Korg Forums

I'm the guy who wrote the blog post (though kronoshacker was first - and seriously, forum admins, get rid of that word replacement, it's insulting and childish; if you have an issue with what we're doing, let's talk).

There's nothing wrong with what we're doing. Modification of musical instruments has been a tradition for millenia. Just because now they're digital doesn't mean anything has changed.

Obviously this isn't for everyone. I deliberately did not write my blogpost "for dummies" or provided an update package that just does everything for you. This could be done if there is enough interest, but it takes a lot of effort to do that in a safe manner, and then you become responsible for keeping up to date with things for your users. Years ago I helped author a similar piece of software that lets you take control of your own hardware, and that one has over 5 million users worldwide with zero reported cases of hardware damage or bricked devices - but that took a lot of effort and a cat and mouse game with the manufacturer (and that was a $250 device, not a $3000 device). I am not quite willing to go there for the Kronos right now

GregC wrote:not an improvement here;

"However, there is one disappointment: with the new hardware, the Kronos has (slightly) less available sample memory than with the old one."

That's before I did the rest of the stuff in the article. By the end, it has more. Specifically, 2277MB of user-accessible sample memory (with nothing autoloaded).

levioter wrote:Just a remark ! I don't know if anyone noticed but one picture shows the cooler on top of the new processor so this is also a difference that can remarked for those that are already ranting against the actual fan noise !

Well, the very first thing I did when I bought my Kronos years ago was replace the fan with a quieter one. The new motherboard is certainly noisier than the silent fan... but the fan speed is dynamic. It's been a while, but I get the impression that, at idle, the new motherboard fan is still quieter than the old case fan that came with the Kronos. And hey, at least now the fan has a good excuse for being there, what with the 200 voice polyphony

enigmahack wrote:The only thing stopping me from going through with this all is the Korg updates. The Korg updates would break these patches, and unless the developer of this update keeps on top of each of the patches with continued support, it's somewhat of a niche upgrade and I'm afraid I'd get too used to having all this horsepower available and then I can't take advantage of it later after an update that might bring new features or stability I want.

Actually it's relatively easy to manually install the bits and pieces of the updates that matter. In fact, I rooted my Kronos at 3.0.2 and then just manually installed 3.0.4. All you have to do is copy out the contents of the encrypted filesystems again to use the new synth software version; there are no actual binary patches to Korg software involved (yet). This could be automated in the future.

SeedyLee wrote:Indont think this is something that can easily be prevented by Korg through a software update either, unless everyone is issued with new auth keys and they fix the underlying vulnerabilities.

Realistically speaking, Korg can't stop any of this without coming up with a completely new model that uses secure boot and a new security model (this is unlikely; it takes companies years and a dedicated security team to learn how to do this properly). These motherboards don't have TPMs and can't be upgraded to proper secure boot. I'm a professional security researcher; from my point of view, the Kronos has no "real" security, just a bunch of obfuscation (i.e. it does stuff to try to hinder hacking, but it can't stop it). I am confident that anything they throw at me in an update, I can work around. This isn't really a case of "fixing vulnerabilities"; the Kronos does not have a comprehensive security model.

jimiki wrote:do you remember, few years ago there was one guy doing this stuff and was banned from this from this forum ? any dejavu situation ??

I read through that thread this morning. It's kind of funny, really. Heider was his name. He provided enough evidence of his work to make it obvious that his results were real, but at the same time, a lot of what he said was completely wrong from a technical point of view. He also had a bit of an attitude

amit wrote:1: Is the current system overclockable? (i doubt that)

If you use an overclockable CPU, sure, but why would you do that? Even the most ridiculous combis don't really exceed 50% CPU usage at max polyphony. It's overkill already.

amit wrote:2: Has the new build been load tested?

I've seen some crashes which I'm investigating. I think some are synthesizer bugs (e.g. so much polyphony that was never tested on the original hardware that it hits corner cases). Obviously time will tell; I only did this literally a couple days ago.

amit wrote:3: what is the temprature like (in contrast)

Haven't installed the sensors tool to check yet, will report back once I have numbers.

amit wrote:4: what is the power consumption like (in contrast)

No power meter, maybe I should get one... the new CPU is much newer than the old Atom, so that is something to consider as a big power advantage (and in fact non-Atom CPUs are much more power efficient, relative to CPU performance, than Atom CPUs; the only thing Atoms have going for them is low max power consumption and that they're cheap). OTOH, I had to disable some of the more aggressive CPU power saving features to avoid instability with the real-time kernel, so that will affect power efficiency.

amit wrote:I am not sure if there is any copyright infringement regarding disclosure/distribution of some information of the system that is non GPL (there might be and korg has the right of objection).

