to-goodness experts in their field. This is a unification thread... what we don't know, hopefully there is enough understanding out there to bring it together.
Before tablet computing, there were enough technically saavy people out there who wanted to make happen bad enough that they started just hacking them together from parts they could buy and then tried developing their proof of concept.
Now, the problem is most people don't understand and don't want to understand a lot of the science behind sound.
But, then again, how many people really understood what a "capacitative" touch screen with OLED technology, smart paper, etc. etc. was 5 years ago. I can tell you that those terms weren't mainstream, that's for sure.
What a number of people will say is that awhile back, sound suffered from a period of terrible snake oil salesman, who would sell ice to eskimos.
When you get people in a room, and you show them what good audio really sounds like and give them an experience that makes it dead simple, your acceptance rate of those terms will still be really low. But, the interest level will go up. And when people have a truly unique attachment, an intuitive experience, that is where a connection is made.
Put a shroud over the product, it doesn't matter. Let them do some blind comparisons, and if it really is good... well I'm just preaching to the choir now.
Right now, there is a developed market out there for custom audio. And people who buy that product or build that product want the "chip foose" of audio. And you can buy that. For a price.
What there isn't out there yet, is an entirely new connection that the code hasn't been broken in on yet.
To the major technology companies, audio is a peripheral... good!
Time for a dsp speaker that you can have the dsp tune to the driver, and then have the base unit sync with a microphone to tune the room response.
With the division of the labor... I see good thing happening. Even if the speaker response isn't super high end, now you can correct the deficiencies in the signal output based on the correction engine for the speaker.
And instead of the central unit trying to run all those clock cycles for the speaker, it's off loaded. And then the load is on the base to do all the room response correction, effects, etc.
And then, imagine having a modular system capable of figuring all that out on the fly, with minimal interaction when it comes to configuration.
1.0 - good driver/discrete amp
1.5 - good driver/discrete amp/tuned dsp
2.0 - above plus wireless interconnectivity
2.5 - above plus wireless programing
and on and on....
and I'm spent.... night guys.
For example, most people have TV's with a "roku" on it. Well, it wasn't long ago that you had to buy a discrete one.
And when Netflix started streaming services, everyone thought "oh that won't work". But here we are, the cord cutting is getting ever so prevalent, and people are integrating technology with their own preferences at an increasing rate.
As far as I can tell, the only real mainstream integration of any of this technology is in the SONOS product, which while probably acceptable, I do not expect it to be cutting edge. In fact they are pretty set on the idea of controlling everything about their experience, they have their own little walled garden there. Keep in mind a full 5.1 Sonos system will set you back 1800, and they aren't even willing to give you any meaningful metrics on the specs of their system. Pretty uneven trade, IMHO.
Also how many people do you know who are honest-to-goodness experts in their field. This is a unification thread... what we don't know, hopefully there is enough understanding out there to bring it together.
Before tablet computing, there were enough technically saavy people out there who wanted to make happen bad enough that they started just hacking them together from parts they could buy and then tried developing their proof of concept.
Now, the problem is most people don't understand and don't want to understand a lot of the science behind sound.
But, then again, how many people really understood what a "capacitative" touch screen with OLED technology, smart paper, etc. etc. was 5 years ago. I can tell you that those terms weren't mainstream, that's for sure.
What a number of people will say is that awhile back, sound suffered from a period of terrible snake oil salesman, who would sell ice to eskimos.
When you get people in a room, and you show them what good audio really sounds like and give them an experience that makes it dead simple, your acceptance rate of those terms will still be really low. But, the interest level will go up. And when people have a truly unique attachment, an intuitive experience, that is where a connection is made.
Put a shroud over the product, it doesn't matter. Let them do some blind comparisons, and if it really is good... well I'm just preaching to the choir now.
Right now, there is a developed market out there for custom audio. And people who buy that product or build that product want the "chip foose" of audio. And you can buy that. For a price.
What there isn't out there yet, is an entirely new connection that the code hasn't been broken in on yet.
To the major technology companies, audio is a peripheral... good!
Time for a dsp speaker that you can have the dsp tune to the driver, and then have the base unit sync with a microphone to tune the room response.
With the division of the labor... I see good thing happening. Even if the speaker response isn't super high end, now you can correct the deficiencies in the signal output based on the correction engine for the speaker.
And instead of the central unit trying to run all those clock cycles for the speaker, it's off loaded. And then the load is on the base to do all the room response correction, effects, etc.
