Some commonsense

[PeterSt]

Ajax, Not sure if this is going to be useful ...

(I did not read the "report" of John Siau to remain objective)

 

I suppose I won't comply because of having the 16/44.1 act together (if all is well, that is). So say that all is explicitly optimized for that and I am not  bothered by HiRes at all. Say it is not necessary ...

With this in mind:

 

HiRes always sounds totally different, and usually not even for the better. The real explanation for that I don't have, but it could well be because I just don't focus on it. All I do know (and this could be important for context) is that ever back I started the DAC project (NOS1) to explicitly create a comparable apples to apples situation. This means, for example, that the DAC itself always runs at the same speed (say always at 32/705.6 or 32/768 depending on the base material). This is crucial for myself because a difference there would imply different electrical behavior and this would make things apples and oranges.

On a side note, and 10 years of more experience later, things still can't compare for 100% real, because there's a lot more file I/O involved for HiRes, which even is so when all runs from memory (we'd still call that file I/O because inherently (C-code) it is).

 

After this somewhat more technical babble, thus Yes, there's a difference discernible and it is totally easy to hear. It is just the character of the sound. Btw, this just as easily shows faked HiRes, because that does not show that character.

 

I suppose (but did not try to interpret for 100%) that this could be opposite to what you expect for responses. Thus in my case

a. I don't comply because of a too "good" system and should not respond in the first place;

b. but I find the 16/44.1 generally better sounding (with no good explanation).

 

Ad b.:

The majority of people will tell you the other way around (HiRes sounds better) but about as many people will work with fake(d) material without knowing it. But then of course the faked material *is* different and who knows may be better sounding than original 16/44.1. This is not strange, because I too (who not, these days) use upsampling (to 705.6 or 768) for the good reasons we can think of. Now careful because what I use for that is a. not very common (Arc Prediction in XXHighEnd) and b. all of my customers (with maybe one exception that I know of) do not like Hires for the really better as well. Or better put: they don't care a hoot like me. 16/44.1 is good enough, and the chance to run into real HiRes is too small (and the repertoire of Aix may not suite you).

 

Peter

 

 

 

[PeterSt]

12 hours ago, rickca said:

How would you describe the character of real HiRes sound compared to redbook?

 

I just upvoted Frank's and George's post because I feel the former leads to the latter.

 

53 minutes ago, gmgraves said:

The Redbook is exciting, rhythmic, and jaw dropping-ly gorgeous sound. The SACD is bland and ordinary.

 

So exactly that but with the reason of additional noise being the suspect.

 

HiRes always seems to be too smooth and in the end (say before halfway through the album) is too boring because of too much of a same sound. Beck's Sea Change could be a good example of that. Nice silk sound but in the end not enough sparkle. And the more gray-ish (but very soft-gray) is an explicit flavor on top of that. Not explicitly gray, but something in our brain which makes us perceive it like that (a result of too silky).

 

For me it is also easily audible that with Hires cymbals are rendered better and maybe even drums in general are (they are quite easy to focus on), but the dynamics lack compared to Redbook.

 

I also realize that it is super dangerous to claim something like the latter because with the (figurative) snip of a finger it could be distortion of Redbook (illegally rendered too high frequency) what one perceives regarding more snap or tinkle or even bell(s). But careful day-in day-out observing does not make me conclude that. It would imply an all-over-sauce again, and this is not so at all.

 

Quote

but the dynamics lack compared to Redbook.

 

The danger is that looking at the (sample) data, Redbook *is* more steep on the micro transients (sample-to-sample) but proper filtering/upsampling should solve that. Thus, because of the fewer samples with Redbook to go from A to B while the time to go from A to B does not change compared with HiRes, the bends in the road are more sharp and squealing tyres are the result. Upsampling makes shortcuts to the bends and all is fine again. With some poorer tyres the sound of the touching of the tyre to the road - a high frequency happening - is also recorded with Hires. During playback we can hear that too. With Redbook, a lower frequency recording means implying better sealed windows cutting off noise from the outside; this noise is now not recorded and thus it is also no continuous same background sound when the car radio is playing and we like to listen to the music of that alone.

