do we need 128GB DDR4 now??

, I think you need to understand how that all works before criticizing the feature, or perhaps you're audio system is not up to the task. There is no more muddying of the sound when using perspectives then there is when moving from 2 channel to 4.

Hello Sonar.
This might tell more about the quality of your system, not hearing the muddying and degradation of the sound.
My first encounter(sample set) with the perspective feature was Brasov. From the moment i played the first notes, i noticed the worsened dull sound using the perspective sliders and really did not like it and i still don't.

Regards Jan

Nope, nothing wrong my audio chain; aren't you very strongly in favour of running only 2 speakers for the front? (Or do I have you confused with someone else)? Perhaps you need to add a few more speakers up front so that the sound becomes more clear All teasing aside, we clearly have different ideas of what sounds good. I have no experience with Brasov, but Rotterdam Transept is crystal clear where I have the slider (somewhere around 70% for practice).

We're going to have to agree to disagree, but what you are asking for is "the moon"; there is no way a producer can produce that many variations of a large sample.

Actually i am playing the transept at the moment. Setting the slider to 70% gives as expected a more direct sound, but also some harshness and loss of detail and spaciousnes is apparent.
I have to admit that with the Transept using the slider is by far better sounding then Brasov.
But still it is not my cup of tea.
And yes i am the one with only two high end speakers in the front.


Regards Jan

"Loss of spaciousness" ... Wouldn't that be what one would expect as you move the virtual position closer to the pipes? In other words the ambient farther back more reverberant sound is lessened.
But why you would get a loss of detail as you move forward is a puzzle. One would expect more detail, not less.

perhaps this is an "american" thing? the bigger the better? (yes I know adri, you are originally dutch, BUT your experience is surely much influenced by the time you have spent in the US). not everybody shares this mentality, and THEY are potential Hauptwerk users too!

Actually, the organs in Germany have historically been extremely large. But the point I want to make here is that your assumption is simply erroneous; just because I live here in America doesn't mean I like big organs. I have always favored quality over quantity. I am not influenced by my American experience whatsoever. I absolutely prefer especially historical European organs.

I simply started a discussion about the future; about what will truly be required for more realism in the future. It's possible that large RAM requirements may become normal even for smaller organs, as we fine tune the sampling even further.

Digital cameras used to start at around 2MP, and now the most expensive models are up to 80MP already!

It is simply a question about the inevitability in the digital world of software and hardware being in a race with each other.

If cars had evolved like computers, a Rolls Royce would give a million miles to the gallon and only cost a nickel.

That kind of discussion: about technological progress and the inevitable. After all, we do need to upgrade our computers about every 2-3, or at least every 5 years. New OS's and new software all require us to do so.

Thanks.

in a blue-sky scenario IMHO for the most absolute realism it will come back to CPU power more than RAM, because I would think the most realistic would be having close-miked pipes with per-pipe individual convolution, that way you would be able to have 'realistic' listening positions wherever you wanted them, but the computational requirements to do that (and possibly in addition some modelling of interference between the pipes and so on) would be huge

However I am thinking that memory is growing faster than CPU power, I am honestly wondering how long it will be until we get solid state storage that has enough bandwidth and low enough latency to allow HW to run entirely "from disk", I wonder if Martin has thought of trying this with some of the really high performance m.2 solid state disks available nowadays

I have similar thoughts about technology moving in unpredicted directions. 128 GB now?? Not for me though if I build a new Windows machine I'll certainly pick a motherboard that will handle 64GB but would initially settle for 32GB with budget considerations in mind. As others have said it's the nature of technology to grow and suck up the existing resources, but I could never justify 128GB at this point in time. The largest sets will load in 64GB in the configurations I'd be likely to consider, that is, less than six channels.

"Loss of spaciousness" ... Wouldn't that be what one would expect as you move the virtual position closer to the pipes? In other words the ambient farther back more reverberant sound is lessened.
But why you would get a loss of detail as you move forward is a puzzle. One would expect more detail, not less.

