Go Back   Home > Forums > Source & Line > Digital Line Level

Digital Line Level DACs, Digital Crossovers, Equalizers, etc.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 5th May 2012, 01:53 PM   #181
diyAudio Member
 
steph_tsf's Avatar
 
Join Date: Mar 2008
For audio enthusiasts like we have on diyAudio, the silicon that we need is the digital engine sitting in the SWEEX SC016 silicon (thus, without ADCs and DACs), upgraded to 24bit/96kHz. No doubt it will come in due time, inexpensive, as this is a mass market.
One may consider a group buy if such silicon remains ignored by Digi-Key, Mouser and Farnell.
In the meantime we may try hooking the SWEEX SC016 on CuBox or Raspberry Pi.
Later on we'll base on this experience for hooking the SWEEX SC016 on a Android 4.0 device.
A purely software affair, thus !
The CuBox and Raspberry Pi solutions are no short-term solution. They are suited to active speakers, embedded into active speakers. Cost remains low because of no graphical user interface..
The Android 4.0 solution would provide a high quality user interface, easing the setup. Especially using a 10 inch tablet and a physical keyboard, as development system for the XOver.
Once the Android 4.0 solution is completed, there would be a way to migrate it on CuBox and Raspberry Pi.
  Reply With Quote
Old 5th May 2012, 02:27 PM   #182
diyAudio Member
 
steph_tsf's Avatar
 
Join Date: Mar 2008
Cmedia CM6620 is a USB2.0 high-speed audio processor that can support the latest USB Audio Device Class Definition V2.0 and 7.1-channel true high-definition audio and Line-In/Mic-In (all up to 192KHz/24bit). The problem with the CM6620 is that it provides an HDA link (Intel High Definition Audio) instead of four I2S lanes or one TDM.

The VIA VT1730 is an USB 2.0 (480 Mb/s) 8-channel, 24-bit/192 kHz audio controller specifically designed to achieve cinema-quality audio recording and playback in high fidelity USB applications. The highly integrated single chip USB audio solution is embedded with a number of crucial modules, including a standard 8032 MCU core, USB transceiver, five I2C masters and one slave controller, an I2S controller, a serial flash controller, four SPI ports, phase locked loop (PLLs) with internal oscillator, and a 3.3 V~2.0V regulator. The VIA VT1730 supports multi-channel PCM digital audio format up to 24-bit resolution and 192 kHz sample rate. An S/PDIF interface with two stereo output ports and one stereo input port supports up to a 24-bit resolution and 192 kHz sample rate.

Are there other USB 2.0 audio controllers nowadays, outputting four I2S lanes or one TDM ? Equipped with a S/PDIF input ?
Are there people writing Android 4.0 drivers for USB 2.0 audio controllers ?
  Reply With Quote
Old 5th May 2012, 08:05 PM   #183
diyAudio Member
 
steph_tsf's Avatar
 
Join Date: Mar 2008
Quote:
Originally Posted by steph_tsf View Post
VIA VT1730 .... an I2S controller
This is what we are looking for. Forget about the five I2C masters.

Last edited by steph_tsf; 5th May 2012 at 08:07 PM.
  Reply With Quote
Old 6th May 2012, 09:16 AM   #184
diyAudio Member
 
s3tup's Avatar
 
Join Date: Dec 2004
Location: Israel
So, what about hardware?
Should we really stick to PC+USB thing? PC+PCI are easily obtainable, with 4 I2S outputs, slaving ability etc, so if we are going with PC, why not? There is no logical difference between USB and PCI - PC is always there.

And if we are going for less than 100% best DACs - then PCI multichannel cards are fine for us...

Yet, PC thing is bad - that's why we want stand-alone hardware DSP.

If we won't target the CPU-hungry FIRs (and we won't - with available cheap-end CPUs and DSPs), we are goind with IIRs.

Why do we need DSP for? Do we want to change presets every day/week? I guess so, till we find the right preset, and leave it for a while...
We need it to be hardware, perfect SQ, and active. Why won't we just stick to analog circuits? These are cheap, somewhat better quality than DSPs (no bits truncations etc), yet painfull to modify. Digital IIRs mimic the analog circuits anyway...

How about PC based crossover to find the right filers, and then build these filters in analog-hardware? It's cheap way (nice sounding soundcards with 8ch outputs are already there and not that pricey, analog boards are opamp price dependant), flexible as your PC software could be, and as long as your digital filter's transfer functions match these in analog cookbooks, you can transfer digital domain filters to analog hardware.

