Click on the Ketron Button

http://www.bck.co.uk/


Bill

I'll believe it when I see it !

I am suprise it only has 200 internal styles. That's all??? Am I missing something? Tyros 2 and Korg P2X are offer more internal styles on their keyboard. I am not quite understand.

coca

There's only so much you can do with an audio loop...

This is a problem I'm not sure we all appreciate, yet. Let's just assume that these 'live audio' styles ARE jaw-droppingly wonderful....

There is NO WAY to include the RAM to be able to have ALL the styles be this good. So how is the arranger going to sound when, after playing a Beguine (for instance) that sounds like a real band, you next get a request for a dance style that DOESN'T have a set of audio loops to go along with it??

Consistency is the hallmark of a great arranger. Everything sounds great. Not just a few styles. And no matter how great the regular arranger sounds, if these 'live audio' styles are all they are cracked up to be, they are going to make the rest of the styles, in comparison, sound lackluster. Or there is little point in this VERY expensive technology.

What do you think?

Wait and see what Roland are doing.

There is NO WAY to include the RAM to be able to have ALL the styles be this good.


Ketron will not be relying on RAM, all of the styles will stream directly from the hard drive.

Currently, my SD1 holds all the added styles I have accumulated,in a master folder, on the hard drive. can access & play styles from that folder,without loading or copying in to RAM or registration.

Dan O'

Dan

To Ketron engineers, it is and always has been, "THE SOUND" , as Don M would put it . The sound is what Ketron cares about the most.

Hearing some one play a guitar riff and capturing that with audio / wav. file , can not be duplicated with numbers (midi) ... in their opinion, that is ...

Dan O

Well, if the samples are streaming constantly from the HD, they'd better do their level best to shock mount and isolate it..! I would be extremely uncomfortable relying on something working that hard, encased in a keyboard that is being pounded on, and wobbling around on an X-Stand

And I still have yet to hear an explanation of how anyone using audio for anything other than drums and percussion are going to get the range of chord types and inversions and extensions that current MIDI-based arrangers can. A loop library for even ONE style, ONE instrument will be HUGE if it has as many chord type/ext/inv's as a normal arranger can do without breaking a sweat.

All well and good if you are willing to let the machine dictate to YOU what chords you can and can't play, but me, if I want a C7#9 over an F bass, I don't want to hear a C over F, thank you very much!

For hiphop and loop based musics, this sounds like the bees knees. But if you want to play ANY standard, and have all the chords correct, not just dumbed down Easy-Play versions, this is going to take some doing...

Does ANYONE have a clue how the Audya is going to deal with this?

I saw what Roland were doing at NAMM... didn't impress me (or very many others either) at all...

All due respect diki, no you didn't...there wasn't even a brochure...then....
I don't reckon you have any concept of what is coming..and that's all I can say at the moment..
Not trying to be machiavellan, but I am sworn to secrecy.

I saw their latest WS, usually the source (after a couple of years) of their arranger technology. A HUGE yawn, all around.

Now, MAYBE they've got a new arranger in the pipe, but I don't see it being released with technology that the WS line doesn't even have, yet.

You don't have to spill any secrets, but let me make a guess...

It will have audio recording. Maybe multi-track. (FantomX)
It will have some of those half-a$$ed SA wannabe sounds I heard on the Atelier.
It'll probably drop the useful SRX slots for the new ARX ones (that they only have two cards out for, yet). BAD MISTAKE.
It will drop all the GS sounds and make compatibility with older styles impossible (did that on the E80 already).
It will have MP3 playback, with tempo and pitch transpose. (Yawn)
It will play synced audio sliced rhythm parts (like the FantomX does).
It will be more expensive than the G70, but add NO real arranger features, just WS ports.
It will probably drop whatever is my favorite current feature (probably the FC-7 support)

Sound likely?

Maybe some of you DO want multi-track audio recording, MP3 playback, synced audio slices (although you can't load them fast enough for a request), etc.. Me, I've GOT two WSs already. And a DAW or two. I want an ARRANGER, not some hybrid WS/arranger. They never add the WS features that WOULD be useful to an arranger player... the arpeggiator and synth controls and contemporary soundset (still under arranger-like controls). They just add the recording bullsh*t that we've already got covered.

Roland.... PROVE ME WRONG!

Make the best arranger since the G1000. Forget ANYTHING to do with DAWs and WSs. G1000 needed none of that to be successful. It just was their best arranger OS for the arranger PLAYER.

i have no intention of stealing the topic, but about G1000... that was what 10 years ago?! of course roland needs to add more than only arranger features into their next arranger. this is general trend, and it happens i agree with it (arranger+workstation approach).

but hopefully they will also refined and improve the arranger section. especially the arranger section.

Diki ...

Does ANYONE have a clue how the Audya is going to deal with this?

