untune's Forum Posts

  • Hey all, sorry I hadn't received any alerts from this - I moved on from Construct a while ago but i've uploaded the file here if anyone needs it http://s000.tinyupload.com/index.php?file_id=10860822105436691304

    Cheers

  • lucid

    Hi Lucid, I've been reading and I'm just been curious about something and how it relates to C2. I have been looking into the lip sync via papagayo and Spriter's support for it - I'm just curious how it might be practically applied. Does each line of dialogue that you want to sync have to be created within the Spriter editor and then exported as individual animations, like you would have a walk cycle, idle, jumping etc... would you be calling say, 'anim_dialogue_134' etc and potentially having hundreds or thousands of individual lines explicitly created like that?

    Does each 'speaking' animation work explicitly, i.e. would your character be stood potentially motionless, idle etc, could you have your character speaking random lines while running, jumping, shooting etc and have them in sync - or would you have to say, export a separate head object and somehow bind them together in C2?

    Sorry this is all speculation for the moment, like I said I'm just curious as to how it would be implemented and how well it's supported currently.

    Cheers!

  • Much as I would love to try this out, it just gives me a black screen (it makes the layer appear black). W10, NVIDIA GTX 750 Ti.

    What else is on your layer? Newest drivers etc? I tested it right before I packaged it up... I think I'm on a gtx 680

  • Hi all, apologies for the delay in getting this up here but a mystery illness since last Thursday has kept me at bay - I managed to get this working a week ago but I'm only just getting round to sorting it out. I'll try to sort a screenshot at some point. A few notes:

    This is adapted from code originally done and kindly shared by Martins Upitis http://devlog-martinsh.blogspot.fi/2013/05/image-imperfections-and-film-grain-post.html with the exception that I have removed the element that crates coloured noise. I did this because I didn't personally feel the extra processing is necessary, and film grain should be kept subtle. I've always preferrred mono noise/grain. It wouldn't be difficult to add this back in, however - this is the first time I've ever looked at a shader and I was just keeping things simple.

    I also borrowed a little bit of code from the implementation here http://www.html5gamedevs.com/topic/8393-phaser-filterkit-vignette-and-filmgrain/ which was ported from the same source but adapted slightly. I can't remember exactly what the additions were but I think there was a slightly different function for randomising/rotating some coordinates - I left it in as I was experimenting with the various methods. There were some issues while testing where the noise would create repeating distracting patterns and I thought this might have been the cause - it likely wasn't, so it might just be adding unnecessary processing.

    Another slight change I made was altering the luminence coefficients to Rec 709 - the original is still commented out in the code, as are a few other things, if you want to experiment. Hope you find it useful.

    Default settings for params are 0.05 for grain amount (0 to 0.10 is a logical range for intensity), 2.5 for grain size (1.5 to 4.0 and beyond if you like, bigger would probably suit a nightvision effect) and luminance stays at 1.0, but can be tweaked - experiment.

    Also many thanks to Gigatron for the effort put into converting despite me having already finished it, apologies for not responding sooner - see above And to chrisbrobs

    Link to zip containing fx and xml file

    If anyone has any optimisations or ideas, please feel free to improve and share with the community.

    Cheers

  • I dont want to disturb you and chrisbrobs

    Here is a quick port of this shader from your source ...

    Tell me if it's ok .. i have not added grain amount a,d other parameters just converted it.

    Here is a demo like i love : http://gigatron3k.free.fr/html5/C2/FX/fgrain/

    Nicely done gigatron! I did finally get this working a few nights ago but I've not been well the last couple of days and haven't had the chance to clean up the code etc

  • untune

    I had a look at your 'FilmGrain_v1.1' file...way too complicated for my brain.

    The best I can come up with is this basic noise/image mix effect-

    I will play around with the sums to see if i can improve it.

    Its a start?.

    https://dl.dropboxusercontent.com/u/22173473/filmgrainrequest.rar

    Thanks, I'll have a look at this tomorrow

    P.S

    ' I don't have a lot of time at the moment and life is getting to me'

    Its only a shader

    Haha yes I know, it's just been a very tough week! Getting better though I've got a much better understanding of the code now so it should be easier to convert as long as a few things are possible with C2...

  • Try Construct 3

    Develop games in your browser. Powerful, performant & highly capable.

    Try Now Construct 3 users don't see these ads

  • Working on it... Not had any fedback on here as to the way c2 works so I'm just working blind at the moment. I don't have a lot of time at the moment and life is getting to me, but I managed to get it working in shadertoy at the weekend (took til 4am though) so now I understand it better it should be (hopefully) easier to port to c2. Will keep you posted.