None of the stuff on my blog is "copyrighted" by Korg; you can't copyright facts or ideas. Obviously if I were actually distributing Korg-authored software or binaries it would be a different story, but I'm not.

amit wrote:1:If it were a 64 bit system (More Ram , 16-32 gig)

The CPU is 64-bit, but the OS isn't. Unfortunately, Korg's kernel module is 32-bit. Getting that to interact with a 64-bit system isn't completely impossible, but it's a much bigger project. Virtualization might help here.

amit wrote:2: Had a Sleep/wake mode (System could be on for months in the studio,ready in couple of secs)

TBH until now I just left my Kronos on 24/7 because I hated the boot times. I might actually start turning it off now that it's more reasonable. Sleep/wake is not impossible, but there are zillions of corner cases with that sadly, and getting it to work stably, especially with the realtime stuff in the Kronos, is impractical. It might be possible to use some kind of trick though. Something like unloading all of the Korg/realtime stuff before going to sleep (to get rid of that problem), and re-loading it on wake, but then patching the synth to NOT reload all the samples into memory since they're already there. Again, bigger project, but perhaps feasible.

amit wrote:3: Perhaps could accommodate external interface. (like a small desktop module handling all the crunching the internal motherboard does. (thinking out loud)

You can already just pull out the cabling (all you need is a USB device socket and a serial port) and hook it up to an external PC and use that as your Kronos' CPU instead of the internal motherboard.

amit wrote:4: USB 3.0 and above support.

That already works. Obviously the built-in ports are 2.0 only, but you could replace them with 3.0 ports or add a new set of ports if you were so inclined. The motherboard does 3.0 and the new kernel supports it, if you plug in something directly into it, inside.

FWIW, if you only care about the polyphony, I think it is possible to do this upgrade without any software modifications. You'd need to use my motherboard (it has an Ethernet chipset compatible with the original), and a PCI Express card with a USB 2.0 EHCI controller. The original kernel would be compatible with that. Then the only thing left is to do the TOLUD fix and replace the CostProfile file to get the extra performance, but you can do that from a rescue boot USB stick and that should not affect future Korg updates in any way (CostProfile isn't part of updates, and there is no reason for Korg to start checking it).

One thing I have planned for my Kronos that may interest you guys - I want to convert the S/PDIF output to ADAT, to get all 6 output channels (L/R and 4 buses) out to my DAW. This would be a hardware mod completely unrelated to any software stuff (already have the design planned out). I'm wondering if anyone else is interested? This mod would be relatively simple to install but would require fine soldering skills (you need to solder onto a few points on the Kronos NKS4 board). I'm wondering if I should look at making a small run of boards (perhaps 10 or so) and sell them. It wouldn't be expensive, maybe $50 or less. The design will be open hardware, but I'm curious if there's interest in this or I should just hack up a single board for myself and call it a day.

SeedyLee wrote:Indont think this is something that can easily be prevented by Korg through a software update either, unless everyone is issued with new auth keys and they fix the underlying vulnerabilities.

marcan wrote:Realistically speaking, Korg can't stop any of this without coming up with a completely new model that uses secure boot and a new security model (this is unlikely; it takes companies years and a dedicated security team to learn how to do this properly). These motherboards don't have TPMs and can't be upgraded to proper secure boot. I'm a professional security researcher; from my point of view, the Kronos has no "real" security, just a bunch of obfuscation (i.e. it does stuff to try to hinder hacking, but it can't stop it). I am confident that anything they throw at me in an update, I can work around. This isn't really a case of "fixing vulnerabilities"; the Kronos does not have a comprehensive security model.

I had always wondered how they secured the key exchange, given they don't have a secure boot environment. Afterall, the keys have to be in memory at some point in order to mount the cryptoloop volumes, and would be obtainable if root access was obtained. Looks like obfuscation was the key here (excuse the pun!).

jimiki wrote:do you remember, few years ago there was one guy doing this stuff and was banned from this from this forum ? any dejavu situation ??

marcan wrote:I read through that thread this morning. It's kind of funny, really. Heider was his name. He provided enough evidence of his work to make it obvious that his results were real, but at the same time, a lot of what he said was completely wrong from a technical point of view. He also had a bit of an attitude

I remember some pretty interesting conversations around that time. I seem to recall one claim he made being that all the audio processing was being done on what he called the SHARK, which was really the NK4 board, and that the Linux OS loaded the EVA img into the SHARK. I have no doubt that he made some progress, but a lot of what he was claiming did seemed implausible (and in direct contradiction to an interview that Dan Phillips gave about the OASYS in 2005):

http://www.oreilly.com/digitalmedia/200 ... oasys.html

It's funny that you've published details about this upgrade now, because I was literally wondering if the same thing was possible a couple of weeks ago when I started a thread about whether an upgraded Kronos is on the way. Whilst it's not something I'm keen to try right now, it may very well be if a motherboard comes along that is more compatible with the D510MO board already in my Kronos (physically, electrically, thermally and in terms of IO).