I do not know all the answers, but I can explain the key on how ketron currently sync's wav. files on the sd1.
The Key is naming a wav file, the same name as the midi file or style...For instance, I have a santana midi file, named "santa"..
I also have a wav file on the hard drive called "santa". When I call up the midi file, the wav file attaches to the midi file.

So we have a midi file playing from one master folder and a wav file streaming from a different folder at the same time. The wav file will respond to transposition or tempo change, without losing musical integrity.

Use midi as a map to manipulate multiple audio loops at the same time ='s AUDYA.
At least that's my understanding.

Dan O'

Sorry, Dan, but these are two totally different things. A single stereo file of an entire song is no problem for even inexpensive arrangers like S900, etc..

But what the Audya touts is the use of audio files for each pattern, each part, each chord. In other words, instead of the arranger playing a guitar sound by MIDI to get the guitar part, it is played using an audio recording of a REAL guitarist, playing that chord. And not in Song (or SMF) mode, but in arranger mode.

Here's your problem... Let's say you've got a four bar picking pattern. That's one file. For one chord, in one key (you can only transpose audio a couple of half steps, and it gets all chipmunk-y on you) you need maybe six or more recordings of the SAME chord (let's just say C Maj, for example) in different keys. Now that's just ONE type of chord. Now multiply that by all that ONE chord's extensions (Csus, C no3rd, C6, Cmaj7, C7, C11, C7b9, C9, C7#9 and so on) and that's just the Maj extensions..! Now add all the Cmin, Cdim, Caug, ChalfDim, and ALL their extensions, you are talking dozens upon dozens of audio files, just for that guitar pattern. Now multiply THAT by the four variations, and the fills, and we start to look at hundreds of audio patterns, just for ONE style. And THAT'S just ONE of the guitar Parts. Multiply THAT by the Bass Parts, the Pianos, the strings, the drums and percussion (although they don't need chord extensions so they are the simplest to produce), the horns, etc., etc..

Pretty scary....

Not to even get into the problem of how well these patterns join into each other, as you move from one chord to another in the middle of this four bar loop.

Now multiply that by the 200 or more styles that you expect in a modern arranger...

I have a sneaky feeling that this product is going to be hard pressed to live up to it's hype. As I said earlier, if the 'live audio' styles are THAT good, you are going to want them ALL to be that good. And that, my friends, is a herculean task for both the arranger to pull off technically, and the company to record.

Time will tell...

I have got a suspicion that they will not have the "audio" for all the instruments, maybe 4 or 5. Say guitars (one of each, ie acoustic+electric+distorted), brass (trumpet and sax), bass, drums and maybe strings..If they were to have audio samples for the entire keyboard it would require a hard drive of at least 200 gig.

My thoughts exactly, but even just a few 'real' instruments can tax even a fairly powerful computer when you need audio streaming lots and lots of small files, all the time. Not to mention the 'beating on the HD' issue when it's encased in a keyboard...

And, of course, there's the final nail, for me... I LIKE tweaking my styles. I like the ability to try a guitar part with a nylon instead of a steel string. I like to tweak the attack and decay, depending on tempo. I like to try the style with brushes instead of rock drums. I like to try tablas instead of congas. I like to be able to make the style MINE...

If a style is NEARLY perfect for a song, but a kick in the guitar part is wrong, or a drum accent is mistimed, I can edit a MIDI style.

You can't do this with an audio file. Everybody that plays the audio based style parts are going to sound IDENTICAL to everyone else's, whether you care to tweak or not...

No, the way to get 'realism' out of a style is not to use audio loops. It's the easy way, the quick way, the preset way. But realism can be achieved with MIDI also. It just involves massively more ROM for the sounds (to get close to GIGA quality), and more elaborate NTT's and rule-based behavior for the patterns. Imagine a cross between Yamaha's SA voices (for note transitions and intervalic rules), Korg's and Roland's Guitar Mode (for more accurate NTT's and instrument specific behaviors), and MUCH larger ROM for multi-velocity sounds.

These are capable of achieving close to audio loop quality, but with all the advantages that MIDI allows us (larger chord choices, editability, smooth transitions).

The GIGA-based arrangers (Wersi and MS) are showing the way (but MS seems lost in loop-land, for now), but seem unable to finish the job, relying on the user to select (and purchase) the GIGA sounds, and do all the work tweaking the styles, working out your own NTTs, etc..

If Ketron want to find a use for (hopefully) robust disk streaming, perhaps just GIGA sized MIDI instruments and more work on advanced NTT's might be a better choice.

It would certainly be a more FLEXIBLE choice...

Diki,

If Audya will manipulate audio loops with in a style,in real time (via midi), than it will , in my opinion, sound better.

If you just want to 'sound better', try getting some really high quality karaoke tracks... Real players on everything! Really GOOD players on everything! That's the best you'll ever sound... Of course, you have NO CONTROL at all, but what the hey! It's the logical end result of this line of technology, anyway.