Hello Rob

As Martin pointed out, a in between setting is not the same as a true recording of the sampless in between.
With 2 perpectives you will get al sort of mixing of the notes, the harmonics and the different reverbs. Each perspective can cover some finer details of the other pesrpective and vice versa.
You wil encounter phase problems and as a result loss of detail.
Not to mention random detuning.
I expect that the same notes played and mixed in different perspectives may not be synchronised regarding detuning. Thus introducing more phase problems and distortion.

Regards Jan

"Digital cameras used to start at around 2MP, and now the most expensive models are up to 80MP already!

It is simply a question about the inevitability in the digital world of software and hardware being in a race with each other."

I take your point but can't resist reporting the following. I like taking photographs, you can see them on my homepage. they have been taken with the same digital camera, which I value very much, for many years. some time ago I noticed that the pictures no longer have the correct information about the date when they were taken embedded in the file. instead the date is now given as 1 january 2000. a friend with whom I shared pictures complained that this tended to mess up his record keeping... right, so I finally got round to trying to adjust the date. it turned out it can only be set for dates between 2000 and...2010! so obviously the problem must have been there for some time...which, as we all know, flies...

point of the story: it has, in fact, often been remarked that more pixels or memory do not necessarily a better camera make. I have another, newer one which I bought because it is very small. its specs are practically the same, the quality of the pictures it takes is somewhat to rather a lot inferior (depending on lighting conditions). SO, it really is not necessarily the case that more bytes improves the product, or that one using less bytes will eventually prove inferior. I cling to that camera because it takes EXACTLY the pictures I want! in what the camera calls "super high quality", i.e. at about 3 MB a throw. I guess a different camera could make them even larger but what would I need that for?

I had a thought for a Hauptwerk feature that might help those of us with limited memory still enjoy these more sophisticated sample sets.

What if it was possible in the "initial settings" dialog to blend channels with certain weights? In other words, you could set your desired listening position, the two channels would be blended into one and cached, and the memory footprint would be that of a two (or four if you keep the rear) channel system instead of six? If you had sufficient disk space you could potentially even support having multiple virtual instruments (virtual virtual? ) which you could switch between by loading a new instrument.

The only downside would be not being able to adjust your listening position in real-time, which would be reserved for those with the memory to load all 6 channels.

What do you think, Brett? Is this something worth putting on the feature request queue?

Jiri, might this be something that Sonus Paradisi and your fellow sample set developers would be willing to support?

"Loss of spaciousness" ... Wouldn't that be what one would expect as you move the virtual position closer to the pipes? In other words the ambient farther back more reverberant sound is lessened.
But why you would get a loss of detail as you move forward is a puzzle. One would expect more detail, not less.

Hello Rob

As Martin pointed out, a in between setting is not the same as a true recording of the sampless in between.
With 2 perpectives you will get al sort of mixing of the notes, the harmonics and the different reverbs. Each perspective can cover some finer details of the other pesrpective and vice versa.
You wil encounter phase problems and as a result loss of detail.
Not to mention random detuning.
I expect that the same notes played and mixed in different perspectives may not be synchronised regarding detuning. Thus introducing more phase problems and distortion.

Regards Jan

In home audio / video, mixing is done on the fly all the time for multi-channel within the various codecs available.

The detuning issue sounds most correct, but it's not a problem; if you suspect that might be an issue then just turn off random detuning for that sample, hardly a deal breaker.

I repeat; if you think the transept organ sounds muddy, you really should invest in more speakers, as that is one of the ways to reduce muddiness and distortion. I'm not saying your 2 speakers are poor quality, I'm just saying adding more to the mix will clear things up.

In home audio / video, mixing is done on the fly all the time for multi-channel within the various codecs available.

The detuning issue sounds most correct, but it's not a problem; if you suspect that might be an issue then just turn off random detuning for that sample, hardly a deal breaker.

I repeat; if you think the transept organ sounds muddy, you really should invest in more speakers, as that is one of the ways to reduce muddiness and distortion. I'm not saying your 2 speakers are poor quality, I'm just saying adding more to the mix will clear things up.

