czerro
19th June 2014, 11:49
I can't survive without you guys, and as much as I try and learn, I can't wrap my head around some of the transforms and masks. But, they work! Amazing.
I primarily do PC LP footage. Autoadjust, Seesaw, mvtools(2), nnedi implementation, anything tritical has ever made, etc, and it's mostly a trial-and-error experience but it's fantastic to even have the options.
As much as I love these things, the documentation is often bad. The coder doesn't do enough due diligence to document dependencies or dependencies on updated libraries or update them thereafter. "if you get this, it will work." Rather than acknowledging that everything else in the chain is going to change as well, but the user has no idea what those dependencies are and likely has conflicting dependencies in a complex chain. So you can get some goofy errors for no reason, simply because you aren't 'in the know' and understand the crosstalk you are picking up on is only relevant to MT, special x64, or special x86 build, dependencies for other features that are now considered 'main', experimental libraries, or some such without mention.
I know you guys are brilliant, and a lot of people have hopped onto avisynth+, and it has good compatablility, but there are alot of people that are scared to do so, and the plugin/script documentation/threads does nothing to alleviate these fears, with all the problems of updating ones workflow I highlighted above.
BTW: I hopped on Avisynth+ today because I accidently overwrote my AutoAdjust 1.90 with 2.20...to discover it didnt work with avisynth 2.5.8 and had to bite the bullet on Avisynth+ because the creator no longer has the binaries :/. I can return to AVS2.5.8 easily and none of my plugins are broken. Aweseome. It's BUILT into the Avisynth+ installer, but I would guess that most people wouldn't investigate it that far. And as far as I can tell everything works, notably mvtools(2). Performance seems to be slightly better. But it takes me back to the original problem of adoption. I understand that avisynth+ will be better, and you can't expect people to come on board while at the same time trying to support implementation of the old guard. What you can do this time around is demand better documentation into a wiki that is controlled and updated concurrently with dependencies so that people know what they are working with without having to bother a maintainer with a question.
I primarily do PC LP footage. Autoadjust, Seesaw, mvtools(2), nnedi implementation, anything tritical has ever made, etc, and it's mostly a trial-and-error experience but it's fantastic to even have the options.
As much as I love these things, the documentation is often bad. The coder doesn't do enough due diligence to document dependencies or dependencies on updated libraries or update them thereafter. "if you get this, it will work." Rather than acknowledging that everything else in the chain is going to change as well, but the user has no idea what those dependencies are and likely has conflicting dependencies in a complex chain. So you can get some goofy errors for no reason, simply because you aren't 'in the know' and understand the crosstalk you are picking up on is only relevant to MT, special x64, or special x86 build, dependencies for other features that are now considered 'main', experimental libraries, or some such without mention.
I know you guys are brilliant, and a lot of people have hopped onto avisynth+, and it has good compatablility, but there are alot of people that are scared to do so, and the plugin/script documentation/threads does nothing to alleviate these fears, with all the problems of updating ones workflow I highlighted above.
BTW: I hopped on Avisynth+ today because I accidently overwrote my AutoAdjust 1.90 with 2.20...to discover it didnt work with avisynth 2.5.8 and had to bite the bullet on Avisynth+ because the creator no longer has the binaries :/. I can return to AVS2.5.8 easily and none of my plugins are broken. Aweseome. It's BUILT into the Avisynth+ installer, but I would guess that most people wouldn't investigate it that far. And as far as I can tell everything works, notably mvtools(2). Performance seems to be slightly better. But it takes me back to the original problem of adoption. I understand that avisynth+ will be better, and you can't expect people to come on board while at the same time trying to support implementation of the old guard. What you can do this time around is demand better documentation into a wiki that is controlled and updated concurrently with dependencies so that people know what they are working with without having to bother a maintainer with a question.