Resampling (statistics) Puns

I asked for an introduction to Statistics, here

https://www.reddit.com/r/statistics/comments/b1o35c/need_recommendation_of_an_introduction_book/

By now, I still feel Chihara's is the most suitable:

Mathematical Statistics with Resampling and R

Laura M. Chihara、Tim C. Hesterberg / Wiley-Blackwell

But it is not widely used, and I can found few reviews for it. And I am not totally sure if learning much resampling at this level would be helpful. Also, I'd like a deeper exposition.

Did any one use it? Could you leave a comment on this book?

Contents and sample chapters can be found:

https://www.amazon.com/dp/111941654X/

I did some AM practice Qs earlier and one was to determine whether the following statement was correct or incorrect, and if incorrect, explain why:

“Resampling is one technique to deal with instability of the efficient frontier, but it has no basis in statistical methodology.”

I believed this statement to be incorrect, because resampling does address EF instability, and as it’s grounded in MVO and Monte Carlo simulation (large scale sensitivity analysis), those are statistical techniques.

The answer key states that the statement is correct, though, and I don’t know why resampling has no statistical basis. Can anyone help?

I am using the ks test to compare two different population distributions. However, the samples of each respective population are different sizes. To avoid sample size difference overpowering the test, I had the idea to repeatedly resample from each parent population to define the distributions, and then average the p-values across the series of tests.

Is this a valid approach? I know the Bonferroni correction is used for "repeated measures" tests. However, I'm uncertain whether this method of bootstrapping sample distributions counts as a "repeated measure", or if there is another correction I should be using.

I'm exporting my sounds in 192/24. I was using audition since now. I'm trying to force reaper to use 192khz on the project but it's going max to 96khz on my macbook (without additional audio inteface).

When I import a 192khz wave file it's saying "resampled". I want to be able to edit and export the file in 192khz, not 96khz (but the mac can't go to 192). So If I export it on 192 will it be downsampled to 96 and upsampled back to 192 or it will be exported straight to 192 (since the original file is in 192)?

When I'm editing sounds on my laptop I don't want to connect a dedicated audio interface.

Update: I've just bought a new audio interface. Obviously, it's the best practice. Thanks for the response. Cheers.

I have started doing it as it puts more load on my gpu and Not only do I get more fps but the visibility is more cleaner and crisp

Hi guys

Iv been trying for a while now to get a decent sound to come out of resampling but can't seem to do it. I'm trying to resample bass to get some dnb wobs and some nice mid range cuts to layer over my bass.

Does anyone have any advice for me or links to some decent tutorials?

Just tried it exactly the same performance if not even sometimes better. 2x might be the play. It upped my gpu ussage drastically.
1xoff 30% gpu usage
2x 50-60% gpu usage same performance maybe slightly better sometimes
Now 4x it went to 80-100% gpu usage NO hit on fps it stayed the same. Game looked crisp on 4x not going to lie.

I run a 3070-ti i9 9900k 16gb ram. I played around with the settings to up my gpu usage.

DOES ANYONE HAVE ACTUALL KNOWLEDGE ABOUT RESAMPLING?

I’m piecing together a techno drum groove by resampling and I’ve resampled the initial kick/cymbal rhythm twice, adding hats on the first and snares on the second. Now, when comparing the third iteration to the original sample, the original is quite a bit louder.

Is this something inherent in resampling? Does the SP (or similar samplers) have, say, an algorithm to reduce the volume of successive resamples so you don’t blow your speakers (I saw some people talking about this online relative to Ableton Live resampling)?

How do I upsample (11025 to 44100) and interpolate data linearly like Wavosaur?

https://preview.redd.it/oijr3pi9y1681.png?width=371&format=png&auto=webp&s=dee76fcb0540fc6760911ee49d37e44591159dbc

I haven't noticed this before, but for some reason when I resample something that has stereo width it gets condensed down to mono. I don't remember having this issue before as I mentioned.

To give an example, I took a small slice of a vocal line and put a huge reverb on it and resampled it, then reversed it so it would lead into the entrance of the vocal, and it's totally mono. I'm not sure why this is happening, I don't have anything on my master. Anyone else experience this?