Software which will do the PC-based sim, and calculate the parts for analog circuits? I already have a program that generates biquads and transfers them to foobar plugin, which passes the sound thru my biquad chains and plays thru multichannel soundcard. Everything is in real time.
Analog circuits sim is already there, but i had issues with modeling the analog transfer functions in the digital domain (analog TF to biquads). I guess i should put bode plots of analog and digital filters next to each other, check whether they match, and say hooray if they do. Same bode plots = same transfer functions = same sound.
  Reply With Quote
Old 6th May 2012, 12:08 PM   #185
diyAudio Member
 
steph_tsf's Avatar
 
Join Date: Mar 2008
Quote:
Originally Posted by s3tup View Post
How about PC based crossover to find the right filers, and then build these filters in analog-hardware?
Done this already using a WinXP PC executing SynthMaker, connecting on the SWEEX SC016 8-channel DAC. Using RBJ-cookbook filters (IIR BiQuads).

Actually it is a little bit more complicated : you need VAC (Virtual Audio Cable) for telling Windows to route the audio (coming from Winamp or any other app) to ASIO, and you need ASIO4ALL for telling Windows to operate the SWEEX SC016 8-channel DAC under ASIO.

This way some time ago I started prototyping crossovers in digital domain, translating them into analog afterwards. I wrote a small Visual Basic application, easing the process. It went published in reply #12 of diyAudio post "IIR_Lab : a design help for digital audio filters" and is still available.

When you do this you realize the importance of using delay lines for properly time aligning the drivers, the woofer emission centre being usually recessed (delayed), compared to the tweeter emission centre.
Even the most straightforward 1st-order crossover needs such time alignment. Builing a crossover without paying attention to time alignment is futile !

How would you implement delays in analog ? You would cascade phase shifters, approximating a pure delay function from DC to some intermediate frequency, ideally 20 kHz. You can compensate tens of microseconds like this, using 4 or 8 opamps in the signal flow, per delay line. If you want to compensate milliseconds, you'll need far more opamps in the signal flow. A possibility is to make a tradeoff, like implementing a pure delay function, not from DC to 20 kHz, but from DC to 2 kHz or so. At the expense of phase distorsion above 2 kHz or so.

What if you want a crossover exhibiting no phase distorsion ? The usual solution is the Lipshitz-Vanderkooy (delay compensated) arrangement, not to be confused with the Linkwitz-Riley. And now guess what ? For the Lipshitz-Vanderkooy (delay compensated) arrangement, you need a delay line in the highpass elaboration, accounting for the propagation delay introduced by the lowpass section.

Crossovers thus intrinsically need delay lines (at least one for time-aligning the transducers, and one more for avoiding phase distorsion), and delay lines are unpractical using analog. The only practical way is digital, especially when wanting to avoid phase distorsion.

This being said, having already done this using a WinXP PC executing SynthMaker, connecting on the SWEEX SC016 8-channel DAC, please tell me how can I connect and program an Android 4.0 device connecting on the same SWEEX SC016 8-channel DAC. This would be my next Open Source DSP XO, if you don't mind.

You see, I'm asking myself why people keep asking about an "Open Source DSP XO", because the WinXP + SynthMaker + VAC + ASIO4ALL + SWEEX SC016 8-channel DAC is already there since years. You can trace it back since the origin in SynthMaker forum. Seems I am the only one publicly experimenting this. Nobody has publicly reported experiments using a better 8-Channel DAC. Nobody has publicly reported experiments using 24bit/96kHz natively within Winamp with audio files supplied by HDracks.

The newest SynthMaker marketing strategy doesn't help : now they disable the 8 ASIO channels on the trial version.

People don't get attracted by digital crossovers, because they know that the crossover alone, is only the emerged part of an iceberg, and can't guarantee a success.

