Atom processor inside Kronos...what the heck ??

[agoz]

... But then you begin to mix down...if you have 128 tracks with a lot of rich high frequency harmonic content, the computer has to fit all those harmonics into a limited number of digital pigeon holes

This is certainly a fascinating mental image of what goes on in a digital system and in its apparent perfect sense it reminds me a bit of "the sun rotates around earth" , but exactly as for our sun rotating around earth, it's simply incorrect.

Andrea

[GiantSonicRobot]

Floating-point numbers use exponents to expand the available range of numbers which can be expressed, from very small to very large - which in audio directly corresponds to dynamic range.
For more, see:
http://emusician.com/tutorials/emusic_daw_math/

Thank you for that link. (It will probably take me some time to fully wrap my head around this.)

In having a separate DAC per sample playback voice, the Synclavier was similar to all other early samplers, such as the Emulator I and II, the Prophet 2000, etc. At this point in the state of technology, the only way to change pitch was to change the DAC's sample rate; sample rate conversion (generally called "interpolation" in this context) was simply not an option. So, to play different pitches at once, you needed to have separate DACs per voice. This in turn necessitated summing in the analog domain.

I wasn't aware thet all early samplers needed to work like that. Makes me even more appreciative of the things a modern workstation like the Kronos can do.

[Zeroesque]

... But then you begin to mix down...if you have 128 tracks with a lot of rich high frequency harmonic content, the computer has to fit all those harmonics into a limited number of digital pigeon holes

This is certainly a fascinating mental image of what goes on in a digital system and in its apparent perfect sense it reminds me a bit of "the sun rotates around earth" , but exactly as for our sun rotating around earth, it's simply incorrect.

Andrea

+1...this is awesome.

To anyone who has bad mixes:
You have bad mixes because you are bad at mixing (and probably have bad source recordings, too). They'll be bad at any bit depth, sampling rate or with whatever high-end/vintage/audiophile analog gear you can throw at them.

I have crappy mixes. Do I blame Nuendo? Uh, no.

It's interesting that the above post talks about high-frequency content. This would seem to be more affected by sampling frequency than bit depth or summing math. And it's interesting to note that a lot of gear, converters and even wires can act as a filter (sometimes purposefully), which may help the situation to your ears.

[synthguy]

I've been a bit busy, but shirking work for a few hours, so I thought I'd finally get back to this.

There certainly are a number of people who advocate what's called "analog summing" - using multiple outputs from a DAW, summing them in the analog domain, and then recording the result as the final mix. However, this remains a controversial topic rather than an industry concensus. Personally, I have yet to read a single technically-sound defense of this approach. There are certainly those who prefer it aesthetically, and I won't argue with them, recognizing that there are typically many different and equally valid aesthetic opinions; see my comments in the previous post as to why that might be so (euphonic distortion, frequency alterations, etc.). That's quite separate from it being technically superior or more "pure," however.

Well, I would agree that if you want the most pristine mixes, going all digital is the way to go, perhaps even to have all the electronic instruments 'in the box.' But you like all the various acoustic flavors your gear provides, or else you wouldn't have so much of it. And flavor most certainly is a part of the package, and has nothing to do with purity, or else you would have a 'cleanest' preamp and compressor, EQ and so on, and just fill your rack with them.

As a counter-example, George Massenburg is one of the most respected audio engineers alive, both for his amazing technical expertise (including design of cutting-edge products in both the analog and digital domains) and his astonishingly great-sounding recordings (http://www.massenburg.com/c/gml/discography.html). He typically uses digital consoles, summing in the digital domain.

Well now, I'm not sure he's the best example to use to support your case, as he's one of those pioneers who's whole drive is to improve the quality of sound in the recording industry, and as far as I can determine, he doesn't consider the job to be done yet. And secondly, the stuff he uses is freaking expensive. Sure, I'd love to use gear made by Euphonix, SSL, SONY and his own stuff, and I have no doubt I'd be in love with the sound, but I just can't afford it. Even to use it. I'm not Jordan Rudess or Eddie Jobson, just Steve the synth guy.

By the way, I'm familiar with Hugh Robjohns article, I'm a subscriber to SOS. But specs don't tell the whole story. If you're into speakers like I am, you would understand that there are certain aspects of sound which you just can't quantify. With speakers, they're things like imaging and soundstage depth. A tweeter or midrange driver which is more pristine on paper might not deliver as good an acoustic image as one which doesn't measure as purely. Ribbon tweeters are notorious for having worse measurements than the traditional dome brethren, but a lot of people rave about the way they sound, and ADAM has a whole line of monitors and home stereo speakers built around their proprietary AIR ribbon. I'm giving these or something similar custom built a serious consideration.

