Samplematic: Unbelievable

[sonicfire]

The most important area of distinction you can make from the "other" auto-sampling gizmos is the price. Samplerobot is VERY expensive the last time i checked. Samplit is pretty cheap and you can get a free version from buying VSampler i believe.


Anyways, your project looks very interesting! Good luck with it.

Also, a slight idea : How about offering automation? I bet thats something no one has ever done yet. Consider this : Your program should be able to sort of, have like a sequencer only for midi CCs. It should be able to offer "layers". That is, one layer of CC07 to automate the cutoff while it is sampling, and another layer to automate pitch lfo frequency, etc., all while sampling a single note. Then, do different types of automation, and be able to layer them into ONE single program. Now, THAT would be a killer auto-sampling app.

About that modulation stuff - nice idea, will note that down for next free update!

@snooky: No worries, it won't be expensive!!

[R.A.W.]

please make this freeware?

I'd love an app like this

http://www.device.arcanedevice.com/?misc

it's called synthcatcher - but i guess it's a bit simpler than this new tool

[sonicfire]

I'm surprised that there aren't more replies on this?

[sonicfire]

hell, i want to see some emoootionnns! some anticipations!!

[grymmjack]

Be prepared for this little "Instrument" to record your external Hardware (or VST's via a virtual MIDI cable) fully automatic into instant playable Instruments!. It records note by note from any Audio Input, correctly applies crossfade or forward/backward-looping and then the final result can be saved as a Soundfont!

I'm currently in the phase of eliminating some little glitches here and there. Hopefully next week i can present you a first Demo of it!

Cheerio!

wonderful, if it works

automatic looping is a tricky one, how would you know this information? also i don't see any options to detect silence, like in samplit, which is one of it's best features -- that way the lower notes can sustain longer and the higher ones can be cut automatically -- lower notes on acoustic instruments like piano will sustain much much longer than the higher ones, as an example.

looks cool though, you've got my attention. need some beta testers?

[grymmjack]

No i meant setting it up the way you want in a Host (rout out of ext. synth into audio in, add fx) - there you go!

So much! Which of these do automatic SF2 creation and automatic looping?

EDIT: How do you sample external Hardware in EXSC?

set source destination as hardware converter it's beautiful in fact, the simplest implementation of this kind of thing i've seen and great mapper right in front. only thing it lacks is the auto-trim silence stuff/auto-detect silence that samplit has.

i'm interested to try your 'auto-loop' algo myself, ESC doesn't have that one either.

[grymmjack]

The most important area of distinction you can make from the "other" auto-sampling gizmos is the price. Samplerobot is VERY expensive the last time i checked. Samplit is pretty cheap and you can get a free version from buying VSampler i believe.

samplit had many bugs in the demo version and i've reported them only to have been ignored by the company.

shitty support + shitty bugs = no thanks on samplit.

otherwise i loved the software though. it was dead easy to use, had nice diskspace saving features (auto detect silence/auto-trim, etc), and for $99 is a decent enough price -- i would buy it if the bugs got fixed and the company acknowledged my report

[grymmjack]

mhh, no it isn't there, but that doesn't bother me much, since i have an app at hand that does exactly this

download the latest version, it's free on his website. put in your reg code and then download it. it's under support link i believe.

[grymmjack]

I'd love an app like this

me too. another thing i noticed the "stop" button -- does this work while the sampling is going on then? ESC has no 'abort' feature as silly as it is to omit it, it's been omitted.

cant tell you how many times i've accidentally forgotten to change the patch on my hardware after setting up a very longggggggggg convert

ah well, i'll email wlodimiers now that i think of it.

[zvon]

Hi Sonicfire,

Looking forward to try this!

[grymmjack]

The most important area of distinction you can make from the "other" auto-sampling gizmos is the price. Samplerobot is VERY expensive the last time i checked. Samplit is pretty cheap and you can get a free version from buying VSampler i believe.


Anyways, your project looks very interesting! Good luck with it.

Also, a slight idea : How about offering automation? I bet thats something no one has ever done yet. Consider this : Your program should be able to sort of, have like a sequencer only for midi CCs. It should be able to offer "layers". That is, one layer of CC07 to automate the cutoff while it is sampling, and another layer to automate pitch lfo frequency, etc., all while sampling a single note. Then, do different types of automation, and be able to layer them into ONE single program. Now, THAT would be a killer auto-sampling app.

About that modulation stuff - nice idea, will note that dwn for next free update!