  • Hi all,

    I've been away from all this for a while and alhough I have ideas swimming around, I'm usually too busy to work on much. I revisited an old project a few days ago and have been playing with various effects to give things a certain look - I had left the 'noise' effect on the layout and although it works nicely I was wondering if there might be a way to implement a more accurate 'film grain' style effect.

    There is already a great shader available here http://devlog-martinsh.blogspot.fi/2013/05/image-imperfections-and-film-grain-post.html (or for the code, see here https://dl.dropboxusercontent.com/u/11542084/FilmGrain_v1.1 ) that I have been trying to convert over to c2 but naturally, nothing is working as expected. I'm at work currently and can't access the code I had been modifying, but I wanted to check I wasn't just hitting c2 limitations as I was getting constant errors - seemingly with the initial declaration of variables, and the c2 effect template was a bit confusing (I also found that there is an additional parameter in the stock c2 effects such as noise that doesn't appear in the effect template XML file - to do with opacity I think). I keep having it crash complaining about a syntax error with 'mediump' - there are a lot of vaiables in there and I'm not sure how it handles things.

    Does every float need to be explicitly assigned a precision? I tried precision mediump float at the beginning but no luck. Is there a way to get the game's X and Y resolution, as it seems to be relevant in this shader? Can C2 handle multiple functions in the shader outside of the main function call? Forgive me if anything is obvious but I just want to avoid wasting time if I won't be able to get it working!

    EDIT I also removed the code relating to coloured noise as I wanted to keep it monochrome, I thought it may speed things up slightly

    Cheers

  • I'm working on this right now:

    Przemek32767 That looks really cool. How are you importing files from your editor?

  • Aurel I'd love to know a bit more about the technical side too

    Funnily enough I was pondering an idea for a sci-fi 2d platformer inspired by Ulysses 31 that captured that outlandish 1980's cartoon vibe but it gradually evolved into a top-down Wipeout-esque future racing idea. Lo and behold, beaten to it

    I did do a little tech demo though to test some ideas with the gamepad - one thing I'd love to know is, have you created your structure using the built-in box2d physics system, or did you code your own movement/collision/physics for greater control?

    Cheers

  • BPM = Beats Per Minute

    I'd be interested in this too as I would like to create a dynamic music/sound system... the webaudio API does apparently have access to very accurate timing information which would be necessary for this kind of set up. Whether or not it is implemented in Construct is a different matter... I'm going to hazard a guess that as Aphrodite has mentioned, it is tied to the overall delta time system since there doesn't appear to be any way to access any other timing info. Which is a shame because sound/music is often overlooked and the webaudio API gives a lot of control that could have a lot of potential if we could get at it!

  • I just realised what I did - I quickly rendered one at 400% and then ran 200% twice on the original and got the same result without realising that I was effectively just halving it twice, ie the same as 400%. It's been a long day, I need to go home and lie down...

  • Fimbul Yes you're right, I knew something didn't seem right but I tested it quickly with different numbers and it sounded correct, typical!

    thehen By that logic what you should really do is run it once at 500% and then once more at 200% to half it again like Fimbul says. But still that's a long way to be pushing a sample so ensure you use sounds with the highest sample rate you can manage, to preserve as much quality as you can. I think Audition is a bit confusing in the way it presents things, there's probably a much simpler pitch shifting tool out there - Audacity is free but the effects are pretty simple, that's probably a better option

  • Audition may be a bit too in depth for what you need to do, as it seems to give you a bit more control than you really need. The stretch is the time domain and the pitch shift is the frequency domain, and it is generally more suited to adjusting speech from what I can tell. But it seems that as long as you keep 'Lock Stretch and Pitch Shift' ticked then it should keep the correct behaviour as long as you change the 'Stretch' percentage. So 100% is regular, 50% is x2 playback etc... I'm not sure off the top of my head why the 800% limit is there but it's safe to say that you don't want to be shifting tihngs that far generally. But if you want to hit 1000%, run it twice on the same sample - 500% the first time and 500% the second time should get you the same result

  • Sorry to bump this, but will this feature be added at any point? Since WAV PCM is supported across browsers (with the exception of IE, which is hardly surprising) and given that it's important to maintain the integrity of all frequencies when appying effects, it seems a bit of a disadvantage not to have the option of at least one lossless sound format available.

    Cheers