hello, im no pd expert at all... im looking for a possibilty to get data from a timecodevinyl (like finalscratch, serrato, miss pinky) into pd and send it over midi or osc to vvvv.

i know that theres miss pinky for max/msp. but is there an opensource equivalent for pd somewhere on the net?

thanks for your help...

I don't know the answer just yet, but I'm in the process of trying to get my Final Scratch v1.0 to work in my Ubuntu Dapper OS. I was going to try to port the Final Scratch software but if that's unsuccessful the next thought I had was piping the record output through a better soundcard (the given "scratchamp" has REALLY bad soundcard chips), recording the timecode and trying to write something myself in PD. My guess is that getting propper audio emulation of turntablist sounds might take some work but simple playback tracking shouldn't be too hard to figure out.

If you were looking to get a set of vinyl, any music store that carries stanton products should be able to order a couple for you. They are sold separately from the entire software/soundcard/vinyl package.

My understanding of this is that it isn't timecode. It's just a recording of a sawtooth wave. In theory it's quite doable in Pd and without a fancy soundcard. You just need to get the pitch (using something like [fiddle~]) and the direction (slope of the input), so that you know whether the turntable is moving forwards or backwards. From those you have the rotational velocity, which you can then use to index a sample playback table.

i was just thinking, that you could probably do a hack job with an old house or techno record. as long as you can find a record with a steady beat and no breakdowns, then you coudl use bonk to sync the beat. obviosuly this would not allow scratching, but might work for beat mixing.

hey elektromeier

I also want to trigger video using timecode records.
I made a pd patch using fiddle~ today to test its capability against the final scratch signal.
at high speed the detection is relatively accurate.
when you lower the pitch you get totally useless results unfortunately.

you mentioned ms pinky. I just realized that they offer an SDK.
hmm. that might be a good option.
otherwise I have to press my own timecode vinyl

So now I discovered vst~ and hooked it up with the PinkyPluggo VST which works great - as long as I don't bring up the editor. As soon as I try to manipulate the UI of the plugin, it crashes pd. Which is annoying because I need not only to read out some parameters but also to load an audio file into PinkyPluggo. Maybe I can exploit a hack they did to support startup file loading. Anyway this would be a very inflexible solution.

Anyone more familiar with the vst~ external?

@obiwannabe said:

My understanding of this is that it isn't timecode. It's just a recording of a sawtooth wave. In theory it's quite doable in Pd and without a fancy soundcard. You just need to get the pitch (using something like [fiddle~]) and the direction (slope of the input), so that you know whether the turntable is moving forwards or backwards. From those you have the rotational velocity, which you can then use to index a sample playback table.

Please see my post about my My Homemade MIDI Turntable

Thank you.

Casainho
http://www.Casainho.net

Nice turntable mod, I don't understand something, is the shaft encoder on the motor spindle? Is the table a belt or rim drive?

Would be nice to hear some scratching of samples on this when you
finish it. nice work casainho

andy

@obiwannabe said:

I don't understand something, is the shaft encoder on the motor spindle? Is the table a belt or rim drive?

Well, I don't have the encoder yet, I just saw his datasheet. I already look at the old vinyl player and I hope to put the encoder directly on the shaft.

Maybe I will have a problem, because the encoder is very cheap, about 15 euros, so, It just give 128 positions per turn

I remember to see a good encoder, for 6000 positions per turn, for about 250 euros!!

I am looking for simple and cheap hardware to detect the position of the shaft.

Other thing, is, should I send position, velocity and acceleration to Puredata?? or just position?? - since the Microcontroler can do all that work...

The hardware is very simple: Old Vinyl turntable + Encoder + Atmel ATmega16 + MIDI -> USB.

. Encoder, can be expensive but starting at 15 euros.
. Atmel ATmega16, the Microcontroler, for 4 euros or less.
. General MIDI <-> USB for about 25 euros.

@obiwannabe said:

Would be nice to hear some scratching of samples on this when you
finish it.

Oh, I don't scratch, I am constructing my toys - So I will need feed-back... I will give the project for free and try to make It simples with a manual hopping that people construct and use It, and I will have feed-back then

Okay, I'm hoping that y'all have heard of the great work of Ms. Pinky? She's developed a system which rivals that employed by final scratch, while staying decidedly DIY/opensourcey. Her records are encoded with something akin to "Trellis Coded Modulation" (wikipedia it) and the decoding was originally done in a MAX/MSP patch! Apparently her system allows both relative and absolute playback of digital files (tables, control values, whatever)

If what Obi says about the stanton vinyl is true (sawtooth waves) it sounds like it would be pretty east to index an array with one. What I never understood about the "timecode vinyl" (misconception?) was that timecode only goes one direction. Imagine if it didn't though! I could scratch my ProTools sessions!

What about a record that was somehow encoded with a bitstream index which began at 0 and just incremented at 44.1k? Or a frequency sweep that started very low and swept higher and higher. We could track the pitch with [Fiddle~] or whatever, then use that as an index.

Any other Ideas?

there are two issues that fight to such an extent they are exclusive afaics, you can have good absolute time encoding so that you can drop anywhere into the timecode and quickly know the position within something like an hour of audio, or you can can have good relative encoding such that you can scratch the actual audio back and forth.