But I must say that I've been reading up on it recently, and more than a few people have gone on and on about their analog firewire mixers, about how sweet it sounds to have an analog hunk of electronics on their front end, either from Mackie or Allen & Heath, so it seems that our affliction is rather wide spread.

I definitely agree that a good analog front-end is important. Here's some of what I use:

http://www.danphillips.com/equipment.htm

Front-end is about recording (input) rather than summing (output), however, so it's rather a different topic.

I guess I should have added the "mixing" analog part too, but I believe I mentioned it already because you commented on it above. From a Sound On Sound review of the Mackie Onyx 1640i, I keep coming across remarks like this:

I don’t know why it should be, but I always find it slightly easier to balance sounds on an analogue mixer than in software, and having 16 inputs allows even a complex a mix to be broken down into the main individual tracks plus any logical subgroups.

Now this is of course from the perspective of people who don't normally own high end studio consoles like George Massenburg does, though they do use control surfaces because no one likes mixing with a mouse. Unless they're very strange.

So this is where I'm coming from, a guy who's budget might be $3000 this year. I have to squeeze a KRONOS of some kind into the equation - well, want to, so $3000 extra is being very optimistic. I'm looking at the Presonus Studio Live 16.4.2, though one person already mentioned that lots of high frequency stuff like strings can sound dense in busy mixes, which mine most likely will be, plus it's only 48khz max. So I'm leaning towards the Mackie and the Allen & Heath ZED-R16 as far as mixers go, and the ZED has the advantage of serving as a great controller. I'm also taken with the Focusrite Liquid Saffire 56 and adding a control surface. If I can kill my Studiomaster's hum, I'm thinking of going that route because the T-2 sounds fantastic. If I had the money though, I could sure go for a TASCAM DM3200 or 4800 with Firewire.

DAW-wise, I'm looking to build an i7 PC since MACs just sbout double the price, so I'm considering Cubase 6 and Presonus' Studio One Pro. I like Pro Tools but jeepers, is it user hostile sometimes. Cubase is kind of an industry standard and has the advantage of more plug-ins and included notation software. Studio One has my attention though, as a number of people have remarked that they've switched to it because of its great sound and sheer ease of use over most anything, and only resort to porting to their old DAW to use some certain feature. And you can load it up with the usual plug-ins.

I'll try some mixing experiments when I get it all running, say do an all stereo mix vs doing every single thing to separate tracks, and do a listening test. No, I'm not expecting dramatic differences, since both Cubase and Studio One have a high bit depth, I think 64 bit for both. But who knows?

Even so, I'm thinking of keeping to 16 tracks or so, simply because of the logistics of a huge mix. Surely that wouldn't make anyone cringe.

It's interesting that the above post talks about high-frequency content. This would seem to be more affected by sampling frequency than bit depth or summing math. And it's interesting to note that a lot of gear, converters and even wires can act as a filter (sometimes purposefully), which may help the situation to your ears.

I'm up on the sound that even cables can provide for a sound. Hosa in particular are just nasty, cheap sounding stuff, so I go for more name brand cable by... (darn, standard cable maker escapes me) or Mogami. I intend to milk sampling freq and bit depth both when I get serious, as 88.2k and 24 bit should be as smooth as you can ask digital recording to be.

[Zeroesque]

If you're into speakers like I am, you would understand that there are certain aspects of sound which you just can't quantify.

Well, that's certainly the opposite approach from any digital signal processor. Too bad Fourier isn't around to tell us about it. I would also suggest that most people building or considering purchasing a flagship workstation (presumably that includes most of the people in this forum) are into speakers in some meaningful way.

...because no one likes mixing with a mouse. Unless they're very strange.

I must be strange. I don't mind a mouse at all, and it means less stuff to dust (it would be very limiting to not have a wheel, ctrl-click and direct numeric entry, however).

I totally understand wanting to try many different signal paths to find your sound, though. As with you, I'd probably try a ton of analog and digital stuff if money were no object. Since it is, though, I'm happy with the pristine -- maybe even sterile -- environment of digital. Then I can alter whatever I want with digital processing. When the mix sucks, it's because I suck (and vice-versa, though the former happens more often). I can sleep comfortably knowing that.

Let's get our damn Kronoses already and start jamming.

