Agree, but question is: who complete it ? Because you must consider that the latest driver (even if faster) was never finished !
In end it means we will have somethings running fast, but with tons of feature less (for example the Altivec and the fullscreen support just to mention the first two)
As certainly i don't see LiveForIt to complete it for free (anytime soon) for now better still with the old driver ...
I don't think that any of the P96_PIP implementations has ever used it.
The old driver support it, but as i don't have any capable machine don't ask me how ...
Quote:
And why not include Joerg's P96_PIP improvements in the bounty instead of using an old and slower P96_PIP outpout ?
I think to have already explained you, because it means a considerable work and that work will cost a lot, see how mutch money he ask just for a few "almost easy" things ... Instead just adding a few fix in current driver will not cost (or just a few)
@K-L If i remember right, that code which in the repo (i.e. latest p96pip from joerg), wasn't finished. I.e. it didn't works in all the cases, and can behave badly in some cases.
Also i do not know for example if it support everything what support that old p96_pip from him. I mean, fullscreen, altivec, osd, whatever else. Will it faster on all machines and all cards ? I mean, will new implementation without altivec be faster that old one with altivec code on machines with altivec ? Is it indeed worth to replace and will be faster in all cases ?
But, in end of all i think about logical solution which will please almost everyone: we can add 2 p96_pips. Old one and new one. And just have 2 drivers in the "vo" output, like p96_pip(v1) and p96_pip(v2), and ppls can choice what they want (while, of course, there will be problem with those default settings, when users just want to run it and have maximum possible speed).
Isn't there any way to merge both versions in one with all the improvement and features?
If only one missing feature in the last joerg's code are fullscreen support, then adding fullscreen should be easy.
And if latest version from Joerg are the faster that previous with altivec, even without altivec, and there is no shortcomings, then probably worth to use this latest version, just with adding fullscreen support.
Is OSD working too with his versino ? And no flickering as in previous one ?
- Doesn't work without changes with the old MPlayer here.
Of course, some integration-adaptation may need, that no problem.
Quote:
- 16 bit RGB code included but disabled (define USR_RGB to enable it).
So what ? I mean, did you have any problems because of that, or will users have any problems because of that in compare with previous version of p96_pip ? I mean who care if some " 16 bit RGB" code disabled, if all works for end users faster than previous version. Or, in some cases it will just not works ?
If i remember correctly there were also some minor glitch with the audio, this happened when you put a video in pause (audio was not in perfect synch)
Quote:
Of course, some integration-adaptation may need, that no problem.
Yep however still to see if LiveForIt will be able to do the work
Quote:
So what ? I mean, did you have any problems because of that, or will users have any problems because of that in compare with previous version of p96_pip ? I mean who care if some " 16 bit RGB" code disabled, if all works for end users faster than previous version. Or, in some cases it will just not works ?
I just retried the Joerg version (just to verify how it works and so on) latest beta from Joerg was the beta 10 and you can find it here:
To summarize, aka to reach the status of complete having all the features of the old p96_pip, the Joerg version may needs those fix:
- Doesn't work without changes with the old MPlayer here. (need to test) - Altivec support missing. - Fullscreen missing - A minor glitch with the audio - 16 bit RGB code included but disabled (define USR_RGB to enable it).
Add to this the things we have planned to complete the old p96_pip
- Drag&drop
But we have also some good news: by using this latest Joerg driver we will be able to remove 3 subpoints:
- OSD bug - White stripes bug (probably it's fixed, but can't test as fullscreen doesn't work) - Menu
Infact those 3 things works already :)
To conclude, if we want to use this version instead of the old p96_pip, LiveForIt should verify that:
- Doesn't work without changes with the old MPlayer here. (need to test) - Altivec support missing. - Fullscreen missing (should work by pressing "F" and with double-click too) - Verify the problem with the audio (glitch) - 16 bit RGB code included but disabled (define USR_RGB to enable it). - Adding drag&drop
(if i don't forget any other eventual thing ! )
As you see now we have a lot of confusion, that was the results of the zero collaboration, different people did their own versions without any coordinations with the others, rewrote the wheel and never merge the code
- Doesn't work without changes with the old MPlayer here. (need to test)
That you even no need to mention. No one can hope that source of drivers will be just copy+paste from version to version. Of course some changes are need it, but its just adaptation changes as do anyone who put some code in one piece.
Quote:
- Altivec support missing.
As K-L say, version from Joerg without Altivec, are faster than version with Altivec. So, that no problem and also can be skiped.
Quote:
- Fullscreen missing (should work by pressing "F" and with double-click too)
Ok, that first problems.
Quote:
- Verify the problem with the audio (glitch)
So, verefy plz, is it here ? If so, then there is second moment to fix.
Quote:
- 16 bit RGB code included but disabled (define USR_RGB to enable it).
And i ask again : so what ? Do you see any problems because of 'missing 16 bit RGB code included but disabled" ? Are end users have any problems because of that ?
Quote:
- Adding drag&drop
Is it works with old version of p96_pip, and didn't with new one ?
As K-L say, version from Joerg without Altivec, are faster than version with Altivec. So, that no problem and also can be skiped.
