I will fully accept it is something with our method and implementation which exacerbates TI parts issues and our designs. I can say this right now today, IF you hand me an Apple phone or tablet it will quite working right that instant. IF I dare take a ride in a VW vehicle history shows over a 90% chance the car will fail and I will need a ride home.
As far as the 5532 type parts. I still do not get why so many insist on using this bad sounding part when so many better cheaper choices are available. To me, it is like insisting that 1947 Ford sedan is in every way better than my 2010 model. What is up with that?
Dynamic range is not S/N ratio. Please do not make that mistake. Please see this attached article for some explanation.
As far as the 5532 type parts. I still do not get why so many insist on using this bad sounding part when so many better cheaper choices are available. To me, it is like insisting that 1947 Ford sedan is in every way better than my 2010 model. What is up with that?
Dynamic range is not S/N ratio. Please do not make that mistake. Please see this attached article for some explanation.
Attachments
I have watched this thread for quite a while. It is interesting to me as I work with similar DSP systems on some of my loudspeaker design work.
And I haven't posted to much as you guys are usually quite constructive.
This is an exception.
Comments like those quoted above make it very difficult to believe what you are writing.
If it your aim to espouse a philosophy then fine. But please don't pretend that you are preaching truths in these sentences.
They are opinions. Nothing else.
This is a thread about an idea, that has grown into something quite interesting.
Let's discuss The work at hand please.
Amen and all of that! And I am not even a religious person. I totally agree with you. Tranquility Bass has been far more patient that what I would be.
Update
I was playing around with the FIR filter settings on the Sabre DAC and measured both the built in slow and fast FIR filters at the zero crossing point. The raggedness of the waveform display is an artifact of this particular DSO when used to capture a single waveform because it uses some kind of dithering to fatten up the trace.
ES9018 Slow FIR Filter
Source:- SPDIF at 96KHz
Waveform:- 1KHz square wave
ES9018 Fast FIR Filter
Source:- SPDIF at 96KHz
Waveform:- 1KHz square wave
cheers
I was playing around with the FIR filter settings on the Sabre DAC and measured both the built in slow and fast FIR filters at the zero crossing point. The raggedness of the waveform display is an artifact of this particular DSO when used to capture a single waveform because it uses some kind of dithering to fatten up the trace.
ES9018 Slow FIR Filter
Source:- SPDIF at 96KHz
Waveform:- 1KHz square wave
An externally hosted image should be here but it was not working when we last tested it.
ES9018 Fast FIR Filter
Source:- SPDIF at 96KHz
Waveform:- 1KHz square wave
An externally hosted image should be here but it was not working when we last tested it.
cheers
As for my comments on August 27th, those are my direct experiences via observation. Certainly there are many things in electronics and life which resist explanation. If this stuff works for you fine. Am simply saying "It never works for me or anyone I know." This is a fact and cannot be argued.
These are my two key points, the Sigma DSP evaluation kit is $200 and available to anyone who wants to work with DSP in their system. Two analog in and eight analog out will go a long way toward a stereo system. The Acorate software with a good sound card will do likewise. My thought was to introduce those here to these two DSP solutions. Of course, one can purchase a Media Matrix NION if that suits them. I use those professionally as they sound incredible good and last a long time. This in no way excludes the home brew DSP by Tranquilty Bass. Simply trying to make others aware of available choices. Is that really all that evil?
These are my two key points, the Sigma DSP evaluation kit is $200 and available to anyone who wants to work with DSP in their system. Two analog in and eight analog out will go a long way toward a stereo system. The Acorate software with a good sound card will do likewise. My thought was to introduce those here to these two DSP solutions. Of course, one can purchase a Media Matrix NION if that suits them. I use those professionally as they sound incredible good and last a long time. This in no way excludes the home brew DSP by Tranquilty Bass. Simply trying to make others aware of available choices. Is that really all that evil?
This is a really interesting product and thread. The hardware looks really nice and should have very good flexibility to implement most normal filtering needed. I think desimation is underated and don't understand why filters for subwoofers are implemented with over a few hundred Hz samplerate. Thats a lot of wasted DSP/MCU cycles.
As there has been some replies about alternative signal prosessing platforms so I chime in with this:
I recently came across the Audio Weaver® - the Only Cross-Platform Audio Design Platform | DSP Concepts software and have tried it on a core i5-5257U processor (PC) and ASIO in/out to a RME soundcard. It actually works and the posibillities for audioprosessing are almost endless.
The only drawback is that it is 1GB installation and that you have to know what you are doing when designing filters because of all the possibilities
I recommend it for everyone trying out soundprosessing as an alternative to the established softwarepackages. (Plugins, filteringsoftware)
I really liked that you can run the algorithms on the PC and hear the result "instantly", but it is kind of mindblowing that you can also transfer them to a cheap ARM4 MCU.
http://www.dspconcepts.com/sites/default/files/white-papers/STM32Journal_DSPC.pdf
(if there is room enough of course. For bass applications you can just decimate)
I even got help to get the ASIO working:
ASIO line in. How to specify channels for input | DSP Concepts
The comments on noise/unlinearity is because I put a 1 Hz 60dB/oct HP filter with 44.1 kHz samplefrequency. And that can not be done without truncation error on the filtercoeffs. Fair enough
To have a look at some of the prosessing modules have a look here.
http://www.dspconcepts.com/sites/default/files/documentation/Audio Weaver Module User's Guide_0.pdf
It is kind of a SigmaStudio on steroids.
