Ah, sorry ... True ... 0.9h somehow can't bear Doubling/Upsampling IMO just the same. I don't know yet why.

After some deliberation with myself and others, it seems much more is going on.

I did not try it myself yet, but a few theories coming together imply that I should leave all like it was for 0.9h, but now change absolute phase ...
What I will do is just present the options so you can play with it yourselves.
Since all is already in there, it just needs the controls to add ... Takes some additional time ... might be tomorrow now ...

So, a bit more of it Dave ...

As you (I think) will find out, people not necessarily will agree with you (and thus me). Not because they don't trust your (or my) findings, but because they just don't agree.

I'm not sure whether I wrote it in here, but I can tell you that the 0.9h version to me is the least disturbing of all. "Distrubing" is an important phenomenon, since for a longer period now I (and a few more) are not talking in terms of disturbing anymore, but in terms of "most realistic" only. So, more realistic, less realistic, and if things really go wrong : disturbing again. But it's a process ... an evolution ...

When a voice is not much realistic at all, you might not be disturbed by a too short decay of it. You might not even be able to hear what's sang or said (of course this varies very much per system, and comes in many gradations). Now, with 0.9h it was the very first time that I even could hear the cutoff of voices. But now the question is, why ?

There's two main reasons obviously :
1. Before the voice was too poorly represented in order to be able to hear the cutoff which has been there all the time;
2. Before it did not cutoff.

It seems obvious to choose for 2, if you think about it without further notice. BUT :
Since 0.9h is the least "disturbing" version of them all (well, it was anyway ) I tend to choose for 1.

Now, there's another IMO important thing, and that is the description of "disturbing" as I use it in this context :
This is about being able to play random music / styles, and none of them disturbs. You could also (or even better) say : suddenly you have more well recorded CDs than you thought before.

The latter is for me an absolute means of judgement. I have always said that, and I still do : the more well recorded CDs you seem to have, the better the playback system is.
This is more imortant (and deep) than you might expect, because it is very very easy to tweak an e.g. speaker to let it sound great for one CD. A bit of a pity is it that suddenly it is the only "well recorded CD" you have in your collection. So something is faking ...
Please trust me on the truth of this, and then use it as an absolute measurerement.

If you're pulling one CD after the other, because "this one" does not sound well, you can bet something has gone wrong with the playback.
"Has gone" because for me (and probably most) all is an active process of tweaking. So don't tune one CD only !!

So, 0.9h is the best of them all ... Right.

Assuming that my (and probably others') means of judgement is correct, and that 0.9h *is* the best of them all indeed, above choice 1 must be applicable :


You don't hear me say this is the final truth, but I do say that logically it is the only option to choose for.
Well, now it is very very easy to say that both versions are wrong. In different areas, but just both wrong.

But mind you, why it is so easy to say that ? not because with both something is the matter, but just because of the stupid simple truth that so many things are wrong with audio playback, that the chance that either 0.9d or 0.9h is the most optimum is plain zero. And I say it again : I did not even start with optimizing playback for Engine#3, although this was the kind of first explicit action of mine to this respect (and I said it in the releasenotes). Btw, I did it as an in between the lines job, as a derival of Gapless.


Yesterday I put back one of the "parameters" that should (?) have caused the anomalies. At a glance, IMO that helped. The sound became more crispy / snappy again, and I looked for the punch in drums and did not judge that as wrong. "At a glance", because I stupidly changed more and listened to that overall result only.
What happend is that suddenly I got the hunch - because of the more or less unexpected wrong way of the "parameter" - that a higher priority of Engine#3 *would* make a difference. So I build that in (was pending anyway) and I really think it made a difference.
Beware though, because I listened to a few tracks only with the parameter reverted, and the remainder of the evening I listened to the version with the reverted parameter *and* the high priority.
Btw as said before, it really takes a week or so for me to really judge, so I can't be conclusive right now anyway.

Today I will finish the version with reverted parameter and priority settings for Engine#3. I can already tell you : if this - without priority setting ! - helps towards the sound of 0.9d, but keeps the never-disturbing advantages (my context !) of 0.9h, I can overshoot the parameter in the same direction as just applied, and the sound should be better again ...
On that matter, I better introduce a slider for just that parameter ... hehe.

Peter


PS: For those who jump in here, and wonder what's happening with parameters and such : this is just about accurate playback, and no means of DSP at all (will never do that ! ).

Hi Dave,

Yes, you asked me politely whether it was okay to post this, and I even urged you to do it because I think it is *very* valuable.

Before responding to your findings, I'd like to do a bit of high-flying first :

People who read this, may think "oh well, this is his speakers and his amps a and his room, and his DAC and I don't even like nos-DACs". Blabla.
IMHO, people thinking like that, won't proceed much on things. BUT, it takes a very special skill to be able to look through all these variables, with one common denomenator only : XXHihgEnd. I think I can do that, and what it needs further is recognizing judgements like yours Dave - in consistency. Maybe difficult to explain, but if you mention a characteristic A and also a characteristic B, I would know they always go together, so if you indeed mention A and B, I'd know you just must be right. Unaccording the speakers and all you use.

