Welcome, Guest
Username: Password: Remember me

NOTE: This is a "Community" forum. Please be mindful that community members are here to help as part of a community effort. We therefore appreciate your effort in keeping this forum a happy place!

If you have a specific issue (e.g. hardware, failure) and want help from our support team, please use our tech support portal (Support menu - > Contact Us).
Thanks a lot of your help in making a better community.

TOPIC: Re: miniSHARC *not yet released*

miniSHARC *not yet released* 8 years 5 months ago #4573

  • KAMIKAZE
  • KAMIKAZE's Avatar
miniDSP share some info: www.minidsp.com/forum/6-diy-hifi-project...imit=6&start=18#4555
- miniSHARC is the PCB only processing board with I2S in/out and with the FIR filter banks. With the same format as all the miniDSP boards, it will indeed allow you to load your custom FIR filters. I think that this could be a good fit for your project. i.e. miniSHARC + miniDSP 2x8 for all 8 outputs.. This is a bit down the line because it will require some new firmware/new software. But we'll release some beta version of the code later on.

Ok, nice. Questions, comments, concerns? Yes! For sure:
Crumboo wrote:
@devteam

Thank you for your kind answer! I'm looking forward to hear more about this new product! :) If the card allows 12k taps, it will mean 2k taps for each channel in my stereo 3-way system. I think it will be enough, but I have to do some more calculations...

I assume this setup would do:

miniDIGI (allowing spdif inputs) -(IS2)-> miniSHARC (multiway FIR crossover) -(IS2)-> miniDSB 2x8 (for more more room correction/eq if necessary + providing multichannel analog outputs)

And from me:

- why we need 2x8 after miniSHARC? Only for analog outputs? If miniSHARC can do FIR filtering, so we have already filtered channels on output? Yes?
- can we get USB input to it?
- what about external DAC's, clock? Without using internal like in miniDSP does.
- no oversampling? can we have two quartz 44k and 48k?

All my questions can be a proposal to hi-end board, so it's will allow to create more quality sound. We need just digital FIR filtering board that allow us to connect external DAC's and then 8-channel volume control and then of course amplifiers.. that can get on input digital signal and have 8-channel digital outputs that this board can filter with FIR, so on output we can have already filtered signals.
Last Edit: 8 years 5 months ago by KAMIKAZE.
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 5 months ago #4608

  • devteam
  • devteam's Avatar
  • Online
  • Administrator
  • Posts: 8748
  • Thank you received: 1123
  • Karma: 161
@ Kamikaze,

Thanks for your interest in this platforms.
At this point of time, a lot of information is preliminary and instead of repeating ourselves multiple times in many forum posts(or saying too much about a product that is quite new), we're planning to release a clear set of information and publicize more details once the products are ready to ship.

By then it will be easier to understand what each of these platforms will be able to do.

Thanks for your patience! :-)

DevTeam
miniDSP, building a DSP community one board at a time.
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 5 months ago #4616

  • Jakster
  • Jakster's Avatar
  • Offline
  • Junior Boarder
  • Posts: 24
  • Thank you received: 4
  • Karma: 3
@ devteam
Thanks for your patience!

I have absolutely no patience what so ever :lol:

This is the suggestion box by which I assume that you´re looking for good ideas from users. So even though you might want us to wait for additional information I´d still like to propose a suggestion.

Are you thinking about doing convolution using fft/ifft instead of time domain convolution??

The reason I ask is that I went on the bruteFIR page and saw his method and he can output 26 channels of 131072 filter taps at 48kHz on a 1GHz athlon. I went into the ADSP 21369 datasheet to see benchmark times and made a little spreadsheet comparing execution times (time domain convolution vs. frequency domain multiplication). The below noted calculations are execution time for filters only - not Input/output handling, interrupt routines, data movement etc etc.