People may welcome a total solution, a step-by-step solution, encompassing everything about multiway speaker design :
- selecting a woofer
- mesuring the T/S parameters using the added mass method
- modelling an enclosure like closed box / bass-reflex / 4th order bandpass / 8th order bandpass
- building the enclosure
- measuring the gain, phase and distorsion of the woofer in the enclosure
- measuring the polar diagram (directivity) of the woofer in the enclosure
- selecting a tweeter
- mounting the tweeter on the enclosure
- measuring the gain, phase and distorsion of the tweeter on the enclosure
- measuring the polar diagram (directivity) of the tweeter on the enclosure
- selecting a crossover topology (1st order, Linkwitz-Riley 4th-order, Lipshitz-Vanderkooy Bessel 4th-order delay compensated)
- selecting enclosure-dependent corrections like time alignment, baffle step, standing wave resonances
- selecting driver-dependent corrections like high frequency dampers, shelving
- selecting a low frequency corrections like a 2nd order Linkwitz Transform
- implementing the crossover and the corrections in digital
- connecting the power amps and the speaker drivers
- measuring the gain, phase and distorsion of the active 2-way speaker
- finely tuning the parameters
- properly documenting the finalized design, for reproduction purposes

Which means that such crossover design software needs a full-size PC (big screen, physical keyboard, mouse), for compiling the software that needs to run on some small and inexpensive crossover hardware like a few ARM Cortex-M0/M3/M4 "S/PDIF DSP+DAC bricks" designed by Richard Dudley. Or Cubox, or Raspberry Pi, or any Android 4.0 device, supplying audio data to an external USB audio peripheral (say the SWEEX SC016 8-channel DAC to begin with).

And you still have many choices about how to implement the digital filters !
You could base on IIR_Lab and program the resulting IIR BiQuads.
You could base on an analog simulation package like LTspiceIV, extract the impulse response of the filter that you want, and program it as a digital FIR (Richard Dudley).
You could enter the unfiltered gain and phase response of your driver (mounted in the enclosure), enter the desired gain response (say a 4th-order Bessel Lowpass), and a Math package like MathCad could calculate the polynomial fraction in p that you need as filter, that you would transform into a z polynomial fraction in digital, then decompose into IIR BiQuads. Or use the FIR way : graph the impulse response of the polynomial fraction in p, that you would use as FIR.
You could play the FIR card since the beginning. Amazing how some 32-tap lowpass FIRs can be selective, and phase linear. With the complementary highpass, quite selective too.
With the FIR approach, there should be a way to enter the unfiltered gain and phase response of your driver (mounted in the enclosure), enter the desired gain response (say a 4th-order Bessel Lowpass), and a Math package like MathCad or Matlab could calculate the FIR coefficients that you need as filter.

That would be the challenge.

Last edited by steph_tsf; 6th May 2012 at 12:36 PM.
  Reply With Quote
Old 6th May 2012, 03:30 PM   #186
ChrisPa is offline ChrisPa  United Kingdom
diyAudio Member
 
Join Date: Jan 2005
Location: Saddleworth
Why not analogue?

Because I want to keep my source signal in the digital domain, unattenuated as long as possible

Because I can get rid of all capacitors, and I can get rid of all component degradation in the analogue components

Because I can change the eq whenever I want without having to build any circuits.

Because the title of the thread is DSP crossovers, not analogue

I can't see the advantage of the analogue circuit
  Reply With Quote
Old 6th May 2012, 07:46 PM   #187
diyAudio Member
 
s3tup's Avatar
 
Join Date: Dec 2004
Location: Israel
Quote:
Originally Posted by steph_tsf View Post
Done this already using a WinXP PC executing SynthMaker, connecting on the SWEEX SC016 8-channel DAC. Using RBJ-cookbook filters (IIR BiQuads).

Actually it is a little bit more complicated : you need VAC (Virtual Audio Cable) for telling Windows to route the audio (coming from Winamp or any other app) to ASIO, and you need ASIO4ALL for telling Windows to operate the SWEEX SC016 8-channel DAC under ASIO.
As long as you need foobar-only platform, my DSP plugin and configurator program makes it easy. I've tried the SynthMaker, but found it cumbersome. kX drivers gave a bit more control in the DSP land.

Quote:
Originally Posted by steph_tsf View Post
This way some time ago I started prototyping crossovers in digital domain, translating them into analog afterwards. I wrote a small Visual Basic application, easing the process. It went published in reply #12 of diyAudio post "IIR_Lab : a design help for digital audio filters" and is still available.
I've tried the program - it really discovers the bottlenecks of hardware IIRs, since then i try to avoid messing with a response on lowish frequencies in digital systems

Quote:
Originally Posted by steph_tsf View Post
When you do this you realize the importance of using delay lines for properly time aligning the drivers, the woofer emission centre being usually recessed (delayed), compared to the tweeter emission centre.
Even the most straightforward 1st-order crossover needs such time alignment. Builing a crossover without paying attention to time alignment is futile !