Any advice would be greatly appreciated!

Edit: Alright I just tried the same thing, and it's stereo now. Perhaps it was just a glitch? Seems weird. Any thoughts?

I was reading the following article, when I encountered this:

>[...] There are other methods for controlling the probability of false results when doing multiple comparisons, including familywise error rate methods (e.g., Holland and Copenhaver, 1987), false discovery rate methods (e.g., Benjamini and Hochberg, 1995), resampling methods (jackknifing, bootstrapping— e.g., Efron, 1981), and permutation tests (i.e., exact tests—e.g., Gill, 2007).

https://dx.doi.org/10.3389/fpsyg.2015.00223 page 3 (box note).

I never heard of using the aforementioned procedures for controlling false results in the multiple comparisons situation. I do know that sometimes we do control the FDR, FWER (Bonferroni, Holm), and use specific tests, like Scheffé, Dunnet, etc.

How can these procedures control de false positive rate in multiple comparisons? If they do, of course.

Thanks

Hi all! I have a raster of land use land cover with a resolution of 100x100m, I need to « resample » it to 25x25m but without any resampling technique, juste divide the cells by 16 but keeping the original value for all of them, so the output will look identical, only the resolution changes. I need that for a further process done on this map. Does any one have a clue on how to do it ?? Thanks a lot.

Edit: typo in the title hehe

I get that Rhythm is mono, but can I play a melodic sample and then resample while playing the same sample in a different note? So like...record playing the 1, resample and add the 3, etc?

Similar to how you may create a triad from a monophonic synth with a looper?

Love the rhythm, but the no poly on midi or tracks gets me...

Like rendering takes so much longer so must be better but in what way?

Thank you :)

I have a curve that is the base of a river. I use the Sort SOP node to ensure the river flows in the correct direction.

However, when I resample the curve at any point it reverts the original point order.

Is there a way to "permanently" reverse the point order of a curve other than bundling a resample and sort node every time?

I watched Au5’s hypergrowl sound design tutorial for Serum which involved resampling a supersquare several times. Anyone know a way to achieve the same results in Vital? I don’t really understand the resampling concept

What are similarites and differences between them? Prons & cons?

I have been making tracks for more than a couple of years now, but was mainly was using drum samples and other kinds as one shots. All the other sounds (bass, pads, melodies) I build with things. The result is that I learned the absolute minimum about sample handling or creating complicated sounds from samples.

Are there any tutorials, plug-ins, etc, that you recommend to learn some advanced sampling/resampling techniques? I use Ableton live and what I call 'basic' manipulation of samples is chopping, warping, transposing...

Can the mpc one recreate the oldschool Lofi sound of lets say a mpc 60?

Can the Erica Synths Sample Drum do resampling? i.e. take a sample playing on one of the two channels and patch into the record input and record it on the other channel. I looked through the manual and didn't see any mention of it. If not, are there any other sampling modules that are 14HP or smaller than can do resampling?

Seriously y'all, it took me awhile to get to it because it doesn't really work the best as a static effect like many other plugins, but if you run your sounds on loop through grain delay and just twist knobs for a few minutes you'll get a wealth of sounds that can be chopped up and moreover, they'll often sound similar enough to the source material that they can fit in the same songs without much tweaking beyond maybe some additional compression.

Edit: To those asking for a tutorial, here's my advice: it works really well on your drum group or your basses. Personally I find it great for transition effects. Certain parameters are more helpful than others e.g. "spray" will just give you a bunch of noise if it's turned up. I usually leave spray at around 10-50 ms and mostly play with pitch, frequency, and feedback. You've also gotta recognize that a lot of the sounds you'll get are unusable, which is why it's good to record maybe 5-10 minutes of just jamming with a section on loop running through grain delay, then go back and listen through, chopping out the pieces that sound interesting. Hope this helps!

Is there an FFmpeg command to upsample (11025 to 44100) and interpolate data linearly like Wavosaur?

https://preview.redd.it/smd5bexay1681.png?width=371&format=png&auto=webp&s=2a9b05dae87c7fd8614fe2f5afb22fb8f9f2b661