Here´s my results:

First time domain convolution
sample frequency: 96000 Hz
filtersize: 131072 taps
nr of ch: 3

time 1 tap 1 channel: 0,00125 u seconds (from datasheet)
time 1 filter: 15728640 u seconds
time filter nr of channels: 47185920 u seconds
time sum (seconds): 47,185920 seconds
Coefficients pr. second 800000000

Now I know sampling frequency and filtersize is too high for time domain execution (it would take 47 seconds to process 1 second) but coefficients pr. second would be a constant (IE. still 800MMAC with 48000kHz and 2000 tap filter)

Let´s continue to frequency domain (multiplication instead of convolution)

sample frequency: 96000Hz
filtersize: 131072 taps
nr of ch: 3

time 1024 point fft: 23,2 u seconds (from datasheet)
time fft filtersize: 2969,6 u seconds
time multiply fft_coef with fft_audio: 0,00125 u seconds (datasheet, 1 tap FIR filter used)
time filter 1 channel: 163,84 u seconds
time filter nr of channels: 491,52 u seconds
time ifft 1 channel: 2969,6 u seconds
time ifft nr of channels: 8908,8 u seconds
time sum (u seconds): 12369,92 u seconds
time sum (seconds): 0,01236992 seconds
time correction filtersize vs. Fs: 0,00906
Taps pr. second 4,16653E+12

The time correction comes from the fact that you´re processing 1,36 seconds of audio (131072/96000)

nr. of times fft conv. Is faster: 5208

Now there is a lot of ifs, buts and maybe´s in this. I made the following assumptions:

-Time execution from 1024 to 131072 is assumed to be a factor of 128 - as Anders Torger (BruteFIR) discovered - large FFT´s take longer - He solved it by "partioned convolution", which also results in lower I/O delay.

-Time for a complex multiplication of filter*audio is one MAC (might take 2)

-ifft takes the same amount of time as an fft

I know fft convolution it will take considerately longer on the mini SHARC than what I came up with, but the numbers do indicate that it should be possible (maybe even for a 4 way). Also the program that has to be developed for this is significantly bigger and more complex than the openDRC program.

Kind regards

Jakster

PS: I haven´t checked my benchmark times against other theoretical benchmarks on the internet - so if someone sees an error in this post - feel free to comment!

PPS: For those who have nothing better to do Anders wrote a paper on crosstalk cancellation using partioned convolution - found here: www.acoustics.net/objects/pdf/article_farina04.pdf
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 4 months ago #4732

  • HYPERTUNE
  • HYPERTUNE's Avatar
  • Offline
  • Fresh Boarder
  • Posts: 8
  • Karma: 0
It would be nice to get some idea of what the minisharc will be capable of.

Can you at least tell us if it could act as a 2x8 crossover, i2s in/out? (I'm asking about the hardware, I assume suitable software for programming might arrive later.)

Edit: basically I want to know if it will have 8 channels of i2s output.
Last Edit: 8 years 4 months ago by HYPERTUNE.
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 4 months ago #4759

  • HYPERTUNE
  • HYPERTUNE's Avatar
  • Offline
  • Fresh Boarder
  • Posts: 8
  • Karma: 0
Looking at the eval board for the ADSP-21369, it seems as though it should not be a problem to get 8 channels of i2s out.

www.analog.com/static/imported-files/eva...nual%20Rev%202.2.pdf

I hope this is available on the minisharc, and I hope there will be a software solution to help programming a stereo 4 way crossover.

I actually don't even need FIR, and would be happy with IIR only. Processing wise, the 2x8 does all I need, I just don't want to use the onboard dac, and don't need analogue in either.
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 4 months ago #4772

  • Crumboo
  • Crumboo's Avatar
  • Offline
  • Expert Boarder
  • Posts: 144
  • Thank you received: 14
  • Karma: 13
I sure hope for 8 ch i2s out as well! It would be nice to see some more info about this product soon...
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 4 months ago #4790

  • devteam
  • devteam's Avatar
  • Online
  • Administrator
  • Posts: 8748
  • Thank you received: 1123
  • Karma: 161
@ All,

Sorry for the delay in giving more info on this... There has been a fair amount of Beta testing that going on for the past few weeks and few changes made it to the products. We just want to release this product with a neat set of features.