@snooky: No worries, it won't be expensive!!

another idea; is entire bank sampling and automatic patch naming using cakewalk instrument definitions, midnam xml format, cubase mixer map format, etc. that would be a killer feature in my book, that along with the ability to dump entire banks unattended after setting up the batch. that's what i was hoping to achieve with samplit since it supported many programs at once but i heard that it was very difficult to use like that and is limited to only 32 open project files at a time (samplit project files also contain program change info -- no bank change is supported though).

if you want to do something unique, that would be a killer feature imo. it seems easy to do it too -- instrument definitions or midnam or mixermap extraction of patches, program numbers, and bank change messages isn't a difficult programming task to achieve. i had considered using a program like autoIT to do the extraction/integration myself but it's not worth the effort for i've only got 8 or 9 banks to do and while 2000+ programs are there, it's going to be a long uphill battle anyway it seems. only automated naming via preconfigured lists, and auto-sensing silence and auto-trim would really help me out.

thinking now, even if you had some 'script' for naming conventions that mapped to program names where the user had to do the footwork, that would help too. but then you'd have to add batch ranges for program change start and end and so on.

i did something like this myself but it doesn't have the sophistication yours does;



of course there is no audio interaction and it requires midiox and windows to work. so it's definitely intriguing me to see where you take your project! how much will it cost or will it be free or what?

looking again at yours though, there are no velocity zone support. ESC has this, so if you charge for yours it would at minimum have to atleast compete in this area i feel. try grabbing the updated version of ESC if you own it, the hardware converter is beautifully implemented with a easy mapper, easy velocity setup, and because it's integrated with ESC it can output to any format supported by that program. hard to beat, to be honest, as it lacks only a few features and i'm sure in time they will be added (hopefully).

also i want to say i didn't post my screeny to compete, my little hack is nothing compared to your nice project, not trying to steal thunder, trying to show how much i support this kind of program because i've even tried to make one on my own with the free visual basic

[sonicfire]

@grymmjack: so much replies! thank you!
I try to answer your questions:

Stop Button: Sure, it does stop the whole process immediately, so if you made a mistake and dont want to continue recording, just press stop and there you go

Betatesters: At the moment i have some betatesters, but i'm sure i'll need some more, soon!

Looping: For now, it simply uses three types of looping, the first two ones are crossfade-loops, that rather start at 25% or 50% offset from the recorded note/sample. The third one uses forward/backward looping starting at 50% offset!
(And of course you can disable the looping, if you dont want it at all)

The benefit of this type of looping is, it will always loop and work. Of course for some samples you just have to manually edit/change the loop points, but that depends on the source material...

Velocity Layers: Yup, coming for next free update i think (or maybe even for v1). But, for now i want to start simple to have one working first version and not overdoing it. (When i program, my code can get messy very fast ). I also think that e.g. 16 velocitylayers would be enough? Is there a limit within e.g. soundfonts on how much velolayers they can handle?

Recording whole banks/diff. presets at once:Excellent idea! I've noted that. Seems reasonable!

Trimming,silence detection: Already on my todo-list. Most important: cut silence at beginning of samples. But the idea to trim them afterwards to cut silence at the END of samples is a very nice idea!!

Instrument definitions:Not sure about this one. Never used / worked with such. Do you have some more information about these and what they (should) look like at hand?

I will also try to implement some kind of information field, i do like that very much! (Where you can see in advance how long the recording process would be, how much samples, disk space needed and so on)

Thank you!

[grymmjack]

I will also try to implement some kind of information field, i do like that very much! (Where you can see in advance how long the recording process would be, how much samples, disk space needed and so on)

Thank you!

thanks for replying to my many posts. PM coming with samples of instrument definitions (cakewalk format - like INI), midnam XML format, and cubase mixmap format.. all ones i hand made (except CW which i modified some sources) for my xp60 and its many banks

[Jeremy_NSL]

Looping: For now, it simply uses three types of looping, the first two ones are crossfade-loops, that rather start at 25% or 50% offset from the recorded note/sample. The third one uses forward/backward looping starting at 50% offset!
(And of course you can disable the looping, if you dont want it at all)

But to my knowledge SF2 and WAV both don't support forward/backward (pingpong) looping... How do you plan to work that? You could destructively append a reversed loop at the end of the wav, but that will increase filesize alot.

[sonicfire]

Looping: For now, it simply uses three types of looping, the first two ones are crossfade-loops, that rather start at 25% or 50% offset from the recorded note/sample. The third one uses forward/backward looping starting at 50% offset!
(And of course you can disable the looping, if you dont want it at all)

But to my knowledge SF2 and WAV both don't support forward/backward (pingpong) looping... How do you plan to work that? You could destructively append a reversed loop at the end of the wav, but that will increase filesize alot.

Sure, it will increase the file size but that's the only method i can think of! And since todays harddisks have enough space, that shouldn't be a problem! (and of course, you dont have to use this loop mode)