What others from above could learn, is that *anyone's* sayings should never be distrusted. But okay, once the person saying it is in the environment of trustees . I mean, you certainly don't hear me say that I'd listen to any one word of x.xxx+ posters on AA, to name one. But it's easy to judge ... for me it is.


Ok, landed again, and what I write below, I already wrote yesterday, so the "yesterday's" in there, are "day before yesterday" by now.
Note that I did *not* listen to 0.9d when I wrote this. But I think I can do without. As an absolute judgement of 0.9h then ...

---

Paper like drums …
I fully agree with that. At first thought (say, 10 CDs or so) this was my major idea of more natural voicing. The first attack of drums just IMO are rather like that. But at second thoughts (the next 50 CDs …) there body missing from the drums.

I’d say that 0.9h is more accurate, which is from the applied theories in the player, and which I also seem to hear. “Accurate” also comes to my brain as more tinny. So another general remark of myself would be “more tinny sound”.
What I applied with 0.9h opposed to 0.9d is two counteractive directions : one of more accuracy and one of less (mixed two different approaches). Possibly it combines in a strange way.


Actually this is the most interesting. Why ? well, since many times it occurred to me that the phase was out of order at drum hits. I never experienced that before.
Sidenote : this is dangerous though, because since the latest 0.9h I have been working on absolute phase control, and it well might be so that I “learned” to pay attention to these things. I have been thinking about this just because I noticed the drum hits, and it’s my conclusion that it’s just there, and I’m not disturbed by it because of the “learning to hear it” process. But still dangerous.


Pwew, if only more people would understand that.
A better expression than analogue I think.


I’m not sure whether your writing can’t be clearly understood by me, but if I try to translate it for myself :

Something like pinpoint imaging does not exist with value (for audio) if it would come to one, say, square (or cubic) inch of space somewhere in front you. An instrument most often expresses the sounds from larger sufaces (like violin), and
a an area of 1 square inch where sound emerges would be far too large to express the instrument
b within 1 square inch the e.g. violin can’t be expressed.
So think of the representation of the instrument in your space (coming from speakers) as many points in space for the one instrument.
When the, say,  30 x 20 x 5 inches needed for that are now expressed as 1x1x1, the instrument is pinpointed the most separate from the others, but it would be no instrument …
It would be the opposite of organic, and it would be better expressed as digital. Not for squary sound, but for perceived “too much accuracy”.
Wrong …
This is not accurate at all. It’s the overshoot of, say, too bloomy.


Yesterday I have been listening for a full CD to music with bagpipes. It occurred to me that this may be a better instrument to judge equal tonality than a piano, just because it occurred to me that it wasn’t so much equal al all. This *is* dangerous, because I can imagine that the horn speakers are use are prone to resonating as a response to these nasal squarish sounds, but it is again something of which I thought : this does not fit.
Later I ran a CD of Joe Henry (never heard of him really, but I found him in the CD rack), and by pure coincidence or not, his voice is in the same areas the bagpipes were unequal and it sounded just bad. Mind you, a guy like this comes to you as one with an irritating voice, but since I have the experience from before (with I think London Beat), this just is not so. Something is just wrong.
It is really un-be-lieve-a-ble what mismatches in this area can do to a person’s voice and the recognition of it. When the mismatch has gone, the person sings 1 octave or so lower ! So all what was creating the irritating voice, were over expressed harmonics.


Many, many times I felt like words were abbreviated. That too, did not occur to me before.
Besides that, also many times I could clearly hear the opposite of the bagpipe thing : frequencies are underexposed. Thus, apart from words abbreviated, it occurs to me that parts of words just fail. That too, I did not hear before.


Not almost, I think it is true, incurred by harmonic distortion.

Btw, the disbelievers of software being able to change the sound just by being (in)correct must be rolling on the floor by now. :lol:


I will put up a new version that undoes the part of the theoretical less accuracy (“theoretical” because I know it does influence, but because of combinations with other things changed, it is hard to predict the “direction” at the other end (say, the speakers).
I will not undo all, because I *have* to learn what does what.

Btw remember, more accurate should always be better, since I believe that the most accurate will e.g. let bloom the instrument as realistic as possible.

---

Thus, the above I wrote yesterday. I can now add a few things to it.
Sidenote : the bagpipes I mention are a coincidence, because they weren't mentioned in your original writing to me.

I decided to add the remainder of what I can say in another post.
Peter

Theories are good I think. There may be (or will be ?) jitter implications though. It's really something to test out ...

Must be so, because that's when it occurred to me the first time (as a choice).

What about the latest RDP client for the PDA ?
I really don't know how to help you here

Actually this is strange ...
But now I think of it ... maybe you're running XP on the remote ? I mean, Vista/Engine#3 would only be more direct than anything else. It goes through about every OS thing. This is not the case with XP (and Engine #1/#2).

PS: You are running XXHighEnd on the remote, do you ? ... you could run it on the PDA, but then it would never work ...

?

Boggie, no ...

Maybe I'm a bit short of time to respond to everything, but let's start with the comparing thing anyway ...