How would you implement delays in analog ? You would cascade phase shifters, approximating a pure delay function from DC to some intermediate frequency, ideally 20 kHz. You can compensate tens of microseconds like this, using 4 or 8 opamps in the signal flow, per delay line. If you want to compensate milliseconds, you'll need far more opamps in the signal flow. A possibility is to make a tradeoff, like implementing a pure delay function, not from DC to 20 kHz, but from DC to 2 kHz or so. At the expense of phase distorsion above 2 kHz or so.

What if you want a crossover exhibiting no phase distorsion ? The usual solution is the Lipshitz-Vanderkooy (delay compensated) arrangement, not to be confused with the Linkwitz-Riley. And now guess what ? For the Lipshitz-Vanderkooy (delay compensated) arrangement, you need a delay line in the highpass elaboration, accounting for the propagation delay introduced by the lowpass section.

Crossovers thus intrinsically need delay lines (at least one for time-aligning the transducers, and one more for avoiding phase distorsion), and delay lines are unpractical using analog. The only practical way is digital, especially when wanting to avoid phase distorsion.
Depends on what do you target. Linear phase, phase coherent, or just something that sounds nice and have no nulls in xover point.
If we are talking about no-compromise system, then full-blown FIRs with lots of taps do the trick by inverting driver's impulse response and thus linearizing every driver as well as whole system. $$CPU$$ .

Acoustical centers matching ain't that bad in 3-4way systems with conventional drivers. I've done measurements on a set of SS18W, 12M and XT25 (Troels's Ekta) and found out the drivers alone align properly in phase when you skew the baffle for 7degrees backwards.
At the crossover corner frequencies. So there is no need to fix anything - just skew the box.
This trick won't work with horns, dome midranges and 2-ways with large/deep (>5") woofer. So digital wins. Especially with horns which have crazy response, and way too sensitive to components involved in the crossover whichever kind would it be.

Filter's phase response is quite another issue and i'm unshure what is best - to keep the phase in "natural way", or flatten it out with symmetric phase-compensated FIRs.

Quote:
Originally Posted by steph_tsf View Post
This being said, having already done this using a WinXP PC executing SynthMaker, connecting on the SWEEX SC016 8-channel DAC, please tell me how can I connect and program an Android 4.0 device connecting on the same SWEEX SC016 8-channel DAC. This would be my next Open Source DSP XO, if you don't mind.
I dream of backsynced SPDIF/I2S on android Perfect media center - wifi, lan, usb host, sweet screen.

Quote:
Originally Posted by steph_tsf View Post
You see, I'm asking myself why people keep asking about an "Open Source DSP XO", because the WinXP + SynthMaker + VAC + ASIO4ALL + SWEEX SC016 8-channel DAC is already there since years. You can trace it back since the origin in SynthMaker forum. Seems I am the only one publicly experimenting this. Nobody has publicly reported experiments using a better 8-Channel DAC. Nobody has publicly reported experiments using 24bit/96kHz natively within Winamp with audio files supplied by HDracks.

The newest SynthMaker marketing strategy doesn't help : now they disable the 8 ASIO channels on the trial version.

People don't get attracted by digital crossovers, because they know that the crossover alone, is only the emerged part of an iceberg, and can't guarantee a success.
I've heard many statements about PC-XO thing. In two words - buggy and complicated.
Making chains with virtual cables, virtual ASIOs, VSTs etc... I couldn't make any VST host work on my PC, and i can't complain i have lack of experience in PC/Microsoft world. If it works now, it dies the next day taking my tweeters.
PCs are for debug only. That's why so lacking hardware platforms such as DCX2496 and miniDSP gained popularity - they don't fail so frequently as PC does, and they are "all-in-one" solutions. Easy, relatively cheap, common and well-studied.

