Hi-Rez Recordings & 24 kHz cutoff
Can anyone who has a Najda comment on the sound quality improvements from hi-rez source material at 44/48, 96, and 192Khz? Does the sound quality improvement expected from a hi-rez recording make it through the 24 kHz cutoff digital input filter to the amps/speakers? In other words, do 96 and 192 khz recordings sound better than a 44 Khz recording of the same musical piece through the Najda?
Can anyone who has a Najda comment on the sound quality improvements from hi-rez source material at 44/48, 96, and 192Khz? Does the sound quality improvement expected from a hi-rez recording make it through the 24 kHz cutoff digital input filter to the amps/speakers? In other words, do 96 and 192 khz recordings sound better than a 44 Khz recording of the same musical piece through the Najda?
For the future i will look to get a full symmetric setup. My Idea is to run two Najda-Board parallel to get the 4 Channels i need for real symmetric work. Is it possible to run one Board as master and a the second as slave? Or is there anybody knowing an alternative for real symmetric work?
Best regards
Dieter
Best regards
Dieter
And what sound quality improvement are you expecting?
I can't answer for Najda specifically, but with any of the other gear, in any of the many listening tests I have done, a reasonably soft lowpass above 22 kHz hasn't had any influence on sound quality except in one special case where a low pass filter actually improved the sound quality - there seemed to be some nasty ultrasonic intermodulation going on.
This may be a little off topic, but I know some of you tried it....
I want to experiment with a Raspberry PI2 as input in I2S mode. How does it compare with a XMOS USB card like a waveio or an amanero card??
I'm now using my Najda with a Squeezebox Duet with the SPDIF connection. Will I see big improvement over this setup???
How easy is the setup with the Najda???
Thank you
Zoula
I want to experiment with a Raspberry PI2 as input in I2S mode. How does it compare with a XMOS USB card like a waveio or an amanero card??
I'm now using my Najda with a Squeezebox Duet with the SPDIF connection. Will I see big improvement over this setup???
How easy is the setup with the Najda???
Thank you
Zoula
Hello Zoula
The Rpi's I2S output is not so good due to master clock issues on the Rpi. (google).
The WaveIO is much better.
I had a Logitech touch and tried a Rpi with a hifiberry digi with isolated spdif out. That was definitive a step up in SQ. I still have those if you want to buy ;-)
After that i chose the waveIO with separate linear power supply and it was super!
It is also quite more expensive than a hifiberry digi. about 8-9x more expensive
I did not mind, but that is up to you...
You can still use the Rpi as streamer in that setup.
I use it with Squeeze on arch, but out of the box there is no Rpi 2 support at this point. If linux is your thing that will be no problem as you can compile yourself (i do not know how...)
The Rpi's I2S output is not so good due to master clock issues on the Rpi. (google).
The WaveIO is much better.
I had a Logitech touch and tried a Rpi with a hifiberry digi with isolated spdif out. That was definitive a step up in SQ. I still have those if you want to buy ;-)
After that i chose the waveIO with separate linear power supply and it was super!
It is also quite more expensive than a hifiberry digi. about 8-9x more expensive
I did not mind, but that is up to you...
You can still use the Rpi as streamer in that setup.
I use it with Squeeze on arch, but out of the box there is no Rpi 2 support at this point. If linux is your thing that will be no problem as you can compile yourself (i do not know how...)
Thank you Shutdown98,
Does the hifiberry as is own clock?? because how different would it be from plunging the Pi directly in the Najda....
Of coarse the WaveIO will sound better, everything on this board is there for maximum audio performance. but like you said it's 3X the price. Not that it's that expensive but I have have 2 Pi in the mail for other project an I was wondering how it sound with the Najda??
Does the hifiberry as is own clock?? because how different would it be from plunging the Pi directly in the Najda....
Of coarse the WaveIO will sound better, everything on this board is there for maximum audio performance. but like you said it's 3X the price. Not that it's that expensive but I have have 2 Pi in the mail for other project an I was wondering how it sound with the Najda??
Hifiberry has its own clock. So has the Najda and i think it reclocks it all and acts as master.
So it could be very good. I think it would outperform your Duet.
But it lacks the jitter reduction the WaveIO has (and more). But if you can experiment and use the pi elsewhere i would certainly try and listen to the outcome.
The pin layout of the I2S output on the rpi you can google. Use also the manual for the Najda. You need 4 short cables (exact same length and as short as possible). Dupont cables are easy.
So it could be very good. I think it would outperform your Duet.
But it lacks the jitter reduction the WaveIO has (and more). But if you can experiment and use the pi elsewhere i would certainly try and listen to the outcome.
The pin layout of the I2S output on the rpi you can google. Use also the manual for the Najda. You need 4 short cables (exact same length and as short as possible). Dupont cables are easy.
Been gone for a while, so not sure how much I have missed on technology advancements.
But I have a simple question I would like to ask, which USB I2S input would work best with the nadja DSP? What options are there currently? I would like to connect my Galaxy S6 / iphone 6 with their now available digital audio out with the best and most reliable solution.
Any help would be highly appreciated! Thank you.
But I have a simple question I would like to ask, which USB I2S input would work best with the nadja DSP? What options are there currently? I would like to connect my Galaxy S6 / iphone 6 with their now available digital audio out with the best and most reliable solution.
Any help would be highly appreciated! Thank you.
I'm using the WaveIO USb tou I2S card for two years now without any problems and I enjoy the sound very much. Lots of peaple are using this card with the Najda DSP.
Luckit | Audiophile grade DIY products
An alternative could be Amanero USB to 2S module.
Amanero Technologies
Hi,
I have the same question.