marcan wrote: One thing I have planned for my Kronos that may interest you guys - I want to convert the S/PDIF output to ADAT, to get all 6 output channels (L/R and 4 buses) out to my DAW. This would be a hardware mod completely unrelated to any software stuff (already have the design planned out). I'm wondering if anyone else is interested? This mod would be relatively simple to install but would require fine soldering skills (you need to solder onto a few points on the Kronos NKS4 board). I'm wondering if I should look at making a small run of boards (perhaps 10 or so) and sell them. It wouldn't be expensive, maybe $50 or less. The design will be open hardware, but I'm curious if there's interest in this or I should just hack up a single board for myself and call it a day.

That's a cool project, but it's not something I'm terribly interested in, to be honest. I just use the analogue outputs into an ADAT converter and that does the job for me. I'd love a future Kronos to have ADAT in and out though: being able to record 8 channels and have 8 outputs would allow me to genuinely use it as a standalone workstation.

SeedyLee wrote:Whilst it's not something I'm keen to try right now, it may very well be if a motherboard comes along that is more compatible with the D510MO board already in my Kronos (physically, electrically, thermally and in terms of IO).

Unfortunately, modern chipsets always go with xHCI, so the best you could do is go with an obsolete chipset/CPU that still has EHCI, or use an external PCIe card with that. Otherwise, you're at least going to need to use a new kernel to get USB to work.

Modulo that, the motherboard I picked ended up being surprisingly compatible with the old one. You just need to figure out the ATX12V connector, but I'm sure someone somewhere sells ready-made adapters to pull that out of the ATX connector...

SeedyLee wrote:That's a cool project, but it's not something I'm terribly interested in, to be honest. I just use the analogue outputs into an ADAT converter and that does the job for me. I'd love a future Kronos to have ADAT in and out though: being able to record 8 channels and have 8 outputs would allow me to genuinely use it as a standalone workstation.

ADAT in is less useful as a mod; the most you could do is get 4 channels, replacing the analog inputs and S/PDIF inputs with ADAT. I'm mostly interested in output. I was thinking of bringing out the 6 output buses, plus mirroring the analog inputs to the remaining two ADAT channels (might as well, it's one extra wire to hook up).

Great news! Can't wait when Windows VSTs will be accessible from Kronos, but it will be no Kronos anymore.. So, it makes sense only for repairing Kronos if its motherboard dies in distant future. Also, may be we'll be able to fix bugs ourselves?)

https://www.kvraudio.com/forum/viewtopic.php?t=414077
http://breakfastquay.com/dssi-vst/
https://github.com/abique/vst-bridge

> So, it makes sense only for repairing Kronos if its motherboard dies in distant future.

So no one wants full polyphony and less than 20 seconds boot up time?

I do!

If I understand correctly, a sub 20 second bootup time should be possible even without replacing the motherboard by simply using a custom kernel and decrypting the images.

With a bit of work, it may even be possible to create an unofficial software update to do this.

Obtaining more polyphony will obviously require an under-the-hood upgrade.

I'm also an embedded Linux developer (A/V set-top boxes, cable modems etc.), so this kind of fiddling would probably be within my capabilities. I work with secure boot systems on systems with MIPS and ARM cores. Mainly disaster recover image download mechanisms at the minute.

Unfortunately, I just don't have the time or mental energy after the day-job. I just want the system to work as a keyboard.

Good luck with the efforts though!

aron wrote:> So, it makes sense only for repairing Kronos if its motherboard dies in distant future.

So no one wants full polyphony and less than 20 seconds boot up time?

I do!

That's all I want. For me, Kronos would be perfect.

SeedyLee wrote:If I understand correctly, a sub 20 second bootup time should be possible even without replacing the motherboard by simply using a custom kernel and decrypting the images.

With a bit of work, it may even be possible to create an unofficial software update to do this.

Obtaining more polyphony will obviously require an under-the-hood upgrade.

Nah, the filesystem check isn't that big a portion of the boot time. A big part of the gain comes from loading banks much faster with the new CPU and SSD. A full set of samples is bigger than the entire Kronos root filesystem.

FWIW, with most of RAM used with samples, my upgraded kronos takes about a minute to boot. I haven't quite seen 20 second boot times, but I haven't tried too hard (I could shave a couple of seconds from the GRUB menu, maybe some BIOS settings, etc). I think with no EXs loaded mine boots in about 30-40 seconds.

marcan wrote:I think with no EXs loaded mine boots in about 30-40 seconds.

Well, I have a bizarre proposition for you.

