Hi, I am new here. I have a question for the people of this site. I see a new converter in the merging site that can convert at 352 khz PCM - the SACD "substitute": less noise and better quality.

They(merging/sony)show in the comparative graph of impulses of response that at 352 khz is almost possible to reproduce, with quality, the impulse of response of the analog sound, more or less 80 %; the great question is perhaps that in 400/440 khz will be possible reproduce 100 % of the structure of the analog sound.

it seems that this goes to "solve" the quality of the digital sound, effects and plug-ins...

and this seems to be a plan(sony) to hide (with SACD format) that is possible to use the converters in high resolutions, in PCM format , to obtain the "analog" sound, preventing more piracy(as it happens in format PCM)

One more think, seems that its realy not possible hear above 20 or 30 khz, but at high samples rates
is possible(?) to capture the "minor particules" or "components" of the analog sound.
Why nobody speaks/argue about this?

thanks, sorry my english

rogerio matos

How are you gentlemen!!! All your base are belong to us.

interesting stuff.
oh, and people do argue about this stuff... believe me!!
as for the human hearing range vs. sample rate, I think those things are different issues; what I mean is that with hearing, it is a measurement of pitch, ie: the higher the pitch, the more frequent the peaks in the waveform one hears. at twenty thousand it is just too high of a frequency for average ears - eardrums cant vibrate that fast or something...? anyway, it is not a magic cutoff point, there are many that can't hear above 13 - 15 khz, and there is evidence that some can hear a surprising amount above 20 khz.
As it relates to sample rate, that is different. It is not number of peaks in a waveform, but rather number of samples per second. You can have a sound at 1 khz, well within human hearing that is at a 352 (?) khz sample rate.

As for replicating the analog sound, I think that at 192 (0r whatever that number is ) it will fool most humans, but there is no way you will get it past the family dog!

Rogerio seems to be talking about this web page: http://www.merging.com/2002/html/pyradsd.htm

Unfortunately, I'm just not sure what Rogerio's exact question is.

It's interesting to note that the impulse responses plotted on that chart do not include an impulse response for the 352.8 kHz format that Rogerio seems to be interested in. It makes me wonder how good the actual impulse response truly is for the 32-bit floating point, 352.8 kHz format? Assuming that the 352.8 kHz impulse response is "80%" as good as "analog", (however you would measure that) I wouldn't expect that the relatively small increase to 400 or 440 kHz would take you to "100% analog" quality levels. I think you will find that 100% is a bit like infinity--you'll never reach it.

(Come to think of it, never reaching something involves the concept of "infinite time", so perhaps you have already reached infinity in a sense.)

As for using sample rates with Nyquist frequencies several times higher than the traditionally accepted upper frequency limit of human hearing, I definitely think there is some merit in that idea. How high is high enough? I'm not sure what the answer is to that question, but when we find the answer, I think a few "Church-of-Nyquist" worshipers are going to be surprised.


take care,
McLilith
(Henceforth, to be known also as The AntiQuist.)

spoonboiler wrote:

As for replicating the analog sound, I think that at 192 (0r whatever that number is ) it will fool most humans, but there is no way you will get it past the family dog!

I agree. You wouldn't likely see a dog attempt to chase the PCM playback of a cat's meow--at least, not until you supplement your surround-sound system system with a surround-scent system.


later,
McLilith (I'm the AntiQuist.)

Hi again!good to see fast reply.

Hi McLilith, the impulse response of the new
Sphynx converter is here

[url]http://www.merging.com/download/Sphynx_2_flyer.pdf

Sphynx can converter at 352 khz PCM(new DXD format) and SACD-2 (1 bit/5.644.800 Mhz)
You can see, in grafic, that at 352 khz the impulse is almost equal that in analog impulse signal (but not completely )[/url]

Hi McLilith

it is not a small increase! 440 khz is equal to the 10 X 44.1 khz !

based in the grafic, it is clearly that the difference betwin the impulse of response of the analog signal versus 352 khz it is minor of that twenty percent. and it(352 khz impulse) tends to increasing in accordance with the increase of the samplng rate.

and also is clearly that each time that the sampling rate fold, the impulse of response practically size fold, approaching if more to the format of the analog impulse of response.

see the grafic here

[/img]http://www.merging.com/download/Sphynx_2_flyer.pdf

Those are interesting charts, but I'm not sure what conclusions I should make from them. It's perfectly natural that subjecting systems with widely different sample rates to the exact same "Dirac" pulse (a 6uS pulse, to be exact) would result in widely different impulse responses. Naturally, the system with the higher sample rate is going to approximate the narrow pulse the best, but I wonder how this actually relates to the reproduction of typical music? I also find it very interesting that the DSD system had the worst noise spectrum, with what appears to be 55 dB more noise at 40 kHz than any of the other plotted sample rates, including the "mere" 176.4 kHz sample rate. It would seem that DSD generates a lot more high-frequency artifacts during A/D conversion, compared to the alternatives. This might actually give DXD the advantage over DSD.

The most important test would involve actually listening to this equipment, and making direct listening comparisons. I wish I were in a position to do that.


later,
McLilith (The AntiQuist.)

joke ON:
Still nobody proved analog sound does not alias in our brain's signal path. Even tape aliases badly. So who knows what sounds better?

