transcoding flac on the fly to Wav

Hi, likesmusuc  I can recommend Horrowitz and Hill - The Art of Electronics  if you want to read  about and learn about these engineering aspects and how to mitigate them. That is the Digital  Electronics engineer bible.. Very good indeed. I spent quite a time on this in my dissertatiorain on signal analysis in image processing hardware//software..developing and researching this with some memorable failures where I completely failed to consider this effect initially ;-)

More recently I have looked at this with SDR, where EMI/RFI from switching circuits causes spectra noise... Which was then audible

Finally for the maths of of spectrum distortion on digital systems, I found Dr Smith's superb text on The Scientist and Engineer's guide to Digital Signal Processing a great guide... It takes you into the maths not too quickly ..

The  EMI reverse engineering (measuring CPU instructions in a smartcard from powerline modulation) was activities 12 years ago with a previous employer, but alas I cannot post specifics about that here...

All these aspects are big standard engineering principles which you have to master if you designing low noise floor applications,, if you are interested in electronic engineering I am sure you agree it's fascinating.. Let us know how you get on, I'd be very interested to hear from you.

Simon

Before you convert your library, just try transcoding on the fly using Foobar, Asset etc. - thus all the processing is done away from the streamer. I'd be interested to hear your findings on FLAC vs  WAV vs transcoded  FLAC. I can hear FLAC is very slightly better than WAV but can't tell between transcoded level 5 FLAC and WAV. YMMV

Don't sweat it - the differences are small IME but it is free and only involves a mouse click or 2 to sort it permanently.

Crikey, now I realise where I have been going wrong. I’ve been relying solely on my ears to decide what I like.

Peter

Simon, I wasn't asking for a general education in electronics. You clearly have one. I was asking you whether you have any direct evidence of EMI causing issues with Naim dacs, or whether you are just speculating.

I take it that you are just speculating.

Oh I see I think, yes an EM field is caused by all electronics (actually the flow of an electric current through a wire in the base instance) , therefore it will be so with Naim as they are electronic devices..... Is that your question? Slightly confused.... Or are you saying electronic design principles such as EMI, ohms law, transmission line principles, small signal principles  etc are speculation....

I think if Naim engineers discovered a way of bypassing the standard laws of physics then there would be far more money to be made than by de signing and building hifi components.... 

 .. Not to mention the international awards...

So to your point I have no evidence of EMI not occurring within Naim components including their DACs, as with almost everything else on the planet

Peter, it's the best way.. 

Unless you have insecurities and need to follow the herd...

What I am saying, and I'm sure you understand it, is that you gave the impression in your earlier post that the reason that FLAC sounds worse than WAV on naim equipment is due to EMI fields and this can be measured on a spectrum analyser. Given your technical competence, that comes over as an authoritative view. I am asking you whether you have actually made this measurement on naim equipment and so have evidence for your hypothesis. Or are you just speculating that EMI might be responsible for the perceived issuse with FLAC on naim equipment. Strikes me you are just speculating.  

Ok.. Yes so recently (in the last 24 months)

A) I have not looked at a spectrum analyzer out of a Naim DAC, but I hear differences in decoding different encodings.. Not just with Naim..

B) I have looked at a Spectrum analyzer of a SDR where programming cycle variations caused a re profiling of the spectra. The audio effect was a slight hardening of the audio.

C) I have also seen the output of a spectrum analyzer looking at intermod distortion caused by EMI from Ethernet  over powerline measured using a local field strength detector. This was audible (with Naim equipment amongst others such as Icom radio equipment ) with increased noise floor and less fluid presentation ie hardening)

Is that what you are  asking.. I still dont quite see what you are getting at.. I suggest perhaps for you  the best way is to see it for yourself on test equipment.. if you are lucky enough not to hear it.... But if you can't hear it why are you worried... Perhaps you can see why I am confused where you are coming from??

BTW I write on this forums in a personal and not professional capacity... Everything is a personal view and perspective and yes with me sometimes it refers to professional or educational experiences where I feelMIT may be of benefit to others ... It certainly isn't authoritative however..  I enjoy reasoned intellectual discussion.. but I do struggle trying to justify standard principles.. Its a bit like saying why is the speed of light what it is.. which is why I refer to great texts on the matters in question  which can explain far better than I ever could on this hifi forum...

So Simon it seems to me that

a) you haven't made any measurements, with a spectrum analyser or otherwise, that demonstrate a difference in a naim streamer when decoding FLAC vs WAV.

b) you are simply speculating that EMI is the cause of the difference you hear.

Yes, based on my expierience of EMI and noise from switching circuits in CPUs and the effect of HF and intermod distortion on delicate audio electronics, I conclude what I have seen and heard elsewhere in non Naim equipment, when I experience such audible differences  with Naim equipment unpacking different data packages such as wav and flac that this is caused by the same influences.