Old frame based timecode such as used on tape, I can say from experience is no use to scratch, it takes too long for the decoder to lock to the absolute position and the divisions between time markers is too big. You can get the decoder to interpolate (or freewheel) between marks but without enough marks per second of code it won't work for scratching.

Sounds like Pinkys system is using FSK, which uses two sine tones like a modem, the trellis encoding is an ECC layer that makes this more reliable and a little more efficient, but basically this is how you encode a binary number into vinyl to tape (if the groove actually had 1s and 0s it would break the stylus very quickly)

If we started at 0 and had one second of codes at 44.1Kz then
at 33 rpm (0.55555 revolutions per second) at 12 inches (30cm) diameter
gives us 1.885m/s (at the best case outside of the disc), which is
0.0000427m for each timecode chunk. In that chunk we would have to
encode at least 24 bits to have an integer time index over about
an hour, making the distance between timecode chunks 1.7e-6m.

Too small!

That's why we have CD technology not not vinyl, if there were a
way to squeeze that much digital info into a vinyl groove Phillips
or Sony would have looked at it

If we use an entirely relative approach, where instead of symbols we encode a constantly
variable analogue signal (what vinyl is best for) and use that to increment or decrement a
counter in the decoder much better results are possible.

Some timecode is reversible, that's kind of the point when you work with tape, you have a buffer and a way to detect frame boundaries and read the frames in reverse.

> Any other Ideas?

Personally I believe the best way is not to encode on the vinyl at all, but to make a turntable
that very accurately measures the angular velocity of the wheel and the angular position
of the tone arm pickup. You can probably do this using bits hacked from a disk drive. It would need
occasional calibrating, but is the best all round solution for high resolution with absolute time. (think of the tone arm and the table rotation like the mantissa and exponent of a floating point number, they're both part of the same number with the tone arm counting the big divisions and the turntable wheel counting the smaller ones.)

If you're going to be building a whole new turntable-esqe in order to scratch samples, seems to me it'd be much more profitable to just re-work the concept of the turntable interface. Why does the solution have to mimic the actual movements of current techniques when one could develop a more purpose-built gestural controller which could be much more able?

I understand that these systems are meant to translate current DJ skills into a much more open-ended paradigm, but to me it seems as though further steps are necessary to push the art.

I think the draw of this quest for most is the aesthetic. Without the turntable and some vinyl, it's kind of pointless. Abandoning the analog element, there are many many options for 'digital scratching' but they can only replicate the Physics of the steel and the Vinyl. If your going to go to the trouble of developing a physical computing environment from scratch, why bother with emulation?

The beauty of Ms. Pinkys system is that the interface is familiar and analog... accessible.

Not everyone wants to play a bunch [boxes~] with an arduino on a laptop.

P.S.

One other interesting project is th D'Groove haptic feedback turntable:
http://www.timothywisdom.com/science/dgroove/dgroove.php

When I said gestural controller I don't mean a laptop with keyboard and mouse control. Likewise, it doesn't have to be digital at every level. What I meant was that one could design an interface that would expand the possibilities of sample scratching without the need for a redundant vinyl element.

On a basic level, one could simply use a modified version of obiwannabe's solution, wherein one wouldn't need to have a vinyl if only the tonearm and shaft positions are significant to the sound.

Taking it a step further, perhaps one could create an interface made up of two orange-sized spheres on shafts that the operator could grab and rotate with a great quickness. Say the left-hand is a coarse-scope adjustment to the album position and the right-hand is fine-scope. You could use the right-hand sphere to scratch normally, and the left-hand one to navigate between samples. when neither sphere is rotating the album would play at normal speed. From here you could add controls for speed changing and any other manipulations that would be useful, and hopefully it would be a more compact and capable system in the end.

nestor: > I think the draw of this quest for most is the aesthetic.

Agree. And a solution that leaves the turntable intact in its function for playing normal records would
be greatly superior to one that renders it unusable for its original purpose. That's why i think Pinkys
method has a lot to be said for it.

mnemonic: > ..two orange-sized spheres on shafts that the operator could grab and rotate with a great quickness. Say the left-hand is a coarse-scope adjustment to the album position and the right-hand is fine-scope.

I like it, very futuristic! One of the creatures in the jazz band at Mos Isely space port would have one of these.

Maybe casainhos method with the shaft encoder is the best all round compromise, because DJs could play existing discs and have the possibility to drop in FX (on that audio) based on the time index.

Ah, of course I never considered that the ability to play existing vinyl would figure into things (doyyyyy)

Slight tangent, hope this is ok...

Has anyone come across a project for making a cdj? A good cdj costs around £1000 whereas it seems to just contain a cd drive, a rotating drive, around 10 buttons and a laptop-esqu touch pad.

I'd love to know of anyone making something with a cd-rom drive, a kaoss pad and a stripped down motherboard...

(I'm very new to pd but feeling very excited by it!)

Tom

I don't know about the CDJ idea, there are some soft dj apps around, such as MIXXX, that might be suitable if you hooked up a controller to them.

But on the original topic, there has recently been some headway! The XWAX libraries have been wrapped into a PD external: http://karhumusic.sesser.at/vinylcontrol/vinylcontrol.html

This will allow PD to read Serato and Traktor "timecoded" (yeah I know it's not actually timecode, but they're marked as such - call it auditory timecode) vinyls.