The aliasing thread

[Improv]

When the subject becomes old and tired.. nothing new left to say.... people invariably start with the he-said/she-said quoting. Look at any other thread on a forum that supports the quote feature.... A few pages and it's information, after that it's an ego battle.

get over it damnit... for goodness sakes.

what he said..

what he said...

What they all said. Time to end this pettyness. The actions of more than one dev here do not look good to the user base. And users tend to remember things like this!

[peejunk]

Wow, some of you people are being waaay too political here. Sure they ALL got carried away a bit, some more, some less. There was a lot of noisy comm-links here (was obvious to me they seriously misunderstood eachother, and they all misunderstood a lot of William's posts, which in turn agitated Rene, George and they agitated him and it goes on..), some out-of-context quotes, some bad (or better yet, stupid) moves -- but why should one take his business elsewhere? How can you judge people like that over something silly as this? I hope you're just kidding, you don't think YOU are so much better than any of these guys?

Personaly, I sometimes like the sound of aliasing, sometimes don't care. Most of the time I hate clinically clean production with too much stereo and polish, which is likely to turn me away from music whereas a dusty lofi sound could lure me into enjoying something I'd consider merely average otherwyse.

However, the most often effect of comparisons like Oskari's, George's and Simon V's that I've seen around on musician forums are a lot of wrong, rushed conclusions by the uninitiated about the overall quality of some products, and then it spreads, and spreads.. And in that regard, William was actually right.

Just my 2c.

[Majken]

I found this thread very interesting until the flaming. However i feel i should write a few last words since i was active in the first half of it.

The test is as far as i can see it 100% accurate for what it does, which is measures aliasing in pitched sinewaves using different samplers. However William had a very good point, a sampler that does good on sinewaves at 15kHz doesn't necessarily have to do good on guitars at 1kHz. It's perhaps likely that it will but that doesn't mean that it actually does. Then again that's not what the test analyzes.

As i see it William hasn't attacked anyone or foulmouthed any products. However he might be a bit clumsy with which words to use. About SFZ sounding filtered, that was something i said which William quoted. And it does sound filtered, compared to W-station's aliased sound which may or may not be a good thing.

And George didn't do anything wrong with publishing the tests. They're very interesting but you have to take them for what they are. You can't judge a sampler from those test, only a specific part of the sampler. It's like measuring sparkplugs in a car, even though they may be better in one car that doesn't mean the entire car is better.

William shouldn't have spoken his mind so fast and George should have waited with the aliasing tests of w-station until the updated version was relesed. However one thing was trigged by the other i suppose.

Zebra is great, SFZ is great, W-station is great so is Rhino and DiscoDSP kicks ass. It's awesome to get to see discussions about improving plugin quality openly on a forum like this. Especially since this ends up giving the big companies a hard time when the smaller companies compete with them.

I really hope you can get along after all this and continue helping eachother with more open discussions. Thanks everyone!

/Majken

[bmanic]

Wow, some of you people are being waaay too political here. Sure they ALL got carried away a bit, some more, some less. There was a lot of noisy comm-links here (was obvious to me they seriously misunderstood eachother, and they all misunderstood a lot of William's posts, which in turn agitated Rene, George and they agitated him and it goes on..)*etc*

What HE said.. except the bit about aliasing .

- bManic

[george]

I've just added aliasing concept for easy understanding, including a brief description.



PDF of the full research is also available.

Well... nothing else to add except big thanks all for your feedback

[tony tony chopper]

George, stop the bad faith, and either stop hiding rendered results, OR do a better CPU vs quality vs RAM usage comparison.

Otherwise, should I just enable FL's sinc interpolator realtime? You'll be able to run 1, or maybe 2 voices realtime. Then what, will you consider this as realtime? It'd be really stupid to do this just for a dumb test page.

[Sicklecell666]

I've just added aliasing concept for easy understanding, including a brief description.



PDF of the full research is also available.