The I/O configuration will indeed allow on the hardware 8ch I2S in and out.(i.e. Hardware configuration) However the firmware (and software) required to make that configuration will take some time since it's not just a matter of building a new config (hardware is VERY different to firmware + software). So it's very much like the miniDSP kit when it started for those who remember, we didn't release 12plug-ins in one day. :-)

Hoping this clarifies, we'll try to release more info (starting by the user manual) by end of this week. Some lucky one already having fun with some room correction or custom engineering FIR filter design.

DevTeam
miniDSP, building a DSP community one board at a time.
The administrator has disabled public write access.
The following user(s) said Thank You: curryman

Re: miniSHARC *not yet released* 8 years 4 months ago #4791

  • Crumboo
  • Crumboo's Avatar
  • Offline
  • Expert Boarder
  • Posts: 144
  • Thank you received: 14
  • Karma: 13
Great news! So, what we know so far:
  • 8 channel I2S in/out
  • Possible to load custom FIR filters
  • At least 12k taps
  • Same format as miniDSP boards

Very promising! :)
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 4 months ago #4793

  • curryman
  • curryman's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 792
  • Thank you received: 181
  • Karma: 101
This indeed sounds very promising :cheer:

One more info that could be interesting:
I think that this could be a good fit for your project. i.e. miniSHARC + miniDSP 2x8 for all 8 outputs..
Not sure though what Devteam has in mind:
  • miniSHARC for overall FIR + miniDSP 2x8 for crossover
  • miniSHARC multichannel incl. crossover + miniDSP 2x8 as D/A converter

I am hungry for more info :whistle:
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 4 months ago #4795

  • Crumboo
  • Crumboo's Avatar
  • Offline
  • Expert Boarder
  • Posts: 144
  • Thank you received: 14
  • Karma: 13
I believe (hope!) it to be be the second alternative (8 ch out indicates that it can be used as crossover). Without the possibility to rout signals internally, I don't really see the point...?

If the miniSHARC will function as I hope, I'd use the following setup for my system:

miniDIGI (as input selector and convert digital signals to I2S) -> miniDSP (for 2 ch room correction - it would cost to many taps to do this with FIR in the bass region I guess) -> miniSHARC (6 or 8 ch crossover/driver equalization) -> 8 ch DAC

Makes sense? :)

Is there any info about when we can expect the release of this?
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 4 months ago #4796

  • Crumboo
  • Crumboo's Avatar
  • Offline
  • Expert Boarder
  • Posts: 144
  • Thank you received: 14
  • Karma: 13
One more thing i really hope for in this product:

That the delay and gain of the channels can be adjusted individually (without changing the FIR filter coefficients). :)
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 4 months ago #4800

  • curryman
  • curryman's Avatar
  • Offline
  • Platinum Boarder
  • Posts: 792
  • Thank you received: 181
  • Karma: 101
Good idea, however actually that is more a matter of the software interface, isn't it?

I wouldn't care whether everything inside the DSP is done in one overall or several filters, if I can adjust delay and gain separately in the user interface.
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 4 months ago #4802

  • Crumboo
  • Crumboo's Avatar
  • Offline
  • Expert Boarder
  • Posts: 144
  • Thank you received: 14
  • Karma: 13
curryman wrote:
Good idea, however actually that is more a matter of the software interface, isn't it?

I wouldn't care whether everything inside the DSP is done in one overall or several filters, if I can adjust delay and gain separately in the user interface.

I totally agree. :)
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 4 months ago #4819

  • Crumboo
  • Crumboo's Avatar
  • Offline
  • Expert Boarder
  • Posts: 144
  • Thank you received: 14
  • Karma: 13
One more thing on my wish list for this product (a software thing probably): the ability to do some simple arithmetic like getting the sum and difference of channels...:)

It would of course be even better with a "custom routing" mode where each output channel could be specified as "function of" one or several input channels, where "function of" would include FIR filters, arithmetics (sum, difference, product, ratio, correlation, ...), delay, scaling, phase inversion etc. Could be used for experimenting with channel upmix/downmix and other interesting things...;)
Last Edit: 8 years 4 months ago by Crumboo.
The administrator has disabled public write access.

Re: miniSHARC *not yet released* 8 years 4 months ago #4832

  • KAMIKAZE
  • KAMIKAZE's Avatar
And price for all this?
The administrator has disabled public write access.
Moderators: devteam