I read Acko's statement as 'if the renderer receives, and can accept, DoP data, by the time the data gets to the DAC it has been unpacked from the DoP packets and the DAC sees 'native' DSD on it's input'?
Ray
Correct!
This morning I've been testing again and have been enjoying DSD64 and DSD128 music files via BBB/Botic. It seems to work consistently and sound quality is excellent. I've been checking the DSD led status and the Asset debug logs to confirm all is well.
I'm still experiencing the occasional truncation of tracks with playback skipping to the next track but, as I've commented previously, I believe that is due to network limitations (wireless network connection and Asset media library is on a laptop that is being used as a laptop).
I did experience one hiccough when things locked up when I transitioned from DSD playback to a FLAC file on my normal NAS hosted library
...
Ray
Ray, please check and compare with locally served DSD music .i.e copy a DSD file to BBB uSDCard or eMMC
Thanks Caaad, we have discussed this before and good to see you have results. Yes, no DAC required as such for extracting audio from DSD
This looks like an interesting thread.
http://www.diyaudio.com/forums/digital-line-level/273474-best-dac-no-dac.html
Ray
Ray, please check and compare with locally served DSD music .i.e copy a DSD file to BBB uSDCard or eMMC
was planning to do that sometime to compare SQ results but I seem to have a dead SD card this morning; I've got a spare one around here somewhere.
Ray
This looks like an interesting thread.
http://www.diyaudio.com/forums/digital-line-level/273474-best-dac-no-dac.html
Ray
And I posted this sometime ago:
DSD "DAC"
My control point, BubbleUPnP does indicate what is being played on it's 'Now Playing' view (for example FLAC 24bit 96KHz) but it doesn't display correctly or consistently for DSD, sometimes for DSD64 it will show something like 2800kbps/384KHz and for DSD128 5600kbps/768KHz. I plan to raise this as an issue with the Bubble developers.
A small correction on the Bubble display info (now I'm looking at it instead of recalling). Playing a DSD128 file and Bubble is showing 5644kbps/705.6KHz and for DSD64 file 2822kbps/352.8KHz.
Ray
And I posted this sometime ago:
DSD "DAC"
Acko, did you see the post I made on at the end of the tread I linked to? I'm pretty sure Allen Wright was using this approach on his SACD mods 12-15yrs ago. Have a look, I would appreciate your thoughts.
http://www.diyaudio.com/forums/digital-line-level/273474-best-dac-no-dac-5.html#post4313909
Ray
And I posted this sometime ago:
DSD "DAC"
Acko, are you taking this concept anywhere?
Acko, did you see the post I made on at the end of the tread I linked to? I'm pretty sure Allen Wright was using this approach on his SACD mods 12-15yrs ago. Have a look, I would appreciate your thoughts.
http://www.diyaudio.com/forums/digital-line-level/273474-best-dac-no-dac-5.html#post4313909
Ray
Acko, are you taking this concept anywhere?
I will eventually. Whilst the concept is simple, the filtering side needs careful attention. DSD produces a lot of ultrasonic noise (starting from as low as 20KHz) from the noise shaping process and this gets pushed out to higher frequencies. So that is why there is interest to go beyond DSD128 as it makes filtering easier and better.
A simple LPF, whilst good enough to get excited from the sound coming out cannot fully attenuate the ultrasonics. Amplifiers and tweeters respond to ultrasonics if you look at the BW and though we cannot hear this it can be very wasteful energy and damaging to the ears. And also pity those poor pets!
As you can see some guys are using Lundahl trafos as filters but afaik they are wideband and respond up to 100KHz, so will not effectively filter the low freg ultrasonics.
So some work required on the filtering side and also looks like we have to consider a DSD only DAC in its own separate box for the time being. Chip manufacturers of course try to shoehorn DSD/PCM DAC for convenience but I would think DSD needs its own path for the best performance.
my 2 cents but any thoughts, ideas or collaboration welcome
I will eventually. Whilst the concept is simple, the filtering side needs careful attention. DSD produces a lot of ultrasonic noise (starting from as low as 20KHz) from the noise shaping process and this gets pushed out to higher frequencies. So that is why there is interest to go beyond DSD128 as it makes filtering easier and better.