"Speaking" of Analog Devices, the software also supports this low cost DSP evaluation kit:
EVAL-BF706M-EZLITE | data sheet and product info EVAL-BF706M-EZLITE | Analog Devices
Regards Torgeir
https://sites.google.com/site/passivefilter/
As there has been some replies about alternative signal prosessing platforms so I chime in with this:
I recently came across the Audio Weaver® - the Only Cross-Platform Audio Design Platform | DSP Concepts software and have tried it on a core i5-5257U processor (PC) and ASIO in/out to a RME soundcard. It actually works and the posibillities for audioprosessing are almost endless.
The only drawback is that it is 1GB installation and that you have to know what you are doing when designing filters because of all the possibilities
I recommend it for everyone trying out soundprosessing as an alternative to the established softwarepackages. (Plugins, filteringsoftware)
I really liked that you can run the algorithms on the PC and hear the result "instantly", but it is kind of mindblowing that you can also transfer them to a cheap ARM4 MCU.
http://www.dspconcepts.com/sites/default/files/white-papers/STM32Journal_DSPC.pdf
(if there is room enough of course. For bass applications you can just decimate)
I even got help to get the ASIO working:
ASIO line in. How to specify channels for input | DSP Concepts
The comments on noise/unlinearity is because I put a 1 Hz 60dB/oct HP filter with 44.1 kHz samplefrequency. And that can not be done without truncation error on the filtercoeffs. Fair enough
To have a look at some of the prosessing modules have a look here.
http://www.dspconcepts.com/sites/default/files/documentation/Audio Weaver Module User's Guide_0.pdf
It is kind of a SigmaStudio on steroids.
"Speaking" of Analog Devices, the software also supports this low cost DSP evaluation kit:
EVAL-BF706M-EZLITE | data sheet and product info EVAL-BF706M-EZLITE | Analog Devices
Regards Torgeir
https://sites.google.com/site/passivefilter/
Last edited:
Hello torgeirs, you must have read my mind !!
I wasn't going to say anything until I successfully ported Audioweaver across to my board so the secret is out now
And you are right - it is a waste of DSP resources to process the bass channels at 192K. You'd only use 192K for the top-end channel of a four-way system. Bass can be done at 24K or even less. Mid-bass coupler can be done at 24-48K whilst midrange-tweeter can be done at 48K-96K etc. The choice is entirely up to you and is only limited by DSP resources.
If you only process bass at 24K it makes it easier to create FIR's for room correction
I envision people creating audioweaver designs and sharing them around with others,
cheers
I wasn't going to say anything until I successfully ported Audioweaver across to my board so the secret is out now
And you are right - it is a waste of DSP resources to process the bass channels at 192K. You'd only use 192K for the top-end channel of a four-way system. Bass can be done at 24K or even less. Mid-bass coupler can be done at 24-48K whilst midrange-tweeter can be done at 48K-96K etc. The choice is entirely up to you and is only limited by DSP resources.
If you only process bass at 24K it makes it easier to create FIR's for room correction
I envision people creating audioweaver designs and sharing them around with others,
cheers
I have never understood why you can't go lower.
Case: Bass unit filtered at 100 Hz LP.
Steep LP filter at 48 dB/oct or more would be about 150dB down at 300 Hz.
I think that is good enough antialising filter. Nyquist then says 600 Hz samplingfrequency.
Then all the room correction or other time and frequency filtering for that bass unit can be done at 600 Hz samplefrequency.
512 sample FFT then gives almost 1 Hz frequency resolution of the filter.
Maybe I chould try it out with Audioweaver;-)
Regards Torgeir.
What sort of parameters do you need to control because there is a tuning interface on Audioweaver that can be run from a pc but that wouldn't be something you'd want an end user fiddling with.
Apart from that the board offers an LCD interface for the standard controls such as volume, input selection, DAC settings etc.
cheers
Oh I see, so with the LCD interface could you map control on the screen to stuff that us set up in audio weaver?
I'm thinking of a situation where an active speaker designer (me) using either your board or any other method out there can provide to tools to be able to say control volume, some eq, room FIR curves etc via some sort of simple user interface maybe on the PC.
I'm thinking more of commercial designs where you would want to keep the user away from all this backend stuff. Like the Grimm Audio stuff.
Maybe i should start an audio weaver thread
I'm thinking of a situation where an active speaker designer (me) using either your board or any other method out there can provide to tools to be able to say control volume, some eq, room FIR curves etc via some sort of simple user interface maybe on the PC.
I'm thinking more of commercial designs where you would want to keep the user away from all this backend stuff. Like the Grimm Audio stuff.
Maybe i should start an audio weaver thread