Quote:
Originally Posted by steph_tsf View Post
People may welcome a total solution, a step-by-step solution, encompassing everything about multiway speaker design :
- selecting a woofer
- mesuring the T/S parameters using the added mass method
- modelling an enclosure like closed box / bass-reflex / 4th order bandpass / 8th order bandpass
- building the enclosure
- measuring the gain, phase and distorsion of the woofer in the enclosure
- measuring the polar diagram (directivity) of the woofer in the enclosure
- selecting a tweeter
- mounting the tweeter on the enclosure
- measuring the gain, phase and distorsion of the tweeter on the enclosure
- measuring the polar diagram (directivity) of the tweeter on the enclosure
- selecting a crossover topology (1st order, Linkwitz-Riley 4th-order, Lipshitz-Vanderkooy Bessel 4th-order delay compensated)
- selecting enclosure-dependent corrections like time alignment, baffle step, standing wave resonances
- selecting driver-dependent corrections like high frequency dampers, shelving
- selecting a low frequency corrections like a 2nd order Linkwitz Transform
- implementing the crossover and the corrections in digital
- connecting the power amps and the speaker drivers
- measuring the gain, phase and distorsion of the active 2-way speaker
- finely tuning the parameters
- properly documenting the finalized design, for reproduction purposes
Pretty too much for single application
Half of enclosure quirks could be handled in DSP, as well as speaker's FR issues. No impedance nor sensitivity quirks in active systems too.
So instead of doing the workflow from scratch, why not fix whatever is already done?
No lowish bass? Throw a Linkwitz transform. Curvy mid's response? A bit of notching cures that too. Drivers not in phase? Spin the phase with all-pass, match timings with delays... Giving a good control over DSP part, and focusing in the DSP as a cure to all speaker problems could be better and much easier, than doing every described point halfway.

I don't think the box design thing should be included in such software.
Speaker design should be split into 2 parts.
First part - is preventing quirks by proper enclosure design and driver selection.
Second part - when the enclosure is already done and drivers already purchased - fixing the quirk leftovers. That's where DSP takes over.

Quote:
Originally Posted by steph_tsf View Post
Which means that such crossover design software needs a full-size PC (big screen, physical keyboard, mouse), for compiling the software that needs to run on some small and inexpensive crossover hardware like a few ARM Cortex-M0/M3/M4 "S/PDIF DSP+DAC bricks" designed by Richard Dudley. Or Cubox, or Raspberry Pi, or any Android 4.0 device, supplying audio data to an external USB audio peripheral (say the SWEEX SC016 8-channel DAC to begin with).
Single software, different DSP platforms i'd add.
Targeting not only DIY devices, but other, commercial units, as popular miniDSP and DCX. Then your designs won't be bound to particular platform, and could be "crosscompiled" to any platform you need. Kind of "Windows" for DSPs.

Quote:
Originally Posted by steph_tsf View Post
And you still have many choices about how to implement the digital filters !
You could base on IIR_Lab and program the resulting IIR BiQuads.
You could base on an analog simulation package like LTspiceIV, extract the impulse response of the filter that you want, and program it as a digital FIR (Richard Dudley).
You could enter the unfiltered gain and phase response of your driver (mounted in the enclosure), enter the desired gain response (say a 4th-order Bessel Lowpass), and a Math package like MathCad could calculate the polynomial fraction in p that you need as filter, that you would transform into a z polynomial fraction in digital, then decompose into IIR BiQuads. Or use the FIR way : graph the impulse response of the polynomial fraction in p, that you would use as FIR.
You could play the FIR card since the beginning. Amazing how some 32-tap lowpass FIRs can be selective, and phase linear. With the complementary highpass, quite selective too.
With the FIR approach, there should be a way to enter the unfiltered gain and phase response of your driver (mounted in the enclosure), enter the desired gain response (say a 4th-order Bessel Lowpass), and a Math package like MathCad or Matlab could calculate the FIR coefficients that you need as filter.

That would be the challenge.
I think the combination of IIR and FIR filters would be best from CPU point of view. Low frequencies would be handled in IIRs - where you can't really handle the room modes nor make some usable-length FIRs (AFAIK). But high-frequencies (mid-hi, from 500 and upwards), where most information comes from, should be handled in best possible way, introducing the FIRs at their best - phase linear, inversed impulse responses of particular drivers etc.


By the way, is it correct that FIRs require lots of taps to dig into low frequencies? Something like Ntaps=Fs/Fo?



So, we need
1. PC Software to configure the different filters - plotting, measurements, filter generator.
- It depends how far we'd like to go. Fist of all, scale-ability, opennes on Math side, user friendliness.

2. DSP core - embedded software in C for MCUs or DSP code for specific DSPs.
- hardware + byte crunching OS.

3. DACs/ADCs/SPDIFs/I2S/Clocking
- loads of solutions over internet - all DACs are built around these blocks. Cooking the DAC and it's output stage is an art of itself, but as long as the signals stay in digital, there are straightforward solutions available.