Sphynx is 24 bit resolution,, only !

maxlef wrote:

Sphynx is 24 bit resolution,, only !

Do you really need anything more?


later,
McLilith

McLilith wrote:

maxlef wrote: Sphynx is 24 bit resolution,, only ! Do you really need anything more? later, McLilith

i do

the graph is marketing-bullshit, check the explanation:
http://www.gearslutz.com/board/showthread.php3?s=&threadid=16684&perpa ge=15&highlight=pyramix&pagenumber=4

they argued a lot about it, many intereseting things to read

I thought high bit/sample rates were snobs matters ...

Some " strict " A/B testing ( listening to a preset then listening the recorded file * just * after, put a crown on 32 bit ) 32 bit integer, as I now record in my Creamware setup : no difference ! in 24 bit : not quite what I listened to, some seconds ago ( on very sharp sounds )

Now on high sample rates : I've run hundreds of sonograms and A/B testing on 44.1 / 96 and it's way more noticeable than bitrates ...

Aliasing, beyond obvious effects described as aliasing is a * lot * of dumb useless background noise ! though not only but ... mainly ...

A disturbing fact : I see near nothing above 20 Khz in some sonograms ( heavily low passed sounds normally ) however I hear a difference in 44.1 / 96 transitions.

Now, I'am really eager to hear many presets at 192 Khz though 96 Khz is very very fine. I'am in to hear some 400 Khz samplings but not eager to pay the bill of the gear behind !!

Oh, it won't help at all in making good music !

Well, I and an awful lot of the rest of the civlised world still listen to my music on CD. So....400KHz helps me how, precisely?

I take it the manufacturing companies are still on the look-out for the next platform for us all to spend money on (and the music-production companies of course would love us all to replace our entire CD collection too)

I wonder how they're going to market this one?
"It's the things you can't hear that make the difference"?

or..."wanna be a dog?"

or..."oi! You deaf bastards! Buy this new toy with extra flashing lights....the Jones don't have one yet!"
(I think the last one would be more effective.)

Nuisances Sonores wrote:

I thought high bit/sample rates were snobs matters ... Some " strict " A/B testing ( listening to a preset then listening the recorded file * just * after, put a crown on 32 bit ) 32 bit integer, as I now record in my Creamware setup : no difference ! in 24 bit : not quite what I listened to, some seconds ago ( on very sharp sounds )

Most of actual 32bit is no more than 24 bit with a huge head room to avoid distorsion.

An audio converter that can't clip, is prety good.

Nah, analog will still sound warmer.

No, I've listened to a good number of synths at 44.1 and 192 khz and while they may sound nicer, they certainly don't sound any more analog. It's not the sample rate, it's an inherent problem in the way we re-create the analog sounds.

We've all listened to many real hardware analog synths recorded at 44.1 on cds and they sound exactly like what they are; analog synths. If you record that analog synth at 192 khz it will sound nicer, but not more analog. Same goes for digital recreations.

In other words, a real analog synth can sound analog when recorded at 44.1, so a digital re-creation should be able to do the same. I think we've been looking for the solution in the wrong places, and no matter how high the sample rate, bit rate or any rate, these instruments will not sound analog.

I firmly believe it's possible to make a digital synth sound analog, but the whole sample rate issue is the right key for the wrong keyhole.

Maybe, but as you say, higher sample rates do tend to sound better. So, don't discourage anyone from wanting to use them or support them. I think that high sample rates are one of the ingredients we need to bake a really good analog cake, but there are obviously other ingredients as well. If you leave any one of them out, the cake won't meet expectations.



McLilith

I think you guys listen to too much Royskopp Melody AM's Eple track.

That's not a intro, it's a engineers hard on!
(The one guy just rambles on about 'good frequencies' over the phone till he says 400. The other dude says '400??? I'm coming over!' )

I would never dream of arguing against higher sample rates. My guitar and bass love being recorded in 96 khz and higher.

The problem is when people say that VSTi's don't sound analog because their sample rate is too low. It may improve clarity and headroom, but it's not going to make it feel any more analog.

I'm still convinced that there is some flaw in our assumption of the behavior of an analog oscillator. Think of it as being like a $10 million mansion built on mud. It may be really nice, but the ground it's built on has some conceptual flaws.

The sooner we accept it and re-examine our methods the sooner we can all have realistic analog emulations from VSTi's.

if anyone in any specific field knows of inexpensive transducers in this range...

I tried high bit rates / samples rates long ago and was quite deterred by the ressources taxing for not that much in return ...
Now I'am becoming more sound sensitive and I'am ready to accept the taxes : enjoy twice, eat a half .. no gain ... great !

BTW its really a per preset thing not a per synth thing so see if it's worth it ... but even 48 Khz is noticeable in some cases.

Now on topic : will it make it more " analog " ? I never owned a real analog but climbing up the ladder will make it sharper and cleaner for the digitals you have.

When I record at 32 bit ( or 24 bit boosted if you like ) I have the exact sounds, at 16 bits it's the girlfriend 20 years later.

For what I tested on all around synths : 96 Khz is just great, 192 Khz should be awesome, above ... I'am curious to hear in some cases, on a good day, on premium monitors !