How to use DSD128 (5.6MHz) and upwards?
Posted by: bicela on 06 September 2015
Dear All,
I'm using 272 with Minim, it can play DSD (DFF and DSF) at 64bit; there is any setting I can use to transcode only the 128bit and upwards to 64bit for allowing higher frequencies to be reproduced by Naim streamers?
Or I must convert (downsampling) off-line (for example with XLD)?
Best, Maurizio
Only DSD64 is supported. DSD128 cannot be played natively.
Hi bicela, I really hope you are enjoying the new toy
+1 feeling_zen, Naim units will only play DSD64
But in your post you mention Its not 64-bit & 128-bit ...........
DSD is all 1-bit, its "bit stream" & is one bit at a time.
CD is 44.1/16 - this means sampled at 44,100 times per second @ 16-bit
DSD uses pulse density modulation (PDM) to encode 2,822,400 1-bit samples per second.
So what is DSD64 ..... DSD64 is 64 times faster than the 44.1kbs CD standard
Also compared to CD, DSD boosts the potential frequency response from 20kHz to 100kHz and the dynamic range from 96dB to 120db.
Ciao ........ Spero yu hanno un sacco di piacere
Thanks dear friends,
Indeed I've made a mistake, of course 64 isn't to 'bit' related! Sorry.
I'm getting fast confused...
I just want to listen to some SACD tracks extracted that seems to be 128.
Someone is aware of valid undersampling programs for OSX? I will so save these tracks to my NAS.
I enjoy VERY much this 272, I will comments by email to you, Mike!
Mike - yes DSD64 is 64 x faster at one bit than CD at 16 bit.. so not really like with like ... I would question the frequency response - sure DSD64 before its filtered will have data upto 100kHz and beyond - but its not really that usable in the real world and becomes increasingly full of noise and digital artefacts as a consequence of noise shaping moving the noise to out of the audio band - hence the preference to often start gently filtering DSD64 at around above 21kHz and also the 120dB dynamic range is between 20Hz and 20kHz and falls away above this frequency...
All said Simon was that it has the potential to go out to 100kHz - I was just keeping it simple - as IMO most prefer it that way.
However I agree your points on the complications of filtering, noise shaping & its effects on the theoretical 120db dynamic range.
Mike, fair do's - however to keep it simple is it not probably best to say DSD64 has an optimum frequency response of between 20Hz and 20kHz and that way one can avoid being technical, avoid mis understandings and over stating its capability etc
Problem with saying DSD is optimised between 20-20kHz is it begs the question "so what is the point of DSD" & how does it sit with the formats competitor, high rate 24-bit PCM ???
I agree it does have all its dynamics in the human audible range between 20-20kHx & it's in that range it can deliver 120dB, however it does have an extended frequency response out to 100kHz & is much the same with high sample rate PCM.
That leads to in theory high sampling 24-bit PCM is technically superior - I believe the equality point is 176/24 - if - that is - your idea of audio is crunching numbers.
And having played a lot of DSD over the last months - IMO the jury is very undecided. To my ears full orchestra is better in PCM, but simple voice & acoustic is better in DSD.
Hi Mike - the equality point is interesting.. given the data rate of 2.8224 Mbits/s of DSD64 this would require the nearest equivalent date rate rounded up of using a PCM framing format of 176.4 kHz / 16 bit.
So DoP takes this 176.4/16 data rate and uses an additional 8 bits as a DSD identifier so as to differentiate it from PCM turning the data framing rate to 176.4/24. Its hugely wasteful but since PCM sample length sizes are either 8,16,24,32 one has to use one of these sample size bit lengths if using a PCM structure, and therefore DSD64 uses DoP 176.4/24 where 8 bits of the 24 bits are used as an identifier.
If however one converted the DSD to PCM, although this is a lossy inexact process, it appears that many consider 96/24 PCM as pretty equivalent..
However I suspect to really appreciate DSD it will be best to use a Delta Sigma compatible DAC as opposed to a Multibit only DAC
Simon
Hello Simon, apart these digression that ignore my question...
Could you please elaborate more about this, it is interesting me, thanks.
However I suspect to really appreciate DSD it will be best to use a Delta Sigma compatible DAC as opposed to a Multibit only DAC
............ I suspect to really appreciate DSD it will be best to use a Delta Sigma compatible DAC as opposed to a Multibit only DAC
............ thats on my future plan list & I'm looking at ESS Sabre32 DAC options, but with various overseas trips & house mods that'll take up to 3 months, it'll probably be mid 2016 before I get into auditions
Hi Bicela, sorry for the diversion....
Naim predominately use the Texas Instruments PCM1704K DAC chip in most of their digital audio products. It is a high precision 24bit multibit DAC chip. Naim select tighter tolerance samples for their higher spec products.
For the latest digital products Naim uses their own digital signal processing software in a SHARC processor ahead of the DAC chip. It is this processor that converts the DSD to multibit PCM suitable for playout by the PCM1704K chips. This process is essentially lossy and is implementation dependent. However despite these limitations clearly we can assume Naim have put a huge amount of care and consideration into this conversion process.
Now there are other DAC chips that are only one bit delta sigma designs, so anything not DSD like PCM needs to be converted by the preceding DSP. So here DSD will be optimum, PCM will be less so.
Finally many modern DAC chips can work natively in DSD or PCM.. so no lossy conversion is required in the preceding DSP. ESS Sabre (as Mike suggests), CS4398 and the PCM4222 devices for example work this way.
I assume Naim have yet to develop a design with these newer devices that matches their PCM performance with the currently used highly regarded PCM1704K device. As the vast majority of digital media is PCM, continued use of the 1704K is understandable.
Simon
One of the best DACs I have heard was in the EMM Labs SACD player. This is a true 1-bit affair, which uses a very simple custom design. From what I have read the ESS Sabre is not a true 1-bit design. There are papers on the Internet from MSB, EMM Labs and PS Audio explain why true 1-bit is better than the way ESS does it. These are an interesting read, but I do not know enough about the subject to filter out marketting input.
Thanks for the tips, I may not be ready to audition yet, or even positively go outboard DAC, but hopefully can get this researched ahead of time.
My concern is DAC looks to be going thru development stages at the moment after a few years of not doing much & that it might well be that this years fav-rave is past sell by date next year.
Hi, I have a number of DSD128 (double DSD) and also DSD256 (quad DSD) - to be able to play them with the current firmware, which only plays DSD64 (standard DSD), they need to be 'down converted'.
The other tool I found is within the Pyramix product from Merging Technologies, which a Pro-Studio toolset, and I had to ask them for an evaluation license.
Thanks,
Simon.
Thank you all for the answers, very appreciated, ciao Maurizio