I was just trying to point out that 'sounding better' at the cost of losing flexibility and chord choices is a poor decision, at least for me. Others may not CARE they can't play any chord they want to. This may be JUST what they need...

I'm guessing they are using some sort of stretching algorith. That way you can adjust both key and tempo of an audio loop without getting chipmunky And if you break up chords for the audio instruments into the four major groups (major, minor, dominant and diminished) you only need four loops per instrument per style. You can then use the MIDI instruments in the arrangement to add those missing notes. In a band it is rare that every member play the full voicing of a chord.

Well that is my theory, without knowing anything about how Ketron actually has aproached this

Techincally if you stream the loops from a drive, this is actually programatically less stress to solve than a full scale MIDI autocomp engine. You can also add a couple extra bass loops where the bass player just stays on one note, that way you can handle inversions by just transposing that.

If you use an SSD drive you woun't have to worry about wobbling about with the board

Doc-Z

[This message has been edited by doc-z (edited 02-22-2008).]

Nice explanation Doc Z...


Diki...

I am glad that you have a grip on all this arranger stuff, cause I sure don't.

Hey Bill ....

Thanks for posting this information. Excellent thread ! Dan O'

------------------
http://www.esnips.com/web/Songs

Hi Diki,
I'm wondering if they're going to do something along the lines of what they did with the drums on the sd1 . A mix of midi & audio.
The "Live Drums" sound to me ( & I'm probably wrong) like a split audio loop.
A similar effect to what I get on the PA800 when I split a drum audio loop & turn it into a groove. Instead of the loop playing back as recorded, I can alter the notes of the midi track that is generated, and change the rhythm of the loop.

The Live Drum's use midi notes, & are supplemented by a normal midi drum track to add additional drum sounds.

Wondering if maybe they're doing something similar with some of the other instruments ie a mix of split audio loop & midi.


Personally I luv the Ketron sounds, & I'm looking forward to seeing what they've done with the Audya, but for me the style editing/creation functions are lacking on my the SD1+ & unless they've included some additional functions like event list editing of a style, I don't think I'd be tempted into upgrading, even if it sounds brilliant.

200 onboard styles, if that's really all there are sort of doesn't sound quite enough..


best wishes
Rikki

[QUOTE]Originally posted by Diki:
[B]

Stretching is for tempo... I've never said they are likely to have a problem with that. But pitch shifting is another thing altogether. Two, three semitones tops without it getting REALLY weird. Remember, we're after realism, here. No point in it if it comes out even more weird than a MIDI track.

You can't derive an augmented chord from any other (amongst many others), so that's one more basic chord type they are going to have to record. And, if you are trying for better realism than current implementations (Yamaha's Mega voices are pretty good already for guitar parts), no, you can't take a recording of a C strum, and add two more notes of MIDI guitar for a C69. That would be an eight string guitar! And the wrong voicing. Totally unrealistic...!

That's the whole problem, in a nutshell. You want REAL, you have to have a recording of it. Anybody use Steinberg's Virtual Guitar? Notice how FEW chord choices you have? And not that many different styles? And it is STILL over a GB in size. And it's several years old technology.

But, don't worry... sooner or later (later, the way things have been going) the Audya will be released. And then you can have IT to be disappointed with, rather than my predictions.

I'm not saying you add more midi guitar notes to make up a C69 on the guitar track. I'm saying for a C69 let the guitar play a C strum, let the string or pad play a D5 and the bass play C. Want a Cmaj#11? ok, let the guitar again play a C, toss in a Bminor on the piano, then you've got your Cmaj#11. Offcourse this is just a theory on how you could do this in the simplest form, you could offcourse use more chords than just the four basic chords. Virtual Guitarist is a little bit of a different beast. The first version used a bunch of loops, the new one uses a blend of loops, samples and midi. But this thing is a Guitar simulator (Much like RealGuitar) And you have quite a lot more chord choices in the newest version of them both. If you want to check out how stretching algorithms can really improve pitch shifting, check out Melodyne and SampleTank 2. I use Melodyne all the time for fixing pitch problems in voice recordings, but I also use it a lot to generate vocal harmonies, transposing recorded instrument parts it's great! SampleTank has a stretching algorithm that transposes samples really good! Both of these algorithm in both applications transpose the audio without changing the lenght of the sample, just the pitch is adjusted. Normal pitchshifting involves running the sample faster or slower to get the desired pitch result, hence the chipmunk effect. This way however does not give you the chipmunk effect, you have to try it out for yourself.

Check out http://www.nativeinstruments.com for Sampletank
and http://www.celemony.com for Melodyne.

If you could incorporate this type of approach into an arranger, you will get more realism than you ever thought was possible. But again this is just my theory on how Ketron could achieve the result they are claiming to produce, I have no affiliation, nor any information regarding what their solution is all about.

Doc-Z

Nothing is coming. Roland have closed the poject in January.

Tin