Well... nothing else to add except big thanks all for your feedback

AFAIC it just demonstartes talent in an image editor..

The results you try & cram down everyone's throat's are directly connected to your public's opinion.

What that means in english is that if you can design a GUI (and nice ones, I admit), you can doctor a demonstration graph..

Previous demonstrations of your kind & coutereous personality doesn't lend much credability to your argument.

[WilliamK]

George, we just released Wusikstation V1.0.6 with Hermite interpolation by default. Please, remove our current image as it is not valid anymore.

We updated the Lite Demo version too. So you can run a test with it now. I tested here and the result was "Average quality" acording to your guide-lines.

Thank you, WilliamK

[george]

George, stop the bad faith, and either stop hiding rendered results, OR do a better CPU vs quality vs RAM usage comparison.

There is no bad faith. The subject is about real time and if your product doesn't excel in that specific area, I'm not going to hide it, although fruity offline rendering is top notch, so why bother?

Stop hiding? If you _read_ the text, there are completely visible links pointing to non-realtime engines of _all_ software tested. And also, RAM vs CPU usage is not the aim of the research.

In case you people feel worried about this, work on it instead blaming me.

[valley]

What that means in english is that if you can design a GUI (and nice ones, I admit), you can doctor a demonstration graph..

But since the methodology is public, just like any valid scientific study, it is third party verifiable.

If you don't trust the results, step up to the plate and to the tests yourself. Either you can prove to yourself that the results are accurate, or you can prove to us that George is lying.

[george]

George, we just released Wusikstation V1.0.6 with Hermite interpolation by default. Please, remove our current image as it is not valid anymore.

William, I will be downloading the new demo in a few minutes and will make the update today, thanks.

[Sicklecell666]

If you don't trust the results, step up to the plate and to the tests yourself. Either you can prove to yourself that the results are accurate, or you can prove to us that George is lying.

Obviously verifiable, & I'd step up to the plate if I actually gave a shit about this hair-splitting crap.

But I don't recall me saying he's a liar. I recall questioning his moral integrity & his glorious propensity & appetitte for slamming other Devs..

[george]

AFAIC it just demonstartes talent in an image editor..

I didn't know cut, copy, paste and drawing rectangles would require talent

What that means in english is that if you can design a GUI (and nice ones, I admit), you can doctor a demonstration graph..

No, GUI design is a work that is highly vinculated with the coder and the human interface. Without one of these things, making some fancy graphics doesn't automaticly makes a good product.

Previous demonstrations of your kind & coutereous personality doesn't lend much credability to your argument.

So you can't discuss this topic properly and start harrasing me instead? If you can't keep this thing civil I'd ask you to move on.

[René]

Regarding the "rendered mode" matter, I do think that it would be really cool to include the results of samplers which can deliver a better interpolation on rendering. That would for instance include HighLife and sfz 'perfect' modes, and the SF2 player. I don't recall any other, but there might be more. Perhaps it's just required to specify those modes.

BTW, is the FLStudio SF2 player demo included into FLStudio supposed to play only some notes? I tried to run the test on it, and this is what I get:

http://www.rgcaudio.com/images/FLStudioSF2demo.gif

It renders well without the 'HQ' mode turned on here, but I might be doing something wrong.

Also, I've finally found the time to download and test Sytrus. Great synth. Here's the test results, playing the last notes on it:

http://www.rgcaudio.com/images/sytrus1.gif

I run the test on 64x oversampling. If it would loaded samples, it might be listed amongst the best on George's list.

-René

[WilliamK]

http://www.rgcaudio.com/images/FLStudioSF2demo.gif
It renders well without the 'HQ' mode turned on here, but I might be doing something wrong.

Remember that I did the same mistake with another program and someone jumped on me like a wolf. (no one never said sorry to me here so far)
It looks like you are rendering one octave up, could that be it?

But yes, I also think that the pages show always show the best quality possible. Even if is not real-time. It would be more fair. What you guys think?

Wk