A simple LPF, whilst good enough to get excited from the sound coming out cannot fully attenuate the ultrasonics. Amplifiers and tweeters respond to ultrasonics if you look at the BW and though we cannot hear this it can be very wasteful energy and damaging to the ears. And also pity those poor pets!
As you can see some guys are using Lundahl trafos as filters but afaik they are wideband and respond up to 100KHz, so will not effectively filter the low freg ultrasonics.
So some work required on the filtering side and also looks like we have to consider a DSD only DAC in its own separate box for the time being. Chip manufacturers of course try to shoehorn DSD/PCM DAC for convenience but I would think DSD needs its own path for the best performance.
my 2 cents but any thoughts, ideas or collaboration welcome
Hi,
For a simple DSD DAC. Try checking signalyst webpage --> Products-->Hardware.
Br
Caad
Hi,
For a simple DSD DAC. Try checking signalyst webpage --> Products-->Hardware.
Br
Caad
Yes, I have seen that. Hardware claims to be open but for it to operate requires the upsampling and sigma-delta modulation process running on his PC software - not so simple and similar to what is happening within the 9018 DAC in DSD mode. I would prefer something similar to what Lampizator has done - nicely tuned filter, without any further signal processing!
After all the Scarlet Book indicates this way also ...
Last edited:
..ideas or collaboration welcome
Sound like you've given it a fair bit of thought & you've hit the nail on the head, HF noise seems to be the main challenge. I believe it starts even lower as 20KHz, for what I could hear from Doede's prototype.
Sound like you've given it a fair bit of thought & you've hit the nail on the head, HF noise seems to be the main challenge. I believe it starts even lower as 20KHz, for what I could hear from Doede's prototype.
Yes, I have been following your developments also. Great work guys! keep it up. I am just piecing together some theories only at this stage. But during some of my DIY lash up and testing of raw DSD outputs, I noticed my pet dog becoming restless, although to my ears is still fine music. Someone then cautioned me and I think we got exposed to some ultrasonic radiation
So please take care!
Last edited:
Hi,
For a simple DSD DAC. Try checking signalyst webpage --> Products-->Hardware.
Br
Caad
Very intresting. Do all the clever stuff realtime in software PCM to DSD convertsion & even accoustic correction.
Fascinating! Our brains process out so much noise that presents itself as fatigue. Sounds like your dog was reacting to stuff out of our audible range, but who knows what damage it does.
And nowadays we are bathed in RF energies from home wi-fi also, round the clock!
Hi,
For a simple DSD DAC. Try checking signalyst webpage --> Products-->Hardware.
Br
Caad
It's not simple because the upsampling process is done by a high performance PC, Intel i5 or i7, using Signalyst HQPlayer.
...... similar to what is happening within the 9018 DAC in DSD mode.
The process performance of Intel i5 or i7 is higher 9018 so that singalyst DAC generates a lower level of noise. However, the sound of singalyst DAC is a weak bass.
Last edited:
Interesting! It goes to show that pure grunt itself is not enough. Besides, upsampling/converting 44.1k to DSD and hoping this will produce Super HiRes music just will not cut , at best only colours the sound! Best if you are going for DSD is to save/get money to buy some professionally recorded material ...
Last edited:
...the sound of singalyst DAC is a weak bass.
I've been reading about 'non-DAC' DSD decoding and that observation has come up a few times but not just with the Signalyst solution; perhaps it's something that isn't quite cracked yet?
Interesting! It goes to show that pure grunt itself is not enough. Besides, upsampling/converting 44.1k to DSD and hoping this will produce Super HiRes music just will not cut , at best only colours the sound! Best if you are going for DSD is to save/get money to buy some professionally recorded material ...
I think it's too early to arrive at that conclusion. Based on my limited experimentation I have preferred the sound of FLAC rips converted to DSF to the original rips.
Ray
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
- Home
- Group Buys
- Amanero Isolator/Reclocker GB