The reason I said it's not easy, was from my perspective of having a "raw" file comparer tool, that's not intelligent.
With the latter I mean that it should leave out the leading (and trailing) bytes, which already would be off because of the CDRom drive not being able to find the exact beginning of the track. Or better, you may have not calibrated it for that (which by itself is a feature of EAC, but which might fail because it needs reference CDs you might not have). So, if the first 100 bytes are off, and the comparer is not inteligent, all will be off.
If you compare the things yourself like with WaveLab, you'd have to find the common denominator of the start, and chop off the bytes from the track that has the additional bytes. For more complicated looking waves it can turn out to be undoable.
So that's what I meant. And I don't know about the bit comparer from Foobar. It might be "intelligent".

Btw Chris, thanks for pointing out the ToolTips. I was already wondering how I came to my choices myself (a few years ago), today not understanding the plain labels on the form. The means of arranging for it all looks still strange to me, but ok ...

Anyway, yes, it is ridiculous that we can't seem to read audio from a CD error free, but that's life for now. Btw, there was a thread on bd-design once where all this was worked out to some extend, and although this thread has been deleted by the owner, I still have it myself, and there sure are means to improve the reliability. I have to make something for it though, and somehow I can't do everything at the same time.
My conclusion for today in that thread was : If you have the opportunity to have two drives and both calibrate them properly, and they both always produce the same data, you're okay with either drive. BUT, you'd have to be very careful in watching the performance of the drive you use; Once the performance degrades (gerenally : it takes longer to rip), do the comparison with the other drive again. You *will* be behind things though, because you'll always decide to check when you ripped several CDs wrongly. But it's a means ...
Oh, and please note that with more poor CDs there's hardly a chance that both drives will read the same data. So on that matter ... have three drives.
Quite awkward ...

The least you will have relatively soon is a means from within XX to check the tracks on "realistic audio data" ... I could say, that if you just can hear the difference from two different rips, 100% sure my checking for realistic audio data will come up with something.
Generally spoken "that you can immediatly hear the difference" would be rare. Or maybe not, but don't make too much fuzz about it until yuo just have proven it (somehow) to yourself.

Oh, and to be clear on things : once the track is on the HDD, it can't be readout wrongly (better : differently) from there.


Looking the way EAC is setup (or presented) this might not be a bad approach. I mean, look how the procedure goes (or how it comes to me) :
You can check the box for "drive supports CRC", and in some later stage you can upload the "findings" from your drive to whereever it is for the next person. So you judge or do wrong, and the next one is in trouble.

As a sidenote, I mention that the best means of ripping (which is not necessarily "insane" mode), often is undoable. I mean, the rip of one CD could take hours and hours, just because EAC wants to make something of something that will fail anyway. Never noticed it ? as soon as it takes 10 minutes to read 5 seconds, you will have a glitch, no matter what. Solution ? rip less secure, so you'd at least have ripped 30 CDs for the night ...

Actually it all IMHO s*cks. No matter how good intentions are.

Might it help somewhat for now : Plextors always have been the best.

Peter

So you don't have this (topmost) setting in there ...
 

Is that CE 5.0 ?
If 4.2, you could try to get 5.0 or find out whether that helps ...
(or Google for registry hacks in the PDA -> not the most simple thing to do, but it can be done)

And Jack, because your data seems inconsistent to me  could you please check whether it's indeed the loading time that's holding things up ...
You could look at the disk activity lamp at the server ...

When you see that loading just starts too late (like 2 seconds before the end of a track instead of approx. 9), could you measure whether the running time of a track is correct ? ... like a track of 4:00 minutes taking 4:10 ...

LOL Jack ...

Besides that it could be done (pre-load a whole, say, Playlist), I don't understand what is happening with you.
First of all, 3 or 4 secs for a track over a 100Mbs network ... hardly believeable. That would be 4 x (roughly) 10MByte equalling 4 minutes playtime or so. But ok, if your tracks are all that short ... possible ...

20 seconds would be 20 minutes playtime. That too is unlikely, since 4 minute tracks would load in 4 seconds. Haha.

All 'n all you currently have 9 seconds for loading a track; if that also doesn't work for a local drive (even PATA) ... what is going on ?
And of course you say it has gone worse, and actually I didn't even change anything ! well, not that I know of. Not from 0.9g to 0.9h.

What Engine are we talking about anyway ?

Peter


PS: Didn't your disks spun down in the mean time ?

I don't know about "client" PDA's, but did you apply the setting "Leave audio at the remote PC" ?

If this setting is not available at the PDA (for RDP settings), that's your problem I guess ...

Peter

Psychological drama ...

That would be a difficult job, because you wouldn't know whether it changed the result.
Better compare the two files then (which is not so easy, for reasons).

That would be correct.
What got into the author of presenting logic like this, I don't know. It feels kind of upside down and unrelated.
I have that box checked as, indeed, per the selfdetections of the program.

The more I think of this, the more I realize that it doesn't say anything ...

Djeezz, I know the cache should not be used, but how is that setting then ?? 

I now, and only just now see this checkbox under the Paranoid Mode radio button. You know, this button we never choose because it's "not recommended".
But let's be fair, when you see *this*, what would be the conclusion ? ...