And then, imagine having a modular system capable of figuring all that out on the fly, with minimal interaction when it comes to configuration.
1.0 - good driver/discrete amp
1.5 - good driver/discrete amp/tuned dsp
2.0 - above plus wireless interconnectivity
2.5 - above plus wireless programing
and on and on....
and I'm spent.... night guys.
Binely,
Nice little synopsis there. I do know many things that may not be audio related, at the same time I am not afraid of technology. I worked with a company developing capacitive touch screen more than 20 years ago, also solar panels for what was then Arco Solar about the same period of time. My first payment from an outside company was actually from Arco Solar. I have just been there to early for most of these things to be commercially viable at the time. It's time for well done dsp controlled speakers. I turned down a chance to work for Sonos, I wouldn't do it. As the young crowd is aging who grew up with internet radio and streaming music start to make some real money they will find they want better sound. I don't want to compete with the $150.00 JBL tower speakers you can see at Fry's every weekend in the paper. I want to give the public what I think they truly want. Today you have to meet the style requirements also, Apple has changed the game on industrial design. No more funky wood boxes are going to meet that desire. I think there are many on this site who are already doing this on a diy basis as far as control for loudspeakers, the Rephase thread is just one example dealing with just the software side of this. The easy way is to just purchase pre-made modules from others but that also entails using there packaging which isn't a good solution for a powered speaker in my eyes.
Nice little synopsis there. I do know many things that may not be audio related, at the same time I am not afraid of technology. I worked with a company developing capacitive touch screen more than 20 years ago, also solar panels for what was then Arco Solar about the same period of time. My first payment from an outside company was actually from Arco Solar. I have just been there to early for most of these things to be commercially viable at the time. It's time for well done dsp controlled speakers. I turned down a chance to work for Sonos, I wouldn't do it. As the young crowd is aging who grew up with internet radio and streaming music start to make some real money they will find they want better sound. I don't want to compete with the $150.00 JBL tower speakers you can see at Fry's every weekend in the paper. I want to give the public what I think they truly want. Today you have to meet the style requirements also, Apple has changed the game on industrial design. No more funky wood boxes are going to meet that desire. I think there are many on this site who are already doing this on a diy basis as far as control for loudspeakers, the Rephase thread is just one example dealing with just the software side of this. The easy way is to just purchase pre-made modules from others but that also entails using there packaging which isn't a good solution for a powered speaker in my eyes.
I haven't followed this thread properly but 16bit is fine as a final delivery medium.
For processing however it is nowhere near enough. The minimum generally used by the cheapest digital eqs is 24bit fixed, the standard 32bit floating point.
George Massenburg (https://en.wikipedia.org/wiki/George_Massenburg) insists on a minimum of 48bit for any digital eq or filter. The eq plug ins which bear his name are 64bit.
GroundSound even uses 76bit internal processing for their digital crossover implementation/plate amps.
Even as a simple volume control 16bit is not really usable without audible negative effects on the SQ. I don't use the volume in itunes or spotify. they are always fully open, I use the 32bit volume control that came with my convertor software as it is substantially more transparent. If I'd use it professionally I wouldn't use that either and invest in an analogue volume controller but for non-critical home-use 32bit is good enough.
But as I said I haven't followed this thread properly and I may have misunderstood and this may not be pertinent to the current discussion.
Kindhornman, Binely,
I see. Your interest is focused on commercializing. Mine is just for myself. I'm certainly not talking about passive crossovers, not even analog. In fact, most of what you describe, except for wireless, seems to be what what the system I describe does albeit with more work on the part of the user - me - which is fine in my case.
Not sure you are listening to what I'm saying at this point. Hopefully someone reads the articles I linked to before and find them inspiring, as I did.
Cheers
I see. Your interest is focused on commercializing. Mine is just for myself. I'm certainly not talking about passive crossovers, not even analog. In fact, most of what you describe, except for wireless, seems to be what what the system I describe does albeit with more work on the part of the user - me - which is fine in my case.
Not sure you are listening to what I'm saying at this point. Hopefully someone reads the articles I linked to before and find them inspiring, as I did.
Cheers
Dsp has disadvantages for active speakers as compared to analog: 1) latency, 2) gain structure, 3) price and 4) quality. My analog filters beat the miniDSP in noise and distortion by 10 or more dB.
Don't over underestimate the importance of point 1. Latency makes a speaker impossible to use as live monitor, difficult to use as an editing monitor, hard to use with video, and unsuitable to integrate in a surround setup with other speakers.
