Friday, 4/10/15, Leawo's Video Converter was on GOTD. As a newer version of Leawo's version of yet another ffmpeg-based video app, I didn't [don't] know of anything new to say. It follows the design trends of the most current Windows apps, with a GUI intended to be friendlier to touch. On that basis they seem to have done a decent enough job. Leawo also followed 2 trends, one more common, one not, using a subscription licensing model & baking several apps into a single program, with the option to buy as many, or few as you want. The interface doesn't matter to me, one way or another, but I tend to dislike the bundling aspect -- I dislike installing stuff I'll never use, & in this case I'm forced to, unless of course I were to buy every option.
Otherwise, as a quick reminder, it's same ol' - same ol'... As it often does, the install routine added an old copy of the C/C++ runtimes [that will need updated via Microsoft Update], & left that installation routine's temporary setup files scattered in the C:\ or Root directory. As a rule I only install Leawo apps in a VM, then copy the software to my regular Windows install to avoid having to clean up that mess.
That said, there were two comments I thought *might* warrant expanding on, as topics in their own right. The 1st one is:
"... found it much too limiting compared to both the plethora of freeware on the net (such as Handbrake) as well as the multitude of great programs like Wondershare DVD Video Converter and Aieesooft Video Converter that is offered here quite often. Uninstalled... I would suggest those that don’t want to use Handbrake wait for one of the two I listed above to come back on GAOTD again, they are both really nice and have more features."
Now my own, personal experience tells me that's mostly incorrect -- Aiseesoft is OK, but not substantially different from Digiarty [WinX/MacX], or Wonderfox, or the majority of Chinese video apps. Wondershare in general would expand on that common model, making things a fair bit more complicated, as far as I can tell currently without much if any added benefit. Handbrake is an entirely different sort of animal. I have a "Video" folder with ~10.5GB of portable & nearly portable video apps, e.g. Wonderfox, Aiseesoft, Leawo & others. I haven't had Handbrake or a recent Wondershare video app on my drives for years.
But that's me... Not everyone is willing or wants to do much in the way of testing or evaluation, & I don't expect everyone will just take my word for it. Yet equally inaccurate statements are made whenever there's a video giveaway. *Maybe* this will help discerning some of the differences, & non-differences between video software A & B?
The quality of the final result converting [transcoding] audio/video, & how long it takes to convert, depends on three phases -- decoding the source, any conversion [e.g. video frame size &/or audio sampling], & encoding. For the best quality use the best tool you can find [or sometimes afford] for each of the 3. Anything else, including the pursuit of faster conversions, is a compromise. The question then becomes when & where are you willing to compromise, & by how much?
Going the more-or-less manual route, perhaps the most challenging source to decode is HD AVC [H.264] -- timing info's lacking compared to other formats, so most software gets it wrong, skipping &/or dropping frames [you may or may not see it but I defy you to get decent audio sync]. VC1 is actually a bit harder I think, but [thankfully] uncommon. Once you find a good decoder, you have to either pair it with AviSynth [preferred] or VirtualDub for the conversion, or save an intermediate using a lossless codec that AviSynth/VirtualDub will accept. Depending on your source & desired output you might need to include frame rate conversions, e.g. HD -> NTSC DVD you might have to go from 24 fps to 29.976, at the same time as conversion or during encoding. Or you might want/need to get rid of duplicate frames that are sometimes present at 29.976, going to 24 fps. And/Or you might have to deal with interlacing in the source &/or output. And then you have to either pair AviSynth &/or Virtual Dub to your encoder, or save another intermediate file, again in a lossless format.
The audio end of things may or may not be complicated, depending on how the source is formatted. If it's a 5.1 [or 7.1 etc.] sound track, you can keep the original number of channels & let the playing hardware, e.g. DVD/Blu-Ray player, handle it, or you can downmix it to stereo for a smaller file & greater player compatibility. If the source was originally a movie that was shown in theaters, there's also a good chance that the rear channels are far too hot, meaning all channels should be volume adjusted. Also depending on your source, it's common for audio to be stored in, but not fill up packets, that themselves may be improperly formed -- specs only call for players being able to play it. It can be a challenge to manage all of that without undue errors, distortion etc., keeping it matched to the video for sync, then merge the audio & video into a single file.
All that above is a PITA, & that was the quick, simplified version(!). It almost makes something like Leawo's Video Converter seem a minor miracle. The expression is that "The Devil's in the details", & in the case of video convertors, they all compromise quite a bit. Some give you more control than others, at the price of being more complicated to use, if only because you have to learn what those options mean to your final result. A good freeware app, strictly in terms of final quality, will basically chain high quality manual tools or steps together so you aren't forced to save huge intermediate files.
