SimpleSample

[Tim Kleinert]

Most basic sampling circuit: no bells&whistles, no tuning -but stable due to time-stamping technology.

Gate inputs for sampling and sample playback.
Audio input and output.
(highlighted modules)

See test modules for implementation.

Edit: This is an old approach that has been superceded by a vastly better technique. See below in the thread. Don't use this one.

---

SimpleSample TK.pch2
Description:	most basic (non-drifting) sampling circuit	Download (listen)
Filename:	SimpleSample TK.pch2
Filesize:	1.92 KB
Downloaded:	1627 Time(s)

---

[iPassenger]

Hi tim,

Spent an hour or so looking at this tonight as I am determined to get my head around it and I think I am starting to get it now (it was the two delays that had me confused but i get that now)

Anyway a few of the modules contained within this patch didn't seem to be doing anything (attached with modules removed), as I tried taking them out to help my understanding and it had no effect (on the patch or my understanding hehe..)...

Would you mind explaining why I might need them, I presume it has something to do with time stamping!? (which I don't really get but have heard you mention before)

Cheers

Ross

---

simplesampleMod.pch2
Description:		Download (listen)
Filename:	simplesampleMod.pch2
Filesize:	1.74 KB
Downloaded:	1450 Time(s)

---

[Tim Kleinert]

I haven't looked at your patch because you don't need time-stamping at all anymore. That approach has luckily gone out of the window.

The time-stamping method worked by syncing the circular readout pointer to the delay line via a LOUD max headroom spike of a few samples, in order to prevent drifting. This brought with it various drawbacks: the inputted audio had to be reduced by 12dB in order to not be mistaken for the sync spike itself (thus scaling down the audio quality from 16 to 14 bits). Also, accidentally playing back the spike wasn't healthy for ears and monitors.

I have since patched a newer method that works by exactly counting out the number of samples contained in a 2.7s loop and thus provides a non-drifting readout base pointer. (Yes, this was a bitch to get right. ) Result: no drift, no bit reduction, no nastily loud audio spike in the loop that has to be navigated around, and more economic too.

I extracted the basic sampling subcircuit and made it into a building block patch with documentation and demo section. See below.

cheers,
tim

---

SimpleSample2 TK.pch2
Description:	Improved basic volatile sampling subcircuit. This is the one to go for.	Download (listen)
Filename:	SimpleSample2 TK.pch2
Filesize:	3.21 KB
Downloaded:	1563 Time(s)

---

[iPassenger]

Arggh!!! hehe just kidding, so I get what you've done here but now the block doesn't retrigger with note on like the old one.

How do I reset the counter back to zero, to restart the sample? I get the feeling it should be easy but I am not sure where to cut into the patch.

EDIT: Also in theory it should be possible to build samplers with a shorter sample time, but increasing poly mem etc.. by truncating the sample counter to a value appropriate the sample time.

[Tim Kleinert]

[iPassenger]

Many thanks for this tim, had loads of fun trying to mod it.

I was trying to make it so I could move the start point of the sample.

Presumably the obvious answer would be to control the start phase of the OSC but I couldn't work out how to do that, so i experimented with affacting the start of the count by adding or subtracting a value.. but this just made it go weird.. but it was fun anyway.

Managed to make it go backwards though by inverting the saw wave, also managed to make it slow down whilst holding a key, almost like a scrub effect but unfortunately the counter circuit wasn't affected so the scrub zone would get smaller as you slowed it down.

[Tim Kleinert]

Hi Ross,

Cool you're having fun with it.

The circuit cannot handle position values above 64 (or below 0 for that matter). So, if you want to change the starting point by adding an offset, you have to make sure that the values wrap around 64 back to zero again. Otherwise it will start to act strangely.

A forwards/backwards-scrub-functionality amounts to having a direction control for the playback sawtooth driver. The "Dir" input in the ShapeLFO does exactly that, but unfortunately doesn't clock at audiorate. So one has to go the DIY route. I made a patch (see attachment below) that does this, and has sample start point control as well.

However, you will notice that the keyboard tracking is bad and the direction control is coarse. This is due to the fact that the values needed to drive a DIY readout pointer at such low frequencies (2.7 seconds middle base) are so minute that they truly challenge the resolution of the whole system.

Of course, there is a method to circumvent this too (as I did in the PolyGranuSyn X algorithm). But we are getting more and more into, erm, weird territory.

cheers,
tim

---

SimpleSample3 TK.pch2
Description:		Download (listen)
Filename:	SimpleSample3 TK.pch2
Filesize:	3.28 KB
Downloaded:	1556 Time(s)

---

[iPassenger]

Tim,

Just realised I hadn't said thank you for this so, thank you, it's very clever.

[Akum420]

Hi,

I can't trigger the loop (by pressing a key) to the start point when I chose 1.0s instead of 2.7s.

[Tim Kleinert]

[Akum420]

ok thank,

I understand better. I don't have all knowledge to do this now but I understand the main idea. I'll watch it closely soon. ++

[Roland Kuit]

this is fun. thanks tim!

---

simplesamplemod tk rk.pch2
Description:	tweak record/trig and level knob	Download (listen)
Filename:	simplesamplemod tk rk.pch2
Filesize:	4.22 KB
Downloaded:	1591 Time(s)

---

[Tim Kleinert]

Oooh, looks interesting. Gotta check this out when I've hooked the G2 back up. Thanks Roland.

[Roland Kuit]

sleeples night tim:)
8 variations

---

simplesamplemod_tk_rk_B.pch2
Description:	totaly punked:)	Download (listen)
Filename:	simplesamplemod_tk_rk_B.pch2
Filesize:	4.81 KB
Downloaded:	1515 Time(s)

---

[Tim Kleinert]

sounds like a drum machine made by aliens amazing.