It would be awesome if there was some PCM to DSD converter that converted everything in realtime, and that would actually make the kusic almost as good as native DSD. I suspect that it will not sound as good native DSD, or even close, but I have not tested it. Seems like a lot if hassle, but if it was possible then why not.
Regarding DAC quality, if $350 is what I have to spend on a DAC, then that DAC will sound good enough to me. 🙂
I believe there are other devices that were available that recorded natively in DSD from at least Sony. I believe the only currently readily available devices are fro AKM. I have several devices capable of recording natively in DSD, and yes they are all based on AKM ADCs. I have an RME ADI-2 Pro FS R that supports native DSD as do my kaamos tech converters.
https://www.alpinesoft.co.uk/VinylStudio/dsd.aspx
My main system is build around PCM hardware running at 2496 so no DSD here, but I do transcode DSD to PCM for playback, my other two systems are analog with dacs that support DSD. I have native DSD and recordings transcoded from DXD.
I am coming to care more about the end result than the means of achieving it, or at least that is what I tell myself. 🙂
https://www.alpinesoft.co.uk/VinylStudio/dsd.aspx
My main system is build around PCM hardware running at 2496 so no DSD here, but I do transcode DSD to PCM for playback, my other two systems are analog with dacs that support DSD. I have native DSD and recordings transcoded from DXD.
I am coming to care more about the end result than the means of achieving it, or at least that is what I tell myself. 🙂
Roon has it in their 64 bit DSP, I recommend you try it. A cheap Intel Nuc or similar computer is all you need to get started. It is not exactly "real time" as there is some latency which is processor and memory dependent. I think latency is irrelevant practically speaking. There are undoubtedly other options for doing the job, but Roon has a free trial option, if you do not like it cancel within 30 days.
My individual dac budget is about 2x - 3x that and I would still consider those inexpensive.
There is a very good FPGA freeware one in the forum (for diy use only): https://www.diyaudio.com/community/threads/simple-dsd-modulator-for-dsc2.370177/ (I use version 3 firmware)
And there are some accessories to go with it: https://www.diyaudio.com/community/threads/general-purpose-dac-clock-board.413001/ Note there is an interconnect board on page 4. Reclocker schematic is on page 3.
Combined with a TheWellAudio or with a MarcelvdG RTZ FIRDAC, you can go in a direction well past anything from Gustard or Topping. https://www.diyaudio.com/community/threads/return-to-zero-shift-register-firdac.379406/ ...The trick is configuration as shown in the Clock Board thread plus a transformer based output stage. ThorstenL hints at one way to do it.
Last edited:
And yet, an analog distribution system would be immensely more lossy then a digital ;-) (psst, see vinyl)
//
So, Tom wrote something about degrading original 16 bit data and acko pointed out that the original recording probably isn't 16 bit anyway. No idea why this derailed into a completely off-topic discussion about analogue tape recorders and vinyl records.
Btw, Tidal did not MQA everything. Original 16-bit recordings left as they are - so nothing to degrade here. Only the newer ones, recorded from 96k -384k 24-bit gets MQA’ed to CD bandwidth and stamped “MASTER”. When unfolded the ‘original’ level is revealed though some loss but close enough
Its nothing compared to a class A/B power amp biased a little more into class A. The only energy needed for DSD upsampling is to power an FPGA that uses less power than a USB board. That said, there are less efficient ways to do the conversion too, just like there are inefficient fully class A power amps.
Last edited:
I have never measured the power consumption of my dedicated Nuc based Roon core, but DSD conversion is definitively processor intensive - the Nuc flat out consumes about 60W which is far less than my four channels of 300B SE class A mids and tweeter amplification and class D amps (bass and subs)
The tape/vinyl analogy appeared due to some earlier comments Mark was responding to.
The widest dynamic range source material I have in my collection is 2-track 15ips master tape reissues, despite the obvious theoretical advantages 24 bit PCM in particular has in this area. IMO there is a point where dynamic range can become excessive, and that is when you set the volume to a comfortable level for quiet passages, and are then forced to turn the volume down on very loud passages due to auditory discomfort.
The tape/vinyl analogy appeared due to some earlier comments Mark was responding to.
The widest dynamic range source material I have in my collection is 2-track 15ips master tape reissues, despite the obvious theoretical advantages 24 bit PCM in particular has in this area. IMO there is a point where dynamic range can become excessive, and that is when you set the volume to a comfortable level for quiet passages, and are then forced to turn the volume down on very loud passages due to auditory discomfort.
My FPGA board (not the same as Mark's) and ASRC chip together draw about 2 W to convert PCM to a 27 Mbit/s noise-shaped single-bit signal.
In interpreted this post as "all digital is lossy; only analog is lossless". I simply pointed out the flaw in that argument. Analog formats are quite lossy, more lossy than current digital formats due to their much lower dynamic range.
My point was actually NOT that the original recording probably isn't 16 bits anyway. My point was that any analog format, be it reel-to-reel, compact cassette, LP, 7" single, wax cylinder, whatever, will have a lower dynamic range than even a 16-bit CD without any of the fancy dithering or noise shaping. I.e., any analog format will be lossier than even a 16-bit CD. So if you are looking for distribution and reproduction that is as true to the original as possible, i.e., as lossless as possible, the CD is the better format. Noise shaping such as MASH only further improves on this.
Of course, these days there's really no reason not to distribute the 24/192 recordings. Bandwidth is plentiful and storage is practically free. Modern DACs support 24+ bits. So I simply don't understand the point of MQA. It seems it's trying to solve a problem that isn't there all while creating another problem ... and charging users for the experience. Yeah... No thanks.
Tom
Be sure that MQA was 95,34% a commercial endeavour disguised in some juicy technical lingo for the gullible to drool on... music had very little to do with it... thats why it was so devilish....
What the industry needs is something new that is demonstrably compleatly different and superior to what we hear today... anything else will just pass by...
This was for sure not M Q A.....
Listening / comparing MQA gave at hand a very small difference... which probably mostly depended on new mastering for the release rather than some "folding"...
Yes, I deplore it.. 😉
//
What the industry needs is something new that is demonstrably compleatly different and superior to what we hear today... anything else will just pass by...
This was for sure not M Q A.....
Listening / comparing MQA gave at hand a very small difference... which probably mostly depended on new mastering for the release rather than some "folding"...
Yes, I deplore it.. 😉
//
Last edited:
I stand behind my comments earlier that extraordinary claims need to be backed by extraordinary evidence. I further believe that those who make extraordinary claims should allow for third-party testing of those claims. MQA failed on both.
Tom
Tom
Its a nice sounding slogan, but one sometimes used as weapon. Such as for example, say, someone's income depends on a claim being deemed as extraordinary. Only point is, the use of such slogans should be evaluated on a case by case basis for appropriateness/applicability.
I cant see how any dependency level of anything could decrease that statement validity. Cant see that income has any bearing on this at all... Your sentences are not really coherent so har to say how you see the income aspect - i would say the higher gain/income, the more need for evidence....
//
//
- Status
- Not open for further replies.