Ok but Altivec by the way was one of the target of this bounty no ? Personally i never test any Altivec things because i don't have a capable hardware, so can't test
Quote:
So, verefy plz, is it here ? If so, then there is second moment to fix.
Here it happen mostly with MP4 videos, but seems a bit better with AVI ... we can test it better later
Quote:
And i ask again : so what ? Do you see any problems because of 'missing 16 bit RGB code included but disabled" ? Are end users have any problems because of that ?
Don't know what will be the implication of this, i simply reported what Joerg wrote , maybe is important, maybe not so
Quote:
Is it works with old version of p96_pip, and didn't with new one ?
Well it's the same here, drag&drop need to be implemented whatever version we decide to use ..
P.S. Drag&drop worked in the old p96_pip driver of afxgroup, but as usual it was never backported by Joerg ....
Ok but Altivec by the way was one of the target of this bounty no ? Personally i never test any Altivec things because i don't have a capable hardware, so can't test
K-L says that last version from Joerg without altivec, faster than old version with altivec. So, its make for final users such result: joerg's version will be faster in all condition, even without implemented altivec support. What mean we can drop old version and use new one, as it will be faster in all cases.
Quote:
Don't know what will be the implication of this, i simply reported what Joerg wrote , maybe is important, maybe not so
If everything works as expected for end users, and no video distortions, then doesn't matter.
Quote:
Well it's the same here, drag&drop need to be implemented whatever version we decide to use ..
P.S. Drag&drop worked in the old p96_pip driver of afxgroup, but as usual it was never backported by Joerg ....
There is only 2 versions of p96_pip. One are old one, that one which used in AfxGroup's version, in muimplayer, and dunno where else. If d&d work in AfxGroup's version, but didn't in muimplayer, then its my fault of integration, but not differences between driver.
There just 2 versions: old one from joerg+afxgroup, and new from joerg only.
Quote:
Here it happen mostly with MP4 videos, but seems a bit better with AVI ... we can test it better later
I.e. in muimplayer and in old afxgroup's mplayer all is fine, but with joerg's mplayer there is problems ?
It's all really interesting, but the speed of mplayer on different video format, depends on FFMPEG different versions included in different build, altivec will always be faster than none altivec, if the codecs are optimized for it.
As far as I know there is no altivec version of p96pip or p96, there is only normal assembler optimized conversion routine.
The bugs in the codecs also comes down to different versions of FFMPEG, if built from nightly build (work in progress) or stable version, there will always be bugs.
The last FFMPEG used in AfxGroup's version had that nasty AAC audio codec bug that effected the AltiVec optimization of the audio codec. For the most part the old FFMPEG was faster than the newer FFMPEG version that might got a bit bloated.
However sticking to an old FFMPEG is also silly because then you might not be able to play all audio codecs and video codecs, and I believe some codecs have gotten faster, even if some have gotten slower. And new video formats comes and so do improvements, and new bugs. That’s just how software development goes.
Quote:
There just 2 versions: old one from joerg+afxgroup, and new from joerg only.
Well not 100% true, more like 99% true, in order for p96pip to behave the same as comp, cgx_wpa, and so on, I have been changing the code, so that different video output works the same, thanks to a common back bone for all video output's on LiveForIt-Mplayer is cgx_common.cpp that I have heavily modified.
The work of reintegrating the menus will be done so the same code is going to work on all video outputs, I am not going to simply copy & past the code into all video outputs, that be too dirty for my taste and horror to maintain.
Same thing with drag and drop, it's go to go into the common video output backbone, or into it's own separate .cpp / .h files. I just hate having to dig through 1000 of lines of code to find what I'm looking for.
The integration of P96 was newer 100% completed, I did some basic work to try integrate it but, but the integration was not as complete as P96_pip.
Edited by LiveForIt on 2015/4/23 16:21:22 Edited by LiveForIt on 2015/4/23 16:22:25 Edited by LiveForIt on 2015/4/23 16:26:15 Edited by LiveForIt on 2015/4/23 16:28:09 Edited by LiveForIt on 2015/4/23 18:55:43
(NutsAboutAmiga)
Basilisk II for AmigaOS4 AmigaInputAnywhere Excalibur and other tools and apps.
p96_pip fixes Implementing menus in mplayer Drag and Drop Adding Screenshotdir and EXTPATTERN New feanor ffmpeg integration 64Bit fix Transferring of bounty will happens after beta-tests and before public release OpenSource
1300 Euro +10% bountysource fee +1,9% + 0,35 Euro per transaction -------------------------------------------------- 1455,05 Euro -------------------------------------------------- 1577,11 US Dollar Since there are always fluctuations, i would collect 1.600 USD
I will create a bounty at bountysource.com After we reach the goal, it takes about a month until payment.
News feed
OS4 Feedback forum
OS4Depot Feedback forum
Amigabounty forum
OpenAmiga forum
AmiCygnix forum
ABC shell forum
AmiKit forum
Cinnamon Writer forum
CodeBench forum
Digital Universe forum
Dopus 5 forum
E-UAE forum
Gnash forum
Ibrowse forum
JAmiga forum
Odyssey forum
OWB forum
Qt forum
SmartFileSystem forum
Timberwolf forum
TouchDevice forum
TuneNet forum
Unsatisfactory Software forum
Donate
)