Logically it is conceivable the audible differences I hear between WAV and FLAC decoding could be due to a software bug in the Naim streamer firmware, but IMO I consider that unlikely.

I have reconstructed WAV and FLAC files and they, or at lease my files, are they are sample by sample identical.

So yes I do speculate that these audible differences are not predominately caused by a software bug.

But if I do get my mitts on a low frequency spectrum analyzer I will have have a play with my NDX and post here.. Just like I did when dispelling the myths about CD  rips a couple of years ago. But of course that won't determine whethere the differences is due to a software bug however unlikely ......

To keep this interesting I am sure we would be interested in your theories and yourconstructive insight  for the audible differences, as you seem determined to discount my supposition for some reason and you have done this previously without adding anything constructive, almost as if you are trolling..  it seems to be all me doing the talking which seems not quit right and to be quite honest it's getting a little dull...    

Simon

Simon, all I am doing, as any decent scientist or engineer would, is to establish the basis for your claims that the difference you hear between FLAC and WAV is caused by EMI and the susceptibilty of a naim streamer to EMI. It turns out that you have never measured this aspect of a naim streamers performance, so all you present is supposition (your own word) and conjecture. You (and I) quite correctly challenged views that naim rips were better than non-naim rips by producing hard evidence. But you haven't produced any hard evidence that naim streamers are susceptible to the (presumed) increased EMI produced by decoding FLAC. If you were to produce some measurements, maybe the guys at Naim could fix the problem.

Correct, but do you agree with my supposition or can add it to it? If not please put forward  your alternative supposition? I'd be genuinley interested either way..

S

dont get into a pissing match with a skunk... words to live by, you two.

I think we just have some different points of view .... thing I don't understand is the Naim DAC has a buffer and you put PCM data in that buffer and it plays it ... we all seem to agree that the data in the buffer is the same .... I use an optical connection which I believe is immune to RFI and EMI ... so I'm not introducing this to the DAC ...

Why does the DAC care (how does it even know?) - if my original PCM was in an AIFF. ALAC. WAV or FLAC container .. 

I would suggest my room is not a controlled environment and is not at constant temperature for example ... and this could change how I hear something. I don't expect identical replay each and every time, but as long as it always enjoyable then no concern. However, this means that I cannot say differences are down to one variable only. 

So I have to introduce some logic from my understanding of data integrity ... my Supernova 6 puts the same PCM in to the DAC irrespective of lossless format, it shouldn't introduce electrical noise, the jitter is very low and no reason why it should be higher with one format than another ... 

I conclude that it really doesn't matter in my set-up which lossless format I store the data in. 

I have read some very strange things on digital audio ... some manufacturers make extraordinary claims - preserving the narrow bits being the most ridiculous in my view - this is very bad practice in my opinion as somebody is trying to make money out of claptrap. 

The views expressed on the forum have no such hidden agenda, but are arguments put forward by genuine contributors ... it is true processors work harder to get PCM out of ALAC/FLAC than AIFF/WAV (even I can see that quite easily on my Mac) ... I can understand that my DAC may not appreciate my Mac's SMPS ... so I can follow Simon's arguments and can definitely see how a Naim streamer may be susceptible to these type of things (I am not saying anything is wrong with the Naim design, as optimising for one format rather than adopting a lowest common denominator makes perfect sense) ... similar I have great empathy for likes' arguments: similar logic to mine, if I'm not mistaken

I guess like all these things ... folk should try for themselves and make up their own mind ... there is no easy definite right answer ..... what is better KDS or NDS? (What answer do people expect? - lots of posts all in agreement that one is better than the other ... I doubt that will happen). 

At least we all seem to agree that the magic ripper is a myth ... some consensus ... by the way if you burn an Audio CD using Roxio Toast does it sound better if you drop WAV files on to the burn box rather than FLAC 

All the best, Guy 

Guy absolutely, it's best to do what sounds best to you, if you are streaming files, then it's a trivial change to transcode to FLAC or PCM as appropriate. If flac sounds the best or the same as everything else then use that, if like me you prefer WAV, then if neccessary transcode, it's a 10 second change, no sweat no fuss..

Simon

all subjective let it rest

Hello Simon

If I were using Asset and streaming to a NDX, NDS or similar then I would definitely transcode to WAV .. I can see nothing against so doing and it will certainly not degrade performance in any logical way I can think of (I guess you do get more traffic on the network with WAV), So it does seem the best option for Naim streaming ... store in FLAC (for tagging) and play in WAV.

In a way I'm doing something not too dissimilar ... store in ALAC/AIFF (I use AIFF for 24 bit as it won't support 192 KHz and it gives me an easy way to tell the bit depth apart) and convert on the fly to PCM before dispatching it to Naim.  

So perhaps we are all converging after all 

All the best, Guy