[PeterSt]

4 hours ago, Ajax said:

One night about 7 years ago my then 12 years old son (now an accomplished musician) came into say goodnight and said "that's spooky Dad, it's like she is in the room with us".

 

One night, about 7 years ago my then 12 years old son (now an almost EE graduate) experienced a farewell party of his class at primary school and the next morning of that (they all camped in a couple of tents in the field over here) one of his class mates came in saying thank you. we did not know her really, and saw her suspiciously look behind the somewhat larger speakers. "What are you looking for ?". Her answer: the saxophone player.

So she appeared to study for playing the saxophone and she was sure there was a sax player in the room somewhere. But it was just music playing through loudspeakers ... (and 16/44.1 :-).

 

4 hours ago, Ajax said:

extremely higher frequency rates aren't required, especially for old buggers like me,

 

That is totally unrelated. The fact that you may not hear up to 20K any more, only implies that some inherently square sounds, change timbre because the square becomes sine. So what should happen is that e.g. a bunch of keys sound different to me my son than to you. But you wouldn't know and you also would not notice the difference occurring over time. One thing though:

 

In my theory it would be so that an instrument like the violin may become more easy to play through loudspeakers because the square sound (of resin'ed bow) will have become more smooth to you and the system does not need to render that squareness. Otherwise all we hear for real square sound is at such low frequency that even when you start to wear hearing aids you will still perceive that.

Notice that the square sound I talk about will be synth originated so it requires old fools like me to play that (kind of music) to begin with. Or Play AC/DC what you may be used to, like us with that guy living maybe 40 kilometers from here. But ZZ-Top is also fine for it. Or Rammstein. Or anything with a nice distortion guitar which, mind you, sounds beautiful when rendered as should (go figure about that because that won't go without amplification, so comparing with on-stage and the on-stage will lose ...). The distortion guitar is full with air and most will not have heard that (it requires "ultimate" speed of the whole playback system).

Now where were we ... Ah right, HiRes may not be needed.

[PeterSt]

3 hours ago, mansr said:

Can we please stop this nonsense of calling PCM lossy?

 

👍

[PeterSt]

4 hours ago, esldude said:

But for the most part Hires files don't use more relaxed filtering.

 

At least the DXD from 2NL does not make use of filtering at all (unless they changed their mind by now).

[PeterSt]

4 hours ago, mansr said:

6 hours ago, PeterSt said:

At least the DXD from 2NL does not make use of filtering at all (unless they changed their mind by now).

At the very least, the ADC will use a filter in its internal resampling from the raw sigma-delta output to DXD. If you were referring to filters in the analogue front-end of the ADC, there probably isn't one.

 

By now I am quite sure that I confused myself with their proposed playback means. Thus the DAC. Back at the time they used the first 384 capable DAC (there was an other brand and I suppose I was the third with one) and they were proud of that sampling rate not needing a digital filter (nor analogue). It was a DAC which was not officially on the market (yet). Btw, this must have been 13 years or so ago and I was talking about it with Morten Lindberg over the phone, because I thought it would be a good idea to have their 352.8 supported in a software player.

Those were the days.

[PeterSt]

9 minutes ago, PeterSt said:

Those were the days.

 

Not important whatsoever (from Mr Lindberg from 2L, April 2008):

 

Dear Peter, for your questions I would like to connect you with the very two persons who defined the DXD-format: Peter Scheelke (Digital Audio Denmark) and Claude Cellier (Merging Technologies).

 

They used DAD. Somehow I had Merging in my mind but not strict. Now I read this back I am surprised that indeed Merging was involved too.

