ConvertX 7.0.0.76 and plans for 2023 and beyond

Latest news and official announcements

Moderators: Maggie, Coral, Cougar_II, JJ, alexia, Forum admin

Post Reply
User avatar
Jacques (VSO)
VSO Team Member
Posts: 129
Joined: Fri May 11, 2007 9:31 am

ConvertX 7.0.0.76 and plans for 2023 and beyond

Post by Jacques (VSO) »

Sorry it a very long post. To sum it up, we are late but we are working.

Last major version of ConvertXtoDVD has been released in November 2019 (7.0.0.69), and since then we have made only a few technical releases . This latest 7.0.0.76 is also a technical release without features added but with several annoying bugs solved. Especially the one that would freeze the software when loading a video, a bug that was caused by the skin system we have used.

We wanted to give something to our very patient users so we rushed to make a ConvertXtoDVD release just a few hours before the end of 2022, and 7.0.0.76 today to add a few other bugs fixed.

What about the future :

In 2019, our plan was to make ConvertXtoDVD a 64 bits application so we can address our main issue: Support of all ultra high definition sources, including those with 10, 12 or 16 bits per color component, and with 4:4:4 color space. These flavor of UHD content are currently not supported in our version of FFMPEG (the video library that we use).

At the same time we wanted to develop VSO downloader as a tool give the ability to save most video from Internet as they are watched on streaming or video website (such as the most well known Youtube). We wanted to make it a 64 bits application, and available for MAC and Linux. We also wanted to include cool features to downloader, such as support for livestream, for playlist.

We had to rewrite everything from the scratch, including one of the coolest feature of VSO downloader which is the automatic (radar) mode: as you browse video from your explorer, they are downloaded automatically. This auto-detection feature requires a local proxy server with support for HTTPS. It took 1 year to write it with the appropriate performances and avoid the main issue of the radar mode in downloader 5 - it was slowing down the browsing experience. This new proxy implementation doesn't suffer from that issue.

long story shot : while we planned 18 month timeframe to have a working multiplatform VSO downloader, it took us twice the time (3 years) to do so. But now it's done, VSO downloader has been released and so we can now concentrate on the other software, such as ConvertXtoDVD.

Therefore our development schedule for ConvertXtoDVD and the other converters is as follow:

1 - 64 bits, that's the top priority to open support for all UHD content. Technically the challenge is to rewrite our own library of graphic rendering which was massively in 32 bits assembler code. To speed this rewriting process, we will use the one from FFMPEG (swscale), which for 32 bits was significantly slower than ours and with less supported pixel formats. But if we start to rewrite all the 32 bits assembly, it will take probably more than 1 year to do so and without the guarantee of significant improvement. Therefore, we are going to use swscale as the main image resampling/merging framework.

2 - Update to latest stock FFMPEG. Historically we have used a very customized FFMPEG library, which make our code incompatible with the stock FFMPEG library "out of the box". That approach allowed us to track down and fix a lot of issues that FFMPEG was suffering from at that time - such as memory leaks or hardware acceleration support - but it has proved to be a burden to keep us up to date with the latest FFMPEG release, which was improving at a very much faster rate than what was possible from our side to keep it up to date.
Now, FFMPEG is considered as a very mature video library (and is litteraly used everywhere) and provide all hardware encoding / decoding facilities, therefore there is no real value added to use a customized FFMPEG. That way it should be also very easy to update it to support latest video/audio format, without the need to change the entire software.

3 - Sandbox all encoding jobs into a separate executable. This approach is either to have more usable memory for each encoding, and to be able to recover from a failing encoding job. This design is to make the software rock solid and less prone to corrupt all the encoding - should a single one fail. All internet browser to date also use this approach, which allow them to continue to operate normally even if there is a page being stuck somewhere. VSO downloader also uses this approach of sandboxed jobs with great success for all the download/encoding jobs.

4 - Multiplatform. Make a MAC and LINUX version. This requires to rewrite entirely the user interface, a work that has been done for VSO downloader but still need to be done for ConvertXtoDVD. It will be especially hard to rewrite for the menu editor for instance.

5 - Merge all converter products - all encoding targets into a single application . Easier to maintain for the small team we are.

And the real question is the timeframe: how much time do we need to achieve all this - knowing the fact that VSO is 1 programmer (Jacques), 1 graphic designer (Alexia) and 1 support person (Katia) - that's all that we can afford as of today :

Step 1 and 2 must be conduced together. It should take at least 6 month to complete. Once done, an ALPHA release should be made available.

Step 3 should take another 3 extra month. Once done, a BETA release could be published.

Step 4 can easily take 1 year. The graphic job can be done during step 1-3, but 6 extra month should be required after.

and Step 5 also can be achieved in parallel of all the other steps as all of our converters share 95% of their code. Yet 3 extra month shall be required for this merge.

So to be clear, we plan to have a workable BETA build of ConvertXtoDVD (64 bits, support for recent UHD sources, sandboxed encoders) by the end of 2023, and the multiplatform and merged converters for mid - end 2024.

We understand it's a long way to go, but it's a realistic one, without excluding the possibility that it takes even more time than that as we will also have to maintain VSO downloader - which is also very time consuming.

But as a good resolution for 2023, we will keep our beloved users updated for our progress in all fields

That's all for now, it's a long post but it was needed to clarify our schedule, and what we can achieve we the resource we currently have.

All the best for the year to come, may 2023 bring peace and happiness everywhere.
Jacques.
VSO Software Team Member
Post Reply