pouët.net

Go to bottom

The art of connecting oldschool gear to your video projector chain at parties

category: parties [glöplog]
Well, everything depends on how you do the doubling, no? Repeating is ugly, proper vertical filtering maybe isn't so far from how it'd actually look on a CRT at some distance.
added on the 2015-09-15 00:46:37 by Sesse Sesse
What kind of vertical filtering do you mean?
added on the 2015-09-15 17:20:33 by yzi yzi
Say, Lanczos interpolation? (Okay, maybe something less prone to ringing, but you get the idea.)

Signal-theoretically, this is relatively straightforward. Your signal had (from the get-go) a certain distribution of vertical frequencies. Now, if you up-convert by adding zeros (black lines) in-between each sample, this has the effect of taking the existing signal spectrum and replicating it in an extra copy the spectrum on top (aliasing). Filtering (or blurring, if you want) has the effect of removing that extra set of frequencies.

Of course, what's theoretically optimal in one sense isn't necessarily what everyone finds aesthetically most pleasing. (I haven't actually made any detailed comparisons for oldschool content.) But one would have to wonder why, if inserting black between samples is the best way to upconvert, why we don't do it horizontally, too.
added on the 2015-09-15 19:30:25 by Sesse Sesse
Bullshit. Signal-theoretically? Optimal? I think it's a weird point of view to talk about vertical frequencies and zero insertion as some kind of legitimization.

I think it is culture-theoretically optimal to try to make prods look the way they were seen by the audience for which they were made for.
added on the 2015-09-15 21:36:04 by yzi yzi
Thank you for your thoughtful reply; I will certainly bow to your “oh, I don't have any arguments, but I like it this way” reasoning.
added on the 2015-09-15 22:05:44 by Sesse Sesse
I'd love to see yzi make a good scaler without any of the signal theory "bullshit" :)

As for normal filtering looking similar to CRT as a distance, that may be true, but most people weren't far enough away, especially not monitors.
added on the 2015-09-15 22:24:31 by absence absence
absence: Might very well be true. Even so, though, it's not a given that the perfect emulation of the CRT look is to just black out N out of every M lines.
added on the 2015-09-15 22:27:31 by Sesse Sesse
That is 100 % certainly not a perfect, or even good, emulation of how a CRT looks ;)
added on the 2015-09-15 22:29:06 by absence absence
I did not say signal theory is bullshit, or how to make a scaler. I said it is bullshit to use signal theory as an excuse for not trying to watch prods in the form their original audience saw them.

Maybe next you'll present a signal-theoretical reasoning for why it's OK to watch prods on Youtube? Maybe something that involves Z transforms and complex numbers?
added on the 2015-09-15 22:40:58 by yzi yzi
And I also do not say that Framemeister's scanlines makes the picture look like a CRT. It doesn't. But it is a step closer.
added on the 2015-09-15 22:44:56 by yzi yzi
spotting this topic for the very first time.. it would be very interesting for some seminar i guess :)
added on the 2015-09-15 22:48:00 by gentleman gentleman
yzi: Sorry, it wasn't clear from your post that what you consider bullshit is to not watch prods in their original form. I disagree, but my reasoning for watching demos on Youtube is entirely non-theoretical and a matter of convenience. :)
added on the 2015-09-15 23:00:38 by absence absence
No.

You may choose to watch Youtube and not try to get the picture close to the original, because you just don't care enough and it's more convenient. And that would be an ok thing to say.

But if you say you don't try to get scanlines right or just watch Youtube, because of signal theory blah blah, THAT is bullshit.
added on the 2015-09-15 23:18:33 by yzi yzi

login

Go to top