Yep. Differences of opinion make the world a more colourful place
(BTW I agree mini dsp really is off topic)
A lot of interessting reading so far but If this thread should carry on lets stay on topic.
Checklist for whats already done:
1: Powersupply for all parts.
2: Software (fir).
3: External dac's (with separate clocks to be used or not).
4: Usb reciver (with separate clock to be used or not).
5: Master clock (to be used for dsp only or as master for all).
So, what parts and how to make them work together ?
Remember we are not aiming for "good enough".
And yes, "Minidsp" is of topic.
Checklist for whats already done:
1: Powersupply for all parts.
2: Software (fir).
3: External dac's (with separate clocks to be used or not).
4: Usb reciver (with separate clock to be used or not).
5: Master clock (to be used for dsp only or as master for all).
So, what parts and how to make them work together ?
Remember we are not aiming for "good enough".
And yes, "Minidsp" is of topic.
Last edited:
re software FIR: I dont know if we should strike it off completely. is there any reason to rule out a cooperative effort between software and hardware? as in correction curves and convolution/impulse responses are worked out in conjunction with an ADC and computer, but then loaded into a DSP which nolonger needs the computer anymore for the realtime processing..
afterall an impulse response doesnt need much by way of memory to be stored in the DSP and then the audio is processed in realtime, rather than having to process your files offline, you just store the IR offline and its then used along with the DSP EQ/filter curves to produce the final result.
afterall an impulse response doesnt need much by way of memory to be stored in the DSP and then the audio is processed in realtime, rather than having to process your files offline, you just store the IR offline and its then used along with the DSP EQ/filter curves to produce the final result.
Last edited:
I didn't mean my comment that way. I meant I think the software (DSP processing) could do with more than just FIR
I had the DEQX HDP-4 at loan recently. I wasn't very impressed considering the price. It was pretty good but not worth the price in my opinion. It did not sound as transparent, open and homegenous as my Lynx Two-B soundcard connected straight to poweramps (using digital volume control).
I have also tried GroundSound and here the difference was even larger. I woudn't be surprised if the most expensive 8 channel miniDSP is better sounding then both of these.
I'm also waiting for the perfect all in unit. I guess I could set up something with PC software and use a high quality soundcard, but I would prefer something separate. The coming DCPL from Hypex is interesting, but I might need FIR filters for my planned bass system. This is also a product I'm looking forward to:
Böhmer Audio
Ok lets go then, source first:
Mac mini/audirvana or Sony scd 777es-with separate dac.
Turntable, a Verdier rip-of and tube riaa "Tubecad".
Preamp also tube based from "Broskie" with Stevens & Billington
tvc's for volume attenuation.
Speakers are 6 feet long ribbons of own construction 95bd 350-20000hz
12 8inch woofers (Seas) / channel 30-350.
Tapped horns for subs (Beyma 12L60).
Crossover 2order, passive for the ribbons (Duelund VSF cu) and dsp/active
for the woofers and subs (4th order).
I will have to leave goal and reference for tomorrow, need some sleep now.
Fair enough - sleep is important
Reason for asking is to get an idear of what you are seeking. I have listened to alot of different system with both passive, FIR and IIR filters. And my experience is that it all can work to a high degree of realism. But the active solution gives a more flexible way of controlling precisely what you want the system to do.
I use active DSP filtering (IIR) for all drivers in my system, and heard others use it for their systems as well, also ribbons and panels, with no problems.
Does my memory serve me right - or is it you that are seeking a DSP with enough dealy, to meet your large horn-subs?
Reason for asking is to get an idear of what you are seeking. I have listened to alot of different system with both passive, FIR and IIR filters. And my experience is that it all can work to a high degree of realism. But the active solution gives a more flexible way of controlling precisely what you want the system to do.
I use active DSP filtering (IIR) for all drivers in my system, and heard others use it for their systems as well, also ribbons and panels, with no problems.
Does my memory serve me right - or is it you that are seeking a DSP with enough dealy, to meet your large horn-subs?
Yes, that's me.
Forgot the amps:
There will be a pair of SET monoblocks ready in a few weeks (C3m/kr300b xls/LL1623) to drive the ribbons. I'm using abletec alc0300 for all other sections.
And for now this is a theoretical disscusion that hopefully will inspier someone more qualified to actually build something.
The flexibility of "Chaparks" filter is what brougth the idea to life.
Forgot the amps:
There will be a pair of SET monoblocks ready in a few weeks (C3m/kr300b xls/LL1623) to drive the ribbons. I'm using abletec alc0300 for all other sections.
