Aliasing test on real music, can you spot it?

[Frankie.T]

We (almost) all know what aliasing sounds and looks like on sine waves, and it's pretty bad, but is so bad on real sounds and music?
Can we really spot it?

Oversampling isn't perfect, and comes with some issue, so is not like "i don't know, let's oversample and call it a day" because there is a trade off between pros and cons.

If you are interested i made a video around this topic you can watch here

If you are allergic with videos, i also made a blind test files you can download here

https://drive.google.com/file/d/1XSt11b ... sp=sharing

You find 4 tests inside. Each test comes with the dry signal, A and B, one is oversampled, the other one comes without OS.

The game is about to write the one who sounds better for you, if any, so each test you should write your fav (A or B) or just "Same" if you don't feel any difference.

I'm also curious to read your observations if you have

[QCGroove]

Love the fact that you’re doing this but the test at the top doesn’t let you select one for each song.

In the first example I feel like I can hear a difference. I like A better. In the last 3 I can’t tell a difference but I would select A on all three of those as well.

[Frankie.T]

Love the fact that you’re doing this but the test at the top doesn’t let you select one for each song.

In the first example I feel like I can hear a difference. I like A better. In the last 3 I can’t tell a difference but I would select A on all three of those as well.

Thanks for the heads up, i tried to change the options, now it should be possible to insert 4 votes, please try it now and let me know if it works

[TMaudio]

I think something is wrong with the FrontDaw test. Either user error, with some kind of phase/polarity issue or FrontDaw O/S is doing something funky. The files are radically different as they are currently.

For the rest as O/S'ed I choose:
Test 1 B
Test 2 B
Test 3 A
Test 4 B

Edited: added my Test 1 answer

[Frankie.T]

I think something is wrong with the FrontDaw test. Either user error, with some kind of phase/polarity issue or FrontDaw O/S is doing something funky. The files are radically different as they are currently.

For the rest as O/S'ed I choose:
Test 2 B
Test 3 A
Test 4 B

I've just checked the files and they are correct, so whatever you are listening is because of OS differencies

[QCGroove]

I think something is wrong with the FrontDaw test. Either user error, with some kind of phase/polarity issue or FrontDaw O/S is doing something funky. The files are radically different as they are currently.

For the rest as O/S'ed I choose:
Test 2 B
Test 3 A
Test 4 B

In my testing I have found that frontdaw uses modulation to change the audio in some way FWIW

[TMaudio]

I think something is wrong with the FrontDaw test. Either user error, with some kind of phase/polarity issue or FrontDaw O/S is doing something funky. The files are radically different as they are currently.

For the rest as O/S'ed I choose:
Test 2 B
Test 3 A
Test 4 B

I've just checked the files and they are correct, so whatever you are listening is because of OS differencies

Thanks, then I'll make my choice

[Ploki]

I've just checked the files and they are correct, so whatever you are listening is because of OS differencies

Oversampling can't shave nearly 5dB off of low-end. FrontDAW test is flawed, or FrontDAW is broken

[WackyZoundz]

Guys, this isn't an aliasing test, it's just trolling or incompetence. I knew the moment I listened to the records. All you have to do is to invert the polarity on either A or B of each "test" and then mix both together. What you get are clipping artifacts but not aliasing because aliasing does not have any harmonic relationship with the source material.

[El°HYM]

To be fair this could also be due to the implemented OS, which doesnt have to be perfect and in the end could cause more (other) problems than it actually should solve. Did the Tester provide the used OS x - factor e.g. 4x or 8x etc. that were used?

[Frankie.T]

I've just checked the files and they are correct, so whatever you are listening is because of OS differencies

Oversampling can't shave nearly 5dB off of low-end. FrontDAW test is flawed, or FrontDAW is broken

Aliasing reflect back distortion adding harmonics. In the zip file you find also the dry signal, if you want you can download FrontDAW demo, and try it yourself.

To be fair this could also be due to the implemented OS, which doesnt have to be perfect and in the end could cause more (other) problems than it actually should solve. Did the Tester provide the used OS x - factor e.g. 4x or 8x etc. that were used?

I don't remember the OS factor, but it's being said in the video

[El°HYM]

Which FrontDAW - Version did you use, as there was a Bug with the OS before 3.3 if I remember right.

[Frankie.T]

Which FrontDAW - Version did you use, as there was a Bug with the OS before 3.3 if I remember right.

I just checked the version and is 3.2

I'm curious now if available what's the kind of issue that is fixed between 3.2 and 3.3 if you know it, thanks

[El°HYM]

A small OS - Bug as it seems:

Liked the plugin and bought it, but there's a little bug in the oversampling: It seems to shift the HPF corner frequency up the higher you go. Screenshots of noise with no oversampling and 8x oversampling.

https://gearspace.com/board/showpost.ph ... stcount=23

[Frankie.T]

A small OS - Bug as it seems:

Liked the plugin and bought it, but there's a little bug in the oversampling: It seems to shift the HPF corner frequency up the higher you go. Screenshots of noise with no oversampling and 8x oversampling.

https://gearspace.com/board/showpost.ph ... stcount=23

Interesting, thanks for the info!