Don't over underestimate the importance of point 1. Latency makes a speaker impossible to use as live monitor, difficult to use as an editing monitor, hard to use with video, and unsuitable to integrate in a surround setup with other speakers.
The facts dont count....!
The above is not correct!
What it is, is a perfect example of out of date technical limitations and old school thinking....Sorry vacuphile but this post and the attitude behind it is actually very dangerous....! It might even (if not corrected) mislead and or put off designers from exploring new product development.
I wont list all the obvious examples, (Google will inform) but suffice to say ALL top flight live sound rigs and most top music and movie mastering suites and live TV broadcasts are run on high end digital hardware / software.
Even the old school "Analog shrine" Abbey Road recording studio has just installed a cutting edge Dolby Atmos movie mastering suite.....All DSP controlled and in perfect sync with the video.
This will also be used to record next gen 3D sound scapes for music only as well as movie scores.
Please dont even think about using Mini DSP as any kind of benchmark or reference.....That performance level is already way off the pace.
I recently attended a demo of Lab Gruppen / Dolby Atmos (they bought the amazing Dolby Lake IP) live sound / video event....Wow! It was cabled so not wirless but end to end latency of under 3 ms and that included steerable DSP controlled line arrays on all 9.4 channels.....Serious DSP power!!
Hope this helps keep everyone focused on what is already being implemented in high end DSP audio / video as well as ways we can scale it down to domestic levels.
Cheers
Derek.
The above is not correct!
What it is, is a perfect example of out of date technical limitations and old school thinking....Sorry vacuphile but this post and the attitude behind it is actually very dangerous....! It might even (if not corrected) mislead and or put off designers from exploring new product development.
I wont list all the obvious examples, (Google will inform) but suffice to say ALL top flight live sound rigs and most top music and movie mastering suites and live TV broadcasts are run on high end digital hardware / software.
Even the old school "Analog shrine" Abbey Road recording studio has just installed a cutting edge Dolby Atmos movie mastering suite.....All DSP controlled and in perfect sync with the video.
This will also be used to record next gen 3D sound scapes for music only as well as movie scores.
Please dont even think about using Mini DSP as any kind of benchmark or reference.....That performance level is already way off the pace.
I recently attended a demo of Lab Gruppen / Dolby Atmos (they bought the amazing Dolby Lake IP) live sound / video event....Wow! It was cabled so not wirless but end to end latency of under 3 ms and that included steerable DSP controlled line arrays on all 9.4 channels.....Serious DSP power!!
Hope this helps keep everyone focused on what is already being implemented in high end DSP audio / video as well as ways we can scale it down to domestic levels.
Cheers
Derek.
I am talking about stand alone speakers, not integrated systems - high end or otherwise - because that is what the design goal is.
Scale.
Me too.....
The only difference in what this thread is looking at and what the big commercial stuff looks like is scale and where the boxes are.....ie external amps with DSP Vs internal amps with DSP.
More and more of the big commercial rigs are including more internal (inside or bolted on to the back the speaker) processing power.
They still tend to have one master control box which is used by the sound engineer to set up the event.
No analog system can come close to achieving what a good DSP system can.
Mini DSP is not an appropriate reference for this thread....Flagship ESS DAC is an appropriate starting point for reference.
All DAC's can & will be upgraded given time and R&D so its good to have the DAC on a plug in card....Same goes for the main DSP chip....Todays pack leader is yesterdays old dog!
Cheers
D.
Me too.....
The only difference in what this thread is looking at and what the big commercial stuff looks like is scale and where the boxes are.....ie external amps with DSP Vs internal amps with DSP.
More and more of the big commercial rigs are including more internal (inside or bolted on to the back the speaker) processing power.
They still tend to have one master control box which is used by the sound engineer to set up the event.
No analog system can come close to achieving what a good DSP system can.
Mini DSP is not an appropriate reference for this thread....Flagship ESS DAC is an appropriate starting point for reference.
All DAC's can & will be upgraded given time and R&D so its good to have the DAC on a plug in card....Same goes for the main DSP chip....Todays pack leader is yesterdays old dog!
Cheers
D.
Overkill, you are right, it is the future, and I am convinced that no loudspeaker energized by fusion generated electricity will still be using analog xovers. But for now .... just the ess chip you mention costs more than an analog xover of comparable quality. And then you don't even have a dsp, just the dac.
Last edited:
- Status
- Not open for further replies.