And for now this is a theoretical disscusion that hopefully will inspier someone more qualified to actually build something.
The flexibility of "Chaparks" filter is what brougth the idea to life.
DSP is math, DSP chips are calculators built for convolution. Data in data out. Latency is dependent on sample rate, filter length, and specifics of how convolution is calculated, and limitations of how calculation may be performed.
For example, PC based convolution with Sourceforge Convolver for audio has partitioning that allows filter to broken down into blocks resulting in dividing down pass through latency. Throughput is optimized based on computational processing power and balancing partition number v filter length. Similar parallels exist with DSP chips, taking advantage of long word lengths and parallelism of architecture. In the end it is all about multiply, add, and carry. Getting output data to a bank of DACs is wheel that has been spinning nicely for quite some time.
Convolver on inexpensive PC can do 65k filter on 8ch at faster than real time using 32bit floating point math. Way more than enough resolution for 24bit integer sources.
24bit integer dynamic range exceeds noise floor of a resistor at room temperature.
This is nonsense. Drivers and measurement setup are limiting factors here. Floor of DAC must be measured at its outputs.
Regards,
Andrew
If someone with the knowledge time and drive to complete something like this is hesitant thinking that no one would be interested in buying such a device I beg to differ.
"Chapark" is charging 375usd incl chipping for his device and I belive that's cheap. I would gladly pay two or maybe even three times that for this (for now) imaginary "High end device".
And there is no need to supply dac's,analog outputs, volumecontrol, connectors (other than u.fl) and so on. A decent adc might actually be what's most interessting if something actually should be included (there are not many external to chose from).
Now don't read me wrong here I'm not dissing "Chaparks" work at all and I will be among the first in line to get my hands on one of his devices. And maybe he will prove me wrong and then there is no real need for this project. On the other hand there will always be evolution and it's fun to search for the next "pot of gold".
I would be very surprised if me and "qusp" where the only two ones in the whole world that think that this is an interessting idea worth to spend a few bucks on.
I'm thinking something quite simular to "Loriens" Wave IO in build quality (it's a piece of art actually
).
That was my Saturday night ramblings
"Chapark" is charging 375usd incl chipping for his device and I belive that's cheap. I would gladly pay two or maybe even three times that for this (for now) imaginary "High end device".
And there is no need to supply dac's,analog outputs, volumecontrol, connectors (other than u.fl) and so on. A decent adc might actually be what's most interessting if something actually should be included (there are not many external to chose from).
Now don't read me wrong here I'm not dissing "Chaparks" work at all and I will be among the first in line to get my hands on one of his devices. And maybe he will prove me wrong and then there is no real need for this project. On the other hand there will always be evolution and it's fun to search for the next "pot of gold".
I would be very surprised if me and "qusp" where the only two ones in the whole world that think that this is an interessting idea worth to spend a few bucks on.
I'm thinking something quite simular to "Loriens" Wave IO in build quality (it's a piece of art actually
).That was my Saturday night ramblings
"The idea is to supply ones own digital reciver (usb), favorite
dac's and analog output stages."
The "DSP" is not the challenge. Nor is the software. Whats left, taking Your prerequisites into account, is oscillator, clock and clock distribution.
Is it the clocks your are not satisfied with in Chaparks XO? Or is it to low cost ?
dac's and analog output stages."
The "DSP" is not the challenge. Nor is the software. Whats left, taking Your prerequisites into account, is oscillator, clock and clock distribution.
Is it the clocks your are not satisfied with in Chaparks XO? Or is it to low cost
?
I actually have no idea how "Chaparks" XO is put together other than two
parts, he has been quite secretive with how it's put together (and thats
actually allright by me even if I'm a little bit curious).
So clock and other stuff is still a bit of a mystery.
Yes maybe I'm just looking for the absolute best parts available combined into
an all digital dsp with a build quality of "Loriens" Wave IO
parts, he has been quite secretive with how it's put together (and thats
actually allright by me even if I'm a little bit curious).
So clock and other stuff is still a bit of a mystery.
Yes maybe I'm just looking for the absolute best parts available combined into
an all digital dsp with a build quality of "Loriens" Wave IO
Last edited:
"The idea is to supply ones own digital reciver (usb), favorite
dac's and analog output stages."
The "DSP" is not the challenge. Nor is the software. Whats left, taking Your prerequisites into account, is oscillator, clock and clock distribution.
Is it the clocks your are not satisfied with in Chaparks XO? Or is it to low cost
I don't see too many here building dsp boards and writing dsp software do you ?? DAC boards are a dime a dozen on ebay
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.