upsampling to 96k ?

Hi Paul, come come, it's bit like riding a bicycle one learnt you never really forget ;-)

James, I agree with your assertion...  I would certainly suggest this be done by adding zero samples to the side of the original sample in the sampled signal. Ie at f add a zero  sample it then bones 2f, add two zeros it becomes 3f etc.
Zeros are great in DSP  as when you multiply them in signals with a filter signal (convolution) they don't add to the output and so are not producing unwanted data (and errors!!). The reconstruction function/filter turns it back to the correct analogue signal. If the maths doesn't need it, why put it in and screw things up?.
I think this is Naims view as well reading their White papers on the NDAC and ndx.
Simon

quote:

Up sampling or oversampling has been around for years in consumer audio equipment. it is usually performed in the DSP in the DAC chip. It's almost been transparent to you.
...
Custom DSP chips are preferable to software as because they are specially designed to perform discrete Fourier transform and related functions, they are a lot faster for a given precision compared to a regular CPU.  In realtime DSP speed is usually important.

Simon, my latest experience with the NAIM gear at hand - the UnitiCute - does not reflect your suggestion regarding which upsampling to use - at least not to my ear-brain system.
Simply, the software-up-sampled 96k files sound way better than streamed 44.1k files with NAIM's internal SRC hardware up-sampling to 96k - anybody highly invented to compare himself...

Besides that - to upsample files by software and then feed the gear at its native sample rate has two benefits (besides the jitter issue)
1. flexibility - if there becomes better software available - no prob. also to switch between choices is easy - and - there are choices at least.
2. software solutions are not restricted in computational time they consume - if it would not have taken 2 days for me to up-sample all files on the NAS but two weeks - so what? - its only once to be done

I agree that DSP horse power has becoming incredibly cheap compared to the times the DSP first chips were introduced and I'm also almost certain that NAIM could do better here (the upsampling form 44.1 to 96k) as is obvious for me since I discovered their FLAC bug.

If you ever offer your service to NAIM I'm your happy supporter at anytime - they definitely could need some help IMO !
:smile:

Michael

Michael, thanks for your kind words. I must admit I am more into communication systems engineering now and I don't need practice my DSP skills other than know which codecs to use and their bandwidth requirements, and I don't think naim have much call for that, But going back to basics through the course of this thread has brought a lot of it back. It does appear you have found something a bit strange withe Qute however. Please do keep your reports coming in, and see if you can borrow a ndac from your dealer, and see what you think with that? Simon