Meridian Unplugged - Powered by MrTechGuy.co.uk Forums Discussion Meridian - Systems WAV music into G68

@Mudge,
Maybe we have here just some confusion ... at least from my part, because you mentioned taking the average of samples. However, you actually say 'weighted average'. In a sense every filter in the time domain is a weighted average of a number of samples. I only never looked at it that way. The question is, where do those weights come from? As I said above, there is a way of transforming a calculation from the frequency domain to the time domain and that will deliver those weights.

I think we both agree that the mean average of sample n and sample n-1 method is computationally cheap. (n added to n-1 then divided by two) No need for FFT and IFT. Lot's of early sample rate conversion chips worked this way IIRC. You can improve on the results by using more samples, taking averages and weighting those averages to produce an intermediate value. This is all less demanding than using a FFT and IFT.

Mudge

I'm with Frank on this one...


Sorry to be pedantic, but earlier you were making quite a bold claim "The G68 rolls off the mid-highs and emphasize the bass." but did not explain it other than to later claim that Meridian were one of "most upsamplers" and upsampling in an inferior way to what would be optimal ie not using fft. Yet you are now saying that you don't know for sure what they are doing, except that they need to use a low pass filter so must be doing it wrong.

How do you know this is all related to upsampling? For anyone who has owned M kit, it is pretty clear that M are applying a 'special sauce' to the signal with or without upsampling. Its clearly not a case of 'bits out = bits in' with many of the signal paths as we know that dither is playing its part with many input signals, as was so clearly demonstrated at HH2.

I can equally guess as much as you that M need to filter for other reasons too yet do not use this to make the claim that "The G68 rolls off the mid-highs and emphasize the bass". I know that a number of us want to know what evidence you have of the G68 doing this in a way that is psychoacoustically (sp?) detrimental to the sound.

Your 'theory' to me seems to me that: M are using DSP's therefore can't be doing it right so have bundled it in the "most upsamplers" category due to the lack of power that you have witnessed from your programming of DSP chips.

Hmmm ... obviously I did understand you correctly the first time. Taking the mean average of only 2 samples is a horrible way of upsampling and should be avoided at all costs ... as I said earlier, only the really bad ones use that kind of method.

I doubt there are many better upsamplers in digital audio than Meridian's. I have no proof, only years of familiarity with Meridian and its abilities.

I don't think anything can be read into this other than the complex "shape" of the preferred filter required the most DSP MIPS. I'm sure that all of the filters that Meridian staff auditioned used a lot of MIPS.

I never made a claim about Meridians upsampling, I explained the result of one user's experience comparing SRC (which anyone can look at the source code for as it's GPL'd or simply study comparisons of it's performance with other upsamplers) with his G68!
Amazing as it may seem, I like the performance of the G68. It is however very DSP limited compared to an offline PC based upsampler. Compared to recent DSP designs it is quite a way off the pace too. The difference is in how good Meridian's DSP code is.

Huh?

That was the quote I was responding to. Does that make more sense? Also, why the negative attitude?

EDIT: 750 MIPS is not a lot of DSP to play with, compared to the latest DSPs from Analog Devices for instance. Hardly surprising since the G68 is a fair few years old and designed not to leach sales from the 861 either.