If you're honestly interested in supporting this, I'm willing to offer up my Kronos 73 as an 'external test device'.

I have an IT background, and understand Linux well enough but I'm more of a sound designer/programmer. That said, I'm willing to put the money down and get the upgrades done to this thing if you're willing to help support getting an automated update process in place.

I could document my experience as well, work with you to share the end-user experience to make this more "publicly accessible", writing guides in less technical language, etc.

Hell, if this works, I'm willing to donate some funds to make this more lucrative a project to make easily implementable for older/out of warranty/broken synths.

Sign me up, I'll be an alpha tester for you.

My email is dave. champagne 1 **at** gmail (dot com), feel free to email me and I'd be happy to help out, order the parts, do the installation and be a second tester for you.

enigmahack wrote:

marcan wrote:I think with no EXs loaded mine boots in about 30-40 seconds.

Well, I have a bizarre proposition for you.

If you're honestly interested in supporting this, I'm willing to offer up my Kronos 73 as an 'external test device'.

I have an IT background, and understand Linux well enough but I'm more of a sound designer/programmer. That said, I'm willing to put the money down and get the upgrades done to this thing if you're willing to help support getting an automated update process in place.

I could document my experience as well, work with you to share the end-user experience to make this more "publicly accessible", writing guides in less technical language, etc.

Hell, if this works, I'm willing to donate some funds to make this more lucrative a project to make easily implementable for older/out of warranty/broken synths.

Sign me up, I'll be an alpha tester for you.

My email is dave. champagne 1 **at** gmail (dot com), feel free to email me and I'd be happy to help out, order the parts, do the installation and be a second tester for you.

Thanks for the offer! I'm glad more people are interested. There are a few ideas that I have about how to convert this into a "proper" user-accessible installer. My standards are really high for this stuff, and that is kind of at odds with the amount of unrelated things I have going on in my life until the end of July or so (I won't have enough time to do what I consider a proper job until then). I also want to get a chance to experience first hand the effects of the mod on my unit for a while.

If you're okay with the wait, I'll let this stuff cook in my head for a while (and probably work on bits and pieces) and let you know once I have time to start cooking up a more user-friendly package for this.

In particular, I think it's necessary to develop a minimal userspace interface for the NKS4 so that we can convert the upgrade process into an interactive one, and actually show menus and proper informational messages/logs to the user, since I'm not comfortable with a "blind" process and single-line messages with no possibility of user input like regular Korg updates. As for future Korg updates, they would be handled through the same process - instead of installing the original update from Korg, you'd effectively copy our installer tool *and* the korg update to a USB stick (with the Korg files in a subdirectory or similar) and it would update the Korg stuff "its own way" (e.g. mounting and extracting the encrypted images instead of just copying them). We could have multiple operation modes, e.g. one where we only use the new kernel but keep all the Korg stuff untouched, one where we install the "non-intrusive root", and one where we fully declaw everything and get rid of loadmod.ko.

marcan wrote: Thanks for the offer! I'm glad more people are interested. There are a few ideas that I have about how to convert this into a "proper" user-accessible installer. My standards are really high for this stuff, and that is kind of at odds with the amount of unrelated things I have going on in my life until the end of July or so (I won't have enough time to do what I consider a proper job until then). I also want to get a chance to experience first hand the effects of the mod on my unit for a while.

If you're okay with the wait, I'll let this stuff cook in my head for a while (and probably work on bits and pieces) and let you know once I have time to start cooking up a more user-friendly package for this.

In particular, I think it's necessary to develop a minimal userspace interface for the NKS4 so that we can convert the upgrade process into an interactive one, and actually show menus and proper informational messages/logs to the user, since I'm not comfortable with a "blind" process and single-line messages with no possibility of user input like regular Korg updates. As for future Korg updates, they would be handled through the same process - instead of installing the original update from Korg, you'd effectively copy our installer tool *and* the korg update to a USB stick (with the Korg files in a subdirectory or similar) and it would update the Korg stuff "its own way" (e.g. mounting and extracting the encrypted images instead of just copying them). We could have multiple operation modes, e.g. one where we only use the new kernel but keep all the Korg stuff untouched, one where we install the "non-intrusive root", and one where we fully declaw everything and get rid of loadmod.ko.

I'm in the middle of selling my house and such, so I'm in no rush either honestly. I just wanted you to know that not only do you have my support but I'm ready to make my device available for testing and such as well. No rush, just don't forget me if/when you decide to push forward. I'll make time for this kind of project

200 voices of poly across all engines, under 30 seconds boot time?

I say: NOW WE'RE COOKING WITH GAS! Great stuff here, must admit. That's the REAL Kronos 2 right there.

If they get this thing to dual boot to windows - I'd do this and use the Kronos as my Windows development environment! hahaha, that would be incredibly awesome!