Ehm, lots of letters





Quote:
Originally Posted by ChrisPa View Post
Why not analogue?

Because I want to keep my source signal in the digital domain, unattenuated as long as possible

Because I can get rid of all capacitors, and I can get rid of all component degradation in the analogue components

Because I can change the eq whenever I want without having to build any circuits.

Because the title of the thread is DSP crossovers, not analogue

I can't see the advantage of the analogue circuit
Why analogue?
Because i want to use my wow-DAC with amazing full-blown TDA1541 in golden sockets etc. I want to use my turntable, radio or whatever, tube preamp etc.
I want to use same DAC for the speaker system and headphones rig.
I don't want to do redundant AD-DA conversion if i'd like to use analog source (me myself don't want to use it ).

And it's a bit cheaper, and analog-ish. Call it Analog-DSP? Linkwitz calls it "ASP" anyway

Keeping signals in analog domain brings some advantages, but kills others... These passive elements are bad. Analog is ugly for modifications/tuning.

Last edited by s3tup; 6th May 2012 at 07:59 PM.
  Reply With Quote
Old 6th May 2012, 09:33 PM   #188
diyAudio Member
 
steph_tsf's Avatar
 
Join Date: Mar 2008
A possible DSP target would be the newest NXP LPC4330 featuring a Cortex-M4 enhanced by a Cortex-M0 servicing the Serial GPIO interrupt for generating the required I2S lanes. I just found the NGX Technologies LPC4330 Xplorer board, soon to be released. See attached pictures. Clearly, the J8 header is conveying SPI1 for the Codec control bus. Can somebody tell me if the J8 and J10 headers are conveying enough Serial GPIO lines for the Codec data bus (implementing four I2S-out lanes, one I2S-in lane, plus the associated bit clock and frame sync)?
Attached Images
File Type: jpg LPC4330 Xplorer (board).jpg (124.9 KB, 185 views)
File Type: jpg LPC4330 Xplorer (block diagram).jpg (144.3 KB, 180 views)
File Type: jpg LPC4330 Xplorer and USB JTAG (ordering).jpg (297.4 KB, 172 views)
File Type: jpg LPC4330 Xplorer (J8 and J10 headers).jpg (252.7 KB, 166 views)
File Type: jpg LPC4330 Xplorer (XO project).jpg (215.6 KB, 169 views)

Last edited by steph_tsf; 6th May 2012 at 09:44 PM.
  Reply With Quote
Old 6th May 2012, 10:15 PM   #189
diyAudio Member
 
steph_tsf's Avatar
 
Join Date: Mar 2008
NXP Serial GPIO on LPC4330, from NXP website NXP’s New Dual-Core Cortex-M4 and M0 MCU Redefines Digital Signal Control :: NXP Semiconductors
NXP’s Serial GPIO, available for the first time on the LPC4000, allows a developer the flexibility to interface to any non-standard serial interface or to mimic multiple standard serial interfaces (such as I2S, TDM for multi-channel audio, I2C and more).
Question : is there Serial GPIO sample code available, for generating four I2S-out and one I2S-in, sharing a bit clock and a frame sync ?
Additional Serial GPIO info : watch this video : http://ics.nxp.com/support/training/sgpio.sct/

Last edited by steph_tsf; 6th May 2012 at 10:20 PM.
  Reply With Quote
Old 7th May 2012, 07:27 AM   #190
diyAudio Member
 
s3tup's Avatar
 
Join Date: Dec 2004
Location: Israel
Pricey... 170$ for just a CPU is plain pricey...

From price point of view, Android device would be much better choice for the same price range. If you stick to some standart I/O bus (HDMI, USB), then you get hardware-independent solution, with Android as DSP and media player (touchscreen, WiFi, LAN, USB Host, memory). Add a squeezebox player app, and you're done implementing LAN DAC.
There is a GPU which could do the DSP processing just like a CUDA does...
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Volume / Source selector - open source project ? AuroraB Analog Line Level 22 22nd September 2012 02:21 PM
Violet DSP Evolution - an Open Baffle Project cuibono Multi-Way 211 18th May 2010 02:26 AM
Open call for suggestions on Open Source DIY Audio Design gfergy Everything Else 1 15th April 2007 07:33 AM
Open Source, Open Architecture! zenmasterbrian Digital Source 185 23rd February 2007 10:35 PM


New To Site? Need Help?

All times are GMT. The time now is 11:42 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2