[synthguy]

I admire you mouse mixers, as well as the headphone mixers, because I just can't do it. I have superb Sennheiser 580s and the one up from that, but something doesn't translate well. Something is always off, and I can tweak mixes on speakers to a tee. Also, mixes are very dynamic and self-affecting. Often if I tweak something, I have to adjust something else, and it's the same way with synth programming. Gotta have speakers, really good ones, which is why I'm likely going the custom built route. I have a couple of sources for renown designs, and with todays speaker elements and crossovers, properly voiced, you can spend a thousand or so and end up with a pair of speakers that sound ungodly expensive.

Yes, I'm spoiled, and I intend to stay that way. Which is why I intend to get a system which doesn't just work well, but sounds world class. Okay, maybe not SSL, Oxford or Euphonix world class, but close. And I think I can manage that.

While I don't have access to an OASYS, I make do with my toys, primarily an M3, Radias, Virus, my luscious Arturia Origin and the new PC3. If I add in my Yamaha V50 FM synth... I'm close. I can certainly do an awful lot with them. Right now, the PC3 is humming over the speakers, so rather than beat this formerly Atom-based thread any more, maybe I'll dive into some more sound creation for the new project, and dream of what's going to happen this summer...

[KapnKrunch]

With the risk of re-opening a can of worms, does anyone with authoritative knowledge (maybe Danatkorg or Rich) know the specific model number of the Atom processor in the Kronos?

Ref: http://ark.intel.com/ProductCollection. ... lyID=29035

(Sorry - just can't help myself. It give's me something geeky to do while waiting for my K-88.)

[danatkorg]

I'm sorry, but I can't currently discuss specific parts usage.

[Jon Lord]

With the risk of re-opening a can of worms, does anyone with authoritative knowledge (maybe Danatkorg or Rich) know the specific model number of the Atom processor in the Kronos?

Ref: http://ark.intel.com/ProductCollection. ... lyID=29035

(Sorry - just can't help myself. It give's me something geeky to do while waiting for my K-88.)

I'm sorry, but I can't currently discuss specific parts usage.

No fear, Mchale is here Im sure he will find out

[McHale]

Luckily, Korg isn't able to link my handle to my real info...

I doubt I'll tear into the Kronos before I get some hours with it, but it will be torn down and pictures taken. If the late June ship date is accurate, I'll have tons of time to tear it down because I won't have enough time to get all my patches programmed to finish the summer "tour".

[EXer]

Just to add some fuel to the fire, please find below 2 links to interesting articles:

¤ Pentium 4 Vs. Atom: A Battle Of The Generations
¤ Efficiency: Core 2 Nukes Atom On The Desktop

Everything I have read about the Atom processor has the same conclusion: Atom is a cheap processor aimed towards lowest-cost laptops where performance doesn't matter


[edit: link to the 1st page of the article]

[X-Trade]

Doesn't matter.

If the KRONOS works then it works.
Interesting the P4 vs Atom article recommends the Atom over the P4. They say that neither are particularly great at performance requiring applications, compared to modern desktop processors. So they perform quite similarly under load, but the Atom's dual core version is much better because of the extra core.
The article was saying that they perform quite similarly and recommends the P4 if you already have one lying around.

Most importantly for those complaining about fans, I know from experience that the P4 runs very hot, whilst the Atom can even get away with passive cooling in most instances.

But yes, the Atom is rubbish when running windows, but KRONOS isn't windows.


The only real thing to be learnt here is that:
If Korg used a Core2 processor then there probably could have been quadruple the polyphony, probably more IFX, maybe more KARMA modules (THAT I would like to see) And it would have cost about £200 more and require noisier cooling.
If they had used an i7 then it would be incredibly more powerful but would cost closer to £500 more. And again still require more cooling than the atom, although not as much as the core2.

[sparkie]

Cheap Atom processor, no tilting screen, buttons and turnwheel have already been complained about ... Starting to look like Korg cut many corners on cost on hardware and design to make this...

[zahush76]

Can you please educate me and expand my general knowledge as to which intel processor is to be found on a yamaha motif? roland fantom? nord stage?

[burningbusch]

Can you please educate me and expand my general knowledge as to which intel processor is to be found on a yamaha motif? roland fantom? nord stage?

To my knowledge, Yamaha, Nord, Roland, etc. do not use Intel or any similar processor but rather use proprietary VLSI/LSI chips for their processing. Though they are digital systems they are not what you would call "computer-based." I also believe these are 24-bit systems in comparison to Kronos which is 32-bits. The advantage of 32-bits is you can mix multiple 24-bit waves accurately. Two full-scale 24-bit waves, for example, require more than than 24-bits to be summed without loss of resolution. I believe Yamaha (maybe Roland) do use an embedded version of Unix for their screen/file handling but not for the general processing of the synth engine, FXs, mixer, etc.

I'm sure Dan at Korg knows all this in greater detail and accuracy.