Ok, well when more info is released we can see what wants and needs are possible or not. Like I said it couldn't be released as a DIY kit only as a finished product which obviously has cost implications. I'll say up front that if I can do this it won't be able to be like a GB, I'd need to make a profit. Obviously if I knew there were 10 people having them then discounting is possible of course and like a GB the more components I buy the cheaper they would be and that could be passed on.
At the moment I am an OEM so that I can include some of thier kit in custom speaker builds for customers. I have no experience of testing for certification a finished product for market, anyone any idea of the process and costs involved?
At the moment I am an OEM so that I can include some of thier kit in custom speaker builds for customers. I have no experience of testing for certification a finished product for market, anyone any idea of the process and costs involved?
Last edited:
Sorry Stephan, I missed the point that you are an OEM. I am OK buying through you. Assuming it is at least 192 kHz internally and analog balanced in and out with some reasonable software then I would take two initially, and two more after testing the response in a 192 kHz DSD setup.
I don't think certification would be a problem and should be the same for OEM or DIY. Whatever markings are done for the DLCP (CE, UL, etc.) should be fine. I would probably take it out of any packaging as I want to put it with the NCore/UcD/PS to minimize power/balanced/output wiring.
Last edited:
Just to be clear, what is "192k DSD"? Standard DSD has a sample rate of 2.8224 MHz, DSD128 has a sample rate of 5.6448 MHz.
My poor wording. Meant 24b/192k music streamed over DSD. It started when I played some 24b/192kHz music and discovered rather accidentally that it only sounded better over USD>DSD. That is over SPDIF/TOSLINK/RCA it was noticeably thinner and not much different than 16b/44.1kHz comparison tracks.
So I decided on DSD and then started messing around with a room tuning project. As soon as I skipped the 3-way crossover in the speaker with a 2x4 miniDSP the sound went thin again. So I borrowed a miniDSP 4x10 HD but noticed little improvement. So my unscientific conclusion (no measurement just the six ears of family testers) is that the DSP downsampling is negating 24b/192kHz music benefit.
Today the best audio chain I can hear is 24b/192kHz music over DSD > USB to High Quality Low Jitter DAC > Balanced Out > Hypex NCore > modified ZRT with 3-Way Crossover. Anything else in that chain (miniDSP, SPDIF, TOSLINK, RCA) thins it out and goodbye DSP room filtering.
I considered the Hypex DLCP but spec-wise it may have the same problem. Which is how I got to this thread - looking for a DSP that would have high-quality balanced analog in and out (forget all the others) and maintain the source sample rate. The only DSP spec I’ve seen that comes close is this project.
http://www.diyaudio.com/forums/digi...-integrated-preamp-crossover-dac-project.html
Of course it could all be BS but for now I’m sticking with crossovers: Why 24/192 Music Downloads make no sense...
https://www.xiph.org/~xiphmont/demo/neil-young.html
One further note: on my audio chain in the previous post the samples in this link are silent with crossovers but hiss with a miniDSP. Does that mean the miniDSP is better (rendering something the crossovers miss) or worse (adding noise)? For now I’m sticking with crossovers and looking for the best DSP solution possible.
Not sure about that, need to look into it. I'm pretty sure it's the boxed product that would need the CE Mark, although this should be made easier by the fact that the components have their own.
You would of course be free to take it out of the box, however if I do this there could be a solution that includes the NCORE channels and the PSU distribution as Hypex will be making an OEM NCORE specifically to work directly with this new module.
I don't know anything about supported sample rates or DSD yet. Though I'd imagine the processing overheads to stay DSD throughout especislly if FIR filters are used would be prohibitive.
It would be impossible. You cannot even control volume without converting to PCM at some point.
And maybe even try doing the DSD over DoP, so that you have PCM converted to DSD encapsulated in PCM frames, converted to PCM for DSP operations in the DAC, and then converted to multi-bit delta-sigma for the actual conversion?
Assuming there is any benefit in 24/192, converting it to DSD is definitely a lossy operation, and any DSP operations will have to happen in PCM.
So what you are saying is that you prefer your music downconverted to DSD from 24/192? It sounds like your DAC somehow happens to produce a subjectively more pleasing result from DSD (maybe as a result of the noise shaping).
Yeah, good points. The DAC is a Wyred DAC-2-DSD-SE and went back and forth with them bit trying to figure the root cause (cables, settings, etc.) but in the end it came down to all things being equal (cables, settings, etc.) that USD sounded un-quantifiably best, jitter and all that.
I should be clear to all readers: not trying to trash the miniDSP... this was a hack job setup with a few UcD trying to figure out the best way forward for a couple of digital crossover and recording projects and I'm and analog guy that doesn't believe in software.
Last edited:
I think the quality of a particular implementation of the interfaces has a huge impact. So the DLCP may not use a high-compute-power DSP but the inputs and analog components around it are very good. Another comparison with the miniDSP 4x10 HD illustrates that: see the 9 June 2015 entry in the Linkwitz what's new page: What's new at LINKWITZ LAB We were all trying to understand if the DLCP or the UcDs were the major factor in the difference, so I appreciate your input on your experiences with the miniDSP 2x4 and UcDs and miniDSP 4x10 HD compared to your best solution. In the end it is the weakest link that defines the system and it is not always obvious which is the weakest link.
Separately I have also seen hints that DSD is a good format as input to a DAC, probably because it is in a format that is easy to integrate into an analog format, so there are fewer possibilities of sub-par processing in the DAC.
Another observation is that we seem to be enjoying the benefits of the ISIS project since the complete rewrite of the software is probably mostly justified by that project and will retro-actively enhance existing products.
Last edited:
That might, to some extent, have been true back in the days of pure 1-bit delta-sigma DACs (not that there ever were that many). With modern multibit DACs, DSD has to be converted in any case, and DSD is much trickier to process than PCM.
Yes. Bruno Putzeys said in an interview that he created a DSD DAC out of discrete components - I think that was when he was working for Philips.
And I still remember making a PCM DAC out of a parallel port and a bunch of resistors back in the 70's.
It is trivial to do a simplistic 1-bit DSD DAC, if you can live with the issues. There is a bunch of reasons pretty much none of the modern DSD DACs are true 1-bit DACs these days.
