Entries by Stephan Paukner

The Boss RC-2 Loop Station is a compact pedal for recording loop phrases with an electric guitar. The pedal has two bigger brothers: The RC-20XL Phrase Recorder and the spacious RC-50 Loop Station. The RC-2 is ideal for practicing and composing. It has 33 different drum patterns available and can store 11 phrases with a sum of at most 16 minutes. You can tap in the tempo of the drum patterns and even the speed of recorded phrases without changing their pitch! The pedal has an AUX IN to record phrases from an external playback device. This makes it possible to e.g. slow down guitar solos you want to learn. In this case you can use the AUTO recording mode that doesn’t start recording before an audio signal is received. Recordings can be overdubbed arbitrary times, and there’s an undo and redo function for the most recent overdub. Playback can be switched from loop to one-shot, making it possible to trigger recorded samples during a performance.

Boss must have put much brain power into the usability design of that pedal to make it possible for musicians to have a reasonable recording workflow available by just using two knobs, a single button and a single footswitch. I don’t want to go into detail here, but rather show what a typical workflow could look like (without an external footswitch):

Now I remember why I rarely go to the movies.

POTC (the first part) was a rather amusing film. But POTC2 was a lame sequel (like they usually are), and POTC3 was trash (like they usually are, see Terminator III). POTC3 takes long and is boring. It is impossible to follow the complex story line. Some scenes have awesome effects, like the world going upside down at sunset. But then there are only weird scenes, like the slapstick marriage during a saber battle. Not recommended, not even as a popcorn movie.

Five murder mystery TV shows with blondes in the lead and starting with “C”: C.S.I., Criminal Intent, The Closer, Close To Home, Cold Case.

No wonder I’m confused. And I always thought that TV is only for people with an IQ < 3.

I used P. Prinz’ toolbox to compute a 2D dual atom on a separable lattice and on a quincunx lattice. It shows that even if both lattices have same redundancy, the dual might be better localized on the quincunx lattice:

In this example, the norm difference of the “quincunx dual” to the original is about 0.1, and that between the “separable dual” and the original is almost 1.0. To achieve more similarity on the separable lattice, one has to take a higher redundancy.

I’m still not sure what the term “separability” actually expresses for 4D lattices: There exist position lattices (on the image space), frequency lattices (on the FT space), and TF-lattices (for both dimensions). If the atom is separable into its two dimensions, then one can compute the two 1D duals and obtain the 2D dual from their tensor product. If one takes two quincunxes for computing those two 1D duals, does this correspond to a non-separable 4D PF-lattice? In the case of 1D signals, a separable TF-lattice means that there’s always the same set of frequencies taken at every position. A quincunx lattice means that there are two different sets of frequencies chosen alternately while walking along the position points. Maybe the same interpretation works for the 2D case. I haven’t analyzed Prinz’ routine yet, but I think that he takes a rectangular (separable) frequency lattice and shifts it alternately by a half of the lattice-point distance while walking along the position spots. Should one try to use a quincunx position lattice? Well, I’ll have to experminent with that.

HGFei asked me to contact Søndergaard, because that topic of non-separability is currently not only “hot” for myself.

I’ve finally improved my script that computes a dilated and rotated 2D Gaussian with respect to low memory consumption. It applies a dilation and rotation matrix to the 2D domain. I published it under the GNU GPLv2, so if you find any improvements, you are forced to share them.

I still have the impression that it is rather slow, but I currently don’t know how to do it better. Of course, it has to be evaluated on 7×7=49 tiles (where tile=interval×interval) instead of just 7 intervals, what means that the computing time will increase significantly even for the non-dilated and non-rotated case. However, it works.

Here’s a comparison of computing time and accuracy:

And some plots using different parameters:

In my notebook, I have a 60GB disk, 27+14=41GB considered for the whole Linux system and 11.5GB left for various personal files. In my backup PC, I have a 112GB /home partition (LVM on RAID-5) and 12GB left, another 12GB are available on the /usr partition and 8.6GB in /opt, and as everything is on LVM, it might be resizable. So there’s currently no real need for additional disk space. The only drawback is that I can’t hold all music files on my notebook and that the available space is splitted into two partitions. But I can live with that, I just have to consider that some things will be in a mounted subdirectory.

When space will get small, I’ll buy a 160GB notebook-disk. And my backup PC will be a new one with two 500GB SATA-disks on RAID-1. I won’t take a kind of commercial network storage array, as these aren’t capable of rsync or the like.

I thought about taking RAID-5 again instead of RAID-1. Three 320GB disks are cheaper than two with 500GB, but data redundancy decreases from 2 to 1.5. Only one disk may fail in both cases, no matter if you’ve got two disks at RAID-1 or three disks at RAID-5. In addition, the probability that two disks fail is three(!) times higher when there are three disks as if there were only two. For me, RAID-5 is therefore just a strategy for expansion, not for starting freshly.

Yesterday I finally wrote a routine which computes a non-separable (dilated and rotated) 2D Gaussian. It takes parameters for horizontal and vertical dilation of the 2D Gaussian which is then rotated by a corresponding parameter. If it is neither dilated nor rotated, it is numerically identical to the pure 2D Gaussian. It has N. Kaiblinger’s gaussnk.m as background when it is computed on an area which is 7 times as large as the desired signal length but then wrapped (summed up) into the smaller region. I won’t publish it yet, as it can be done better; currently I separated the tasks of computing and wrapping, but they could be done at once.

With this I could finally reproduce a graphic from the paper “2D-GA Based on 1D Algorithms”: It shows (1) a non-separable 2D atom with relatively prime height and width, (2) the atom mapped to vector shape, (3) the dual of that vector with regard to some combined time and frequency steps, and (4) the dual vector reshaped to an image.

The question remains about how to finally do GA using these two 2D atoms. The only ways I found so far was either building the 1D Gabor system (of the atom shaped as vector) or computing a sampled STFT by that 1D vector; this works because modulations stay modulations. However, in the case of separable 2D atoms, this can be done in another way, as I’ll show in another article.