But I found two more adapters:
I2SoverUSB - I2S over USB Audio
and
Isolated XMOS DSD DXD 384kHz high-quality USB to I2S/DSD PCB with ultralow noise regulator - DIYINHK
Is anybody of us using this two modules with najda?
And what are the differences between this modules?
I have the same question.
But I found two more adapters:
I2SoverUSB - I2S over USB Audio
and
Isolated XMOS DSD DXD 384kHz high-quality USB to I2S/DSD PCB with ultralow noise regulator - DIYINHK
Is anybody of us using this two modules with najda?
And what are the differences between this modules
?WaveIO has started has a diy projet on diyaudio
http://www.diyaudio.com/forums/digital-source/188902-xmos-based-asynchronous-usb-i2s-interface.html
WaveIO seems the only one to have onboard I2S isolator.
Hello!
I'm going to order WIMA capacitors and Its need to be clear how much I need to buy. On your photo I see you changed only two capacitors for each OP amp, but the other guys changed all four. We don't have diagram, but as I understood you changed only input and output capacitors - is it? Do you think, it's not very important to change the other two?
Hi, we changed 4 capacitors per OPAMP based on a circuitry recommended by Cirrus Logic fot that DAC. We don't really know if Nick used the same topology, but it seems logical to suppose that he adhered to the suggestions of Cirrus Logic. You can find the schematic of the output buffer in the data sheet of the DAC. If you only change the output capacitor you only need 2 per OPAMP
Hi! There is a schematic of the output filter/buffer from the manual. As we can see, there are only one capacitor in the signal path. The other is in the filter. If we change 22uF to 6.8uF, we will change the cutoff frequency for this RC-filter from 3.8 Hz to 12,5 Hz and I don't know if does it matter.
Attachments
I'm done with my box, and it works great. Thanks Nick, for a great product. I do think I deserve the slowest build award, since I started using it two weeks ago in my system.
As an experiment, I tried using rePhase to rebuild my Orion crossover. Since rePhase can do minimum as well as linear phase filters, I thought I would do everything minimum phase instead of the crossovers as a first step.
I remember reading that the taps can be freely assigned, and I think if they were, I could get a working FIR-only setup without decimation for the Orion with the available taps. But in my NUC it seems it's 1023 taps per channel.
Perhaps even 2500 taps is really too low for the dipole compensation down to 20Hz - but in rePhase the response looks reasonable. So I thought 2500 for the bass, 1000 for the woofers and the rest for the tweeters (about 500 I suppose).
The woofers are crossed over at 92hz, LR 24db/oct. I suppose with decimation, one can cut this signal down to nearly nothing and get really good results with FIR filtering without using a whole lot of taps.
However, this kind of compensation really needs to be minimum phase, so using FIR for it is kind of a waste. So it seems to me that to make the board really useful (at least for dipole speakers), one should either have a mixed fir/iir mode or decimation to allow shorter filters in FIR-mode only.
So the question: Is there any progress on decimation and/or mixed FIR/IIR mode?
Another question: I have a miniSharc that I bought a while back. I was thinking to perhaps stick it before the Najda and do FIR filters there. It has about 6000 taps available per channel when run as Opendrc-DI. Did anybody try this setup so far?
As an experiment, I tried using rePhase to rebuild my Orion crossover. Since rePhase can do minimum as well as linear phase filters, I thought I would do everything minimum phase instead of the crossovers as a first step.
I remember reading that the taps can be freely assigned, and I think if they were, I could get a working FIR-only setup without decimation for the Orion with the available taps. But in my NUC it seems it's 1023 taps per channel.
Perhaps even 2500 taps is really too low for the dipole compensation down to 20Hz - but in rePhase the response looks reasonable. So I thought 2500 for the bass, 1000 for the woofers and the rest for the tweeters (about 500 I suppose).
The woofers are crossed over at 92hz, LR 24db/oct. I suppose with decimation, one can cut this signal down to nearly nothing and get really good results with FIR filtering without using a whole lot of taps.
However, this kind of compensation really needs to be minimum phase, so using FIR for it is kind of a waste. So it seems to me that to make the board really useful (at least for dipole speakers), one should either have a mixed fir/iir mode or decimation to allow shorter filters in FIR-mode only.
So the question: Is there any progress on decimation and/or mixed FIR/IIR mode?
Another question: I have a miniSharc that I bought a while back. I was thinking to perhaps stick it before the Najda and do FIR filters there. It has about 6000 taps available per channel when run as Opendrc-DI. Did anybody try this setup so far?
I have had thoughts along similar lines recently.
@Nick: is there a possibility to make a capable add-on board dedicated to room correction that can function in-system? This way, we can use basic Najda with IIR filters and the add-on board can do the FIR processing.
Hi Nick,
I have a question regarding the NUC software. I'm now using my Najda with a RaspberryPI 2 with I2S connection. It's a way better than my toslink connection!!! I probably won't use the Pi for long, I will probably use a real USB to I2S converter (WaveIO ; Amanero) But the Pi may stay in the enclosure depending what's the answer to my question.
My question is : Is it possible for you to compile your code in linux so that It will be possible to have the programming interface integrated with the Najda. This would be really handy!!!
My second question is about the I2S output, I need 8 channels, Do I need a special code for the reassign I2S output???
Thank you
I have a question regarding the NUC software. I'm now using my Najda with a RaspberryPI 2 with I2S connection. It's a way better than my toslink connection!!! I probably won't use the Pi for long, I will probably use a real USB to I2S converter (WaveIO ; Amanero) But the Pi may stay in the enclosure depending what's the answer to my question.
My question is : Is it possible for you to compile your code in linux so that It will be possible to have the programming interface integrated with the Najda. This would be really handy!!!
My second question is about the I2S output, I need 8 channels, Do I need a special code for the reassign I2S output???
Thank you