 

Now on a somewhat other note:

It was this Peter Scheelke who told me about no filter being necesssary for DXD (= 24/352.8). Back then I took this for granted, but also kind of was lead by this when the Phasure NOS1 DAC was created, which was going on at that time already (first going to market end of 2010). Now, knowing that the NOS1 inherently is filterless (the PC software is supposed to do the filtering), it 100% clearly is so that the only material which is without any doubt better at "HiRes" than 16/44.1, is DXD from 2L.

 

With all other HiRes that I ever ran into, it is from "so-so" up to "worth nothing". But with the DXD from 2L it is an all over WOW like being in another universe suddenly.

Funny thing is that I never heard anyone else raving about this so much, unless at an audio show when I show it myself.

 

Anyone any special opinions on DXD regarding this ?

(I never even considered this myself, but now reading back those emails it all comes back how I was working on that DXD market, which btw never got to live really, if you ask me)

 

 

[PeterSt]

Maybe the idea that DSD would be Walhalla ?

(not my idea of it)

[PeterSt]

2 minutes ago, Paul R said:

the clipping is from beastly compression after it was mixed

 

Try the Beastly Boys.

[PeterSt]

11 minutes ago, esldude said:

Surprise, surprise. 

 

Not really. I once dedicated Ill Communication as one of the really better sounding albums.

11.4 this one.

 

I don't want to shock people, but this one sounds even better:

5.1.

 

[PeterSt]

52 minutes ago, esldude said:

2L downloads don't convince me the hires is some leap forward.  They are the most honest comparison files I know of available for people to listen for themselves.  I personally can't hear they are an improvement. But I'm old enough higher frequencies aren't going to be audible to me

 

This is dangerous because yesterday I said quite the opposite. And you can't blame me for hearing above 13KHz. I think I also said it is not about that at all. Still it is the super most obvious that exactly that HiRes (DXD from 2L) works out. Nothing works out for me, but DXD from 2L does.

But then I made a DAC for it ... (sort of, see yesterday's post).

[PeterSt]

9 minutes ago, manisandher said:

If I upsample the 16/44.1 to 705.6 (or 352.8) but play the DXD natively (no 'upsampling', filtering or SDM) then the DXD sounds vastly superior.

 

Hi Mani - So what you're saying is that when DXD is upsampled to 705.6, it actually destroys sound. Right ?

[PeterSt]

44 minutes ago, marce said:

Again, wouldn't it be better to concentrate on frequencies we can hear....

 

Wouldn't it be better if you read what the thread is about ?

[PeterSt]

55 minutes ago, mansr said:

Here's a file containing a series of 40 kHz bursts, each 1 ms long.

 

I did not look into it, but this could imply a tad more than 40KHz. Thus if these indeed are dirac like "bursts" (as in on/off each other 1ms (??)) then I wouldn't trust my tweeters with it.

 

Remember the telephone dial ? (I forgot the name of the file) - that fried a lot of tweeters. NOS/Filterless and such ...

[PeterSt]

udial.wav ...

(this worked on IMD distortion IIRC)

 

So point is: the 40KHz you may not hear. Turn up the volume, and your tweeters may start to smoke.

Hearing distortion (like IMD again because of the distortion itself (the IMD folding down in the audio band) and you might think "nice sound". Meanwhile the real HF tone is XX dB higher.. But you can't hear it ...

 

With proper analogue filtering nothing should happen.

 

[PeterSt]

39 minutes ago, marce said:

you don't have to prove your ignorance every time I post.

 

Ignorance eh. Ach so.

[PeterSt]

1 hour ago, mansr said:

They are shaped such that the frequency content is restricted to about ±3 kHz from the centre.

 

You are correct. But I didn't make the combination with 40KHz, On/Off and that. It is still hard to digest.

So you have 40KHz on/off which comes down to ~3KHz and then we shouldn't hear anything of it ? How does that work ?

I just don't get it, that's all. Not smart enough. And of course ignorant.

[PeterSt]

Ah ... I didn't read the thread. So this was in the context of the sampling rate being 44.1Khz, right ?

[PeterSt]

I'd say it unreal. Only imaginary.