Hello Sonar
You are making quite a lot of assumptions in your answer that are wrong.
1e the Transept is by no means a muddy organ. It is a delightfull organ with a very transparant sound which i use a lot but only with the slider in the diffuse setting.
2e detuning is not a deal breaker but it is there and should be used, but if it somehowe has any worsening effect then i hope Martin will correct this issue in next updates, synchronising front ,diffuse and rears to have the same tuning. It's not a very big issue however.
3e You don't hear muddiness so its not there!!
Well i also don't hear muddynes either in al my other sample sets.
They are crystal clear. I only hear muddines(some loss off transparancy might be a better discription) when i use the slider. So adding more speakers, with in my opinion to much broadening of the organ front just in front of my nose and corrupting to some degree the stereo image is not a solution i want to pursue.
With the oldfashioned surround sets the sound is always ultra transparant in my setup.
It might even be possible that using more speakers with their broad sound front, and their interactions with each other, corrupts in some degree the subtle spacious information recorded with stereo mics. This might be the reason you can't distinguish the things i described with your setup.
Don' t take me wrong i think many sets with variable perspective will give a lot of playing plessure by lots of people. But audiophiles like me (which is an other hobby of me) can hear the drawbacks of using variable perspectives and i think it's a pity.

Regards Jan

1e the Transept is by no means a muddy organ. It is a delightfull organ with a very transparant sound

Great, glad we agree

2e detuning is not a deal breaker but it is there and should be used, but if it somehowe has any worsening effect then i hope Martin will correct this issue in next updates, synchronising front ,diffuse and rears to have the same tuning. It's not a very big issue however.

Exactly, there is no problem with variable perspective in regards to tuning. You can turn it off.

3e You don't hear muddiness so its not there!!
Well i also don't hear muddynes either in al my other sample sets.

You hear muddiness so its the sample / feature's fault and not your system which is "high end"

They are crystal clear. I only hear muddines(some loss off transparancy might be a better discription) when i use the slider. So adding more speakers, with in my opinion to much broadening of the organ front just in front of my nose and corrupting to some degree the stereo image is not a solution i want to pursue.
With the oldfashioned surround sets the sound is always ultra transparant in my setup.
It might even be possible that using more speakers with their broad sound front, and their interactions with each other, corrupts in some degree the subtle spacious information recorded with stereo mics. This might be the reason you can't distinguish the things i described with your setup.

That's some nice, vague hand waving. You don't corrupt the stereo image by adding additional stereo pairs, you simply create additional sound sources. Each source has it's own stereo field. Yes, it's a bit of a compromise, but all of HW is a compromise. You compromise by attempting to play too much sound through 1 speaker, when you can simply add another stereo pair and suddenly a speaker that was trying to play 400 pipes is only playing 200; this gains you clarity and reduces distortion.

Remember, we're talking about a pipe organ with thousands of pipes, not a women singing with guitar (where a soundstage becomes important). You're not going to corrupt the field by adding additional pairs, the pipes just move around a tiny bit. Many high end speakers do that anyway with tweeters, midrange, and woofers.

Anyway, this has gone on long enough. Enjoy your audiophile system, I certainly enjoy mine

hello Sonar

Nice meting you again on this forum.
It's al a compromise. You chose a other compromise then i did.

Enjoy!!

Regards Jan

[mdyde]

Yes, gentlemen, I think it's time we drew this to a close, please. We all have our individual preferences.

Replying to first two posts:

My advice would be that when you plan to upgrade your computer, take an X99 motherboard (an MSI X99 SLi PLUS is one of the 'cheapest' but has everything you need - including great looks - and not much more), use 2 or 4 16GB DIMMs (they come with a price premium, at least here in Holland, but it is not extremely expensive and much more cheaper than replacing the RAM later on) and you will have the option of 128 GB for the future. Windows can handle that without problems, so no direct need for server like solutions (unless you want 2 CPU's).

Regards
Pim

I had also found recently while looking on Newegg, that many of the LGA 2011-v3 MB's take 128 gigs of memory. I also found that one can get 4- 16 gig modules for about $520 and 8- 8 gig modules for around $400. So it's not that much more to go with the 16's and have 4 more spaces left to add another 64 gigs later.

Eric