What do you want to see next in Media Preview?

[BabelSoft]

Here's a poll so all of you - users of Media Preview - can tell me what you want to see first implemented in Media Preview.

There are so many requests but I cannot work full time on Media Preview anymore. I got some donations (mostly from Northern America, then Australia / New Zealand, and then good old Europe -mainly Germany-) and some licensing fees from a French company, and I really want to express my gratitude to all those people that spontaneously supported my work. But sadly it's really not enough to make a living out of these Still, I want to keep on maintaining Media Preview by fixing bugs and adding new features when I have spare time off from "real life" (that's how v1.4 got out).

So tell me what you want first with the poll. This poll contains every request I got till now. Don't hesitate to elaborate. But I guess that the top request will be the fix for the thumbnail cache on Windows 8, and actually I already engineered a design to fix this issue. But this time, it shouldn't take a full year to get an official version for this Just let me know what you people think about how Media Preview should evolve.

[Awesome Donkey]

Yeah, I voted for the thumbnail fix for Windows 8. It's a no-brainer to me!

[theoneofgod]

Yeah, I voted for the thumbnail fix for Windows 8. It's a no-brainer to me!

+1 I think this is the most important thing for Windows 8+ users.

[unoentremil]

Hi,

My vote goes to Media file metadata

In fact, I was searching for a solution to see MKV metadata when I found Media Preview (other than installing the full DivX package or using MediaInfo).

I hope you find the time to implement this feature.

Best,
Diego

UPDATE: I have just discovered that metadata support (at least for MKV files) it's been removed from DivX 10! Any suggestion? Is anyone aware of an alternative? TIA.

[BabelSoft]

Quite a lot of people now is asking me to add metadata. So it will probably be added but I don't know when yet. I'll try to add this not on the next version but on the version after.

[krackaboi]

3 things.

1. I don't have win 8 but I will move to it and other people are already there. Win 8 support.

2. Multi-core support so it takes A LOT less time to get the thumbs.
*** edit: better multi-core support

3. An options to stop caching when you come across a screwed up file. I don't know how many times in only the last two weeks that one file being messed up and no thumbable has locked up the rest of the files from getting thumbnailed and jacked up other folders. It should just skip files after it finds problems instead of trying it ad nauseum and forcing you to restart and deal with the file so it doesn't screw it up again.

[theoneofgod]

3 things.

1. I don't have win 8 but I will move to it and other people are already there. Win 8 support.

2. Multi-core support so it takes A LOT less time to get the thumbs.
*** edit: better multi-core support

3. An options to stop caching when you come across a screwed up file. I don't know how many times in only the last two weeks that one file being messed up and no thumbable has locked up the rest of the files from getting thumbnailed and jacked up other folders. It should just skip files after it finds problems instead of trying it ad nauseum and forcing you to restart and deal with the file so it doesn't screw it up again.

2). The bottleneck for generating thumbnails should be the drive i/o not the CPU.

[BabelSoft]

@krackaboi: thanks for your really interesting suggestions.

1. Media Preview is already working on Win 8+. But there's this annoying cache issue where the system wipes out the thumb cache too frequently. I made an attempt to fix this with a Level 2 cache with the latest beta, which is working for me.

2. I don't know if a better multi-core support would improve that much. The slowness of Media Preview isn't actually coming much from the video frame decoding, but mostly from the luminance and entropy calculations. I don't think I can improve the entropy method. However, the time taken to compute luminance is currently directly proportional to the square of the requested thumb size (the used algorithm has a quadratic complexity). With some better maths, I could probably improve this to get it run with a linear or even logarithmic time instead of a quadratic time.
But for this, I would need to delve back into advanced maths topics, and I currently don't even have time to improve the L2 cache...

3. It's a really sensible request, and it should be really easy to implement it with the L2 cache enabled. I'll be sure to include this with the next beta when I get time to work on Media Preview again.

[krackaboi]

3 things.

1. I don't have win 8 but I will move to it and other people are already there. Win 8 support.

2. Multi-core support so it takes A LOT less time to get the thumbs.
*** edit: better multi-core support

3. An options to stop caching when you come across a screwed up file. I don't know how many times in only the last two weeks that one file being messed up and no thumbable has locked up the rest of the files from getting thumbnailed and jacked up other folders. It should just skip files after it finds problems instead of trying it ad nauseum and forcing you to restart and deal with the file so it doesn't screw it up again.

2). The bottleneck for generating thumbnails should be the drive i/o not the CPU.

For the most part but I think that he has it figured out, right below your comment. Computation seems to be the biggest issue when you have a faster HD.

@krackaboi: thanks for your really interesting suggestions.

1. Media Preview is already working on Win 8+. But there's this annoying cache issue where the system wipes out the thumb cache too frequently. I made an attempt to fix this with a Level 2 cache with the latest beta, which is working for me.

2. I don't know if a better multi-core support would improve that much. The slowness of Media Preview isn't actually coming much from the video frame decoding, but mostly from the luminance and entropy calculations. I don't think I can improve the entropy method. However, the time taken to compute luminance is currently directly proportional to the square of the requested thumb size (the used algorithm has a quadratic complexity). With some better maths, I could probably improve this to get it run with a linear or even logarithmic time instead of a quadratic time.
But for this, I would need to delve back into advanced maths topics, and I currently don't even have time to improve the L2 cache...

3. It's a really sensible request, and it should be really easy to implement it with the L2 cache enabled. I'll be sure to include this with the next beta when I get time to work on Media Preview again.

Thanks for your reply and thanks for keeping up with this great software! You seem to know what's going on and have the technical know-how to do this and find work arounds. I myself don't know much about the technical aspects of this. One other thing. Does the multi-core support work by doing one file at a time or does it work by doing multiple files at a time? I'm just wondering what the difference would be for speed of doing multiple files.

Again, thanks for your very personable response.

[BabelSoft]

@krackaboi: Doing multi-core support for multiple files at once would actually accelerate rendering globally. The thing is Media Preview is a shell extension, and is thus called by the system, which sends a thumbnail request for one file each time. So, Media Preview is only notified of new files one after the other, never as a file set, and then can really only improve multi-core processing for a single file. The system could probably decide to call several instances of Media Preview on several files in parallel, but it never does so and for good reasons: shell extension is a general purpose mechanism, not specialized for video-processing, so the system cannot decide if this kind of optimizations can sensibly or even safely be used for each registered shell extension.