Most video converters today give comparable results with comparable speed -- if you want best speed, look at encoders/convertors focused on using your graphics hardware [e.g. Cyberlink Espresso], but realize you'll take a quality hit. There aren't that many free decoders/encoders, so most converters use ffmpeg or a fork of ffmpeg. Ffdshow &/or LAVfilters might accelerate decoding, but also can break stuff so be careful [backup]. X264 is a great AVC encoder, & it's included in ffmpeg, but the version of the MainConcept mpg2 encoder that Sony uses is so much better than the mpg2 encoder in ffmpeg that there's hardly any comparison. If you want more quality, some freeware converters use AviSynth for resizing, rather than the QT code in most of the converters out of China. AviSynth &/or VirtualDub are also [sometimes Much] better resizing than any non-linear video editing app [NLE]. That said, if the result's going to be viewed on a 7"-8" hand-held, or smaller, you most probably couldn't tell the difference.
Lots of people don't like [particularly video] software from Chinese companies -- that's just something to be aware of when/if you read reviews or comments or whatever. Use what works for you, the type of video source you usually convert, the output format you normally need, & the level of control you're comfortable with, bearing in mind that it's rather silly to bemoan the app that doesn't let you make a manual setting, but has the same value you'd use as a preset. It doesn't hurt in the least to learn about Direct Show filters & what it can mean to you when software includes them. Failing that, always check out video-related apps before installing in your main copy of Windows [I use VMs], or do a backup 1st -- it's easily possible to install an app & find it broke at least some aspect of video & or audio handling in Windows. It's also unfortunately easy to have software break something like that, & not find out about it for days or weeks afterward. Note that a *break* doesn't have to mean something fails entirely -- a break can mean reduced quality noticeable only in certain circumstances, e.g. video might be too light/dark &/or stutter or even shimmer but only in certain scenes using certain players, software or hardware.
* * *
The 2nd post is/was:
"H.264 video codec does include GPU hardware acceleration support – not sure about output quality yet."
I'm not aware of anything you can do with a computer where your mileage will vary more than when it comes to graphics hardware acceleration. A ways up above I mentioned 3 steps, decoding, conversion, & encoding -- the GPU may help with all 3. Plus there are a few ways a GPU might help out -- there's DXVA [Direct X Video Acceleration], which is so common many [most?] apps use it for their program windows or GUI; there are 3 proprietary & one open set of standard implementations: Intel QuickSync, Nvidia CUDA, ATI [AMD] Stream, & OpenCL; & there's plain ol' Direct X & OpenGL. Code may also directly address the GPU at a lower level, e.g. AMD's Mantle, though AFAIK it hasn't been exploited yet for video apps.
A graphics card may also have added electronics dedicated to video decoding &/or encoding, addressed or activated by the drivers &/or by software that comes with the driver setup files. Usually this sort of thing is found on graphics cards starting at the low end & progressing all the way up to, but not including the mainstream model. They can both speed things up & impose restrictions on output video file bit rates & or formats.
Using GPU acceleration in one of those 3 steps may mean it's not available to the other 2. Using one of the types of GPU assist may mean you can't use the other 2. This isn't an issue with most Chinese converters because they're pretty self contained. If you're using video software that works with VFW &/or Direct Show components you've installed, it can have a very definite effect. You may find for example that if you use ffdshow for decoding capabilities, setting it to use GPU assist *may* make the GPU unavailable to your encoder. Depending on the input & output formats, & the software you're using, that may or may not matter.
It's also possible to use more than one GPU, though there are restrictions... AFAIK the free Davinci Resolve is the only app that'll address multiple graphics cards, & using both a graphics card & a CPU-based GPU can be iffy, & is said to be improved in win10 when it hits its final form. I had it working to some slightest extent with Virtu in Vegas Pro with an i7 + an AMD graphics card. [Davinci Resolve can work as an editor, though that's not why it was developed, & it's something a lot of folks would at the least find interesting to check out. It does however have a learning curve, & isn't something you'll instantly be productive with.]
To tell if this sort of thing is working or not run GPU-Z, or possibly HWMonitor or MSI Afterburner. All will show a graphics card's GPU utilization, but GPU-Z will show you more details that can indicate other parts of the GPU are helping out, rather than just how much brute force it's applying. Measuring the activity of a CPU-based GPU is harder -- you can try all 3 apps & may or may not see anything telling you.
One constant -- figure on GPU assist reducing quality. Further, unless you're using something like Espresso or A's Video Convertor, that are focused more purely on using the GPU, you'll take a hit with processing coordinated between the CPU & GPU -- when you use both, the CPU % will go down. Because of that it may or may not be faster to use both -- it depends on your hardware & software. Finally, GPU assist can be Really flaky -- it depends on software & drivers installed, & well, just plain luck... I've seen people with what appeared identical hardware & software, with some saying yes & some complaining no.
Now, acceleration with ffmpeg... CUDA may work for you [with an Nvidia card], because ffmpeg developers had Nvidia graphics cards to test with. They've been trying to improve AMD [ATI] GPU support in later releases, but you normally won't see the latest ffmpeg releases in Chinese video apps. Personally I've always gotten best results in these apps with AMD acceleration off -- if it had any effect at all, the CPU would drop but I never saw much if any increase re: the GPU, & overall conversion times were always greater with GPU assist on. But mileage will vary.