Handbrake/Mediafork vs Visualhub
Handbrake/Mediafork vs Visualhub
How does Handbrake/Mediafork and Visualhub compare with one another? Aside from the ability to rip protected DVDs, is there a difference in the quality or speed of the encoding?
I used MTR to rip (full disc) my 007 Goldfinger DVD. I then used Visualhub and Mediafork to encode the movie to h.264 for playback on my Apple TV. The first thing I noticed is that Visualhub created an 862x480 movie and Mediafork created a 720x416 movie.
Why would Visualhub create a movie that appears to be higher resolution than the source DVD?
After ripping one of my Firefly DVDs with MTR (full disc) I discovered a feature I really like in Mediafork. It has the ability to encode individual episodes from the VOB files... something Visualhub doesn't support.
I used MTR to rip (full disc) my 007 Goldfinger DVD. I then used Visualhub and Mediafork to encode the movie to h.264 for playback on my Apple TV. The first thing I noticed is that Visualhub created an 862x480 movie and Mediafork created a 720x416 movie.
Why would Visualhub create a movie that appears to be higher resolution than the source DVD?
After ripping one of my Firefly DVDs with MTR (full disc) I discovered a feature I really like in Mediafork. It has the ability to encode individual episodes from the VOB files... something Visualhub doesn't support.
VisualHub is a closed-source, commercial, graphical wrapper around the open-source video encoding tool, ffmpeg.
HandBrake is an open-source, stand-alone application that uses some of the core libraries of ffmpeg, as well as several other open source libraries. It is intended specifically for ripping DVDs.
Because VisualHub is just a front to ffmpeg and some other cli tools, it is beholden to the ffmpeg team for core library updates that improve speed/quality/features. HandBrake isn't. This means, when we want a feature like chapter markers, someone can hack it into the mp4 library without dealing with another project's bureaucracy. On the other hand, because VisualHub doesn't have to write everything itself, it is easier for Tyler to add new features. And because VH is just ffmpeg, it means it can support many more video formats as input and output than HB can, because HB is written around DVDs.
As far as the output resolution from VisualHub...right now VisualHub is promoting itself with a very misleading graphic, that purports to give better quality/resolution than HandBrake. However, using the latest source code's anamorphic option, HB will indeed present video at the same size as VisualHub. I don't know what VH is doing at the moment, but I can only assume from your comment that it truly is upscaling the video and storing it at the larger size. What HB does is store the video at the same distorted size as the DVD, and let QuickTime upscale it on playback, which is more efficient.
HandBrake is an open-source, stand-alone application that uses some of the core libraries of ffmpeg, as well as several other open source libraries. It is intended specifically for ripping DVDs.
Because VisualHub is just a front to ffmpeg and some other cli tools, it is beholden to the ffmpeg team for core library updates that improve speed/quality/features. HandBrake isn't. This means, when we want a feature like chapter markers, someone can hack it into the mp4 library without dealing with another project's bureaucracy. On the other hand, because VisualHub doesn't have to write everything itself, it is easier for Tyler to add new features. And because VH is just ffmpeg, it means it can support many more video formats as input and output than HB can, because HB is written around DVDs.
As far as the output resolution from VisualHub...right now VisualHub is promoting itself with a very misleading graphic, that purports to give better quality/resolution than HandBrake. However, using the latest source code's anamorphic option, HB will indeed present video at the same size as VisualHub. I don't know what VH is doing at the moment, but I can only assume from your comment that it truly is upscaling the video and storing it at the larger size. What HB does is store the video at the same distorted size as the DVD, and let QuickTime upscale it on playback, which is more efficient.
Just to make sure I understand what's happenning...
1. DVD resolution is 720x480
2. For widescreen, the DVD contains anamorphic video that must be stretched horizontally to 853x480 so it displays in the intended 16:9 aspect ratio
The approach that Visualhub makes is that since the intended native resolution of the original video is 853x480 this is what it creates. Is it fair to call this upscaling?
The approach Mediafork makes is not to upscale/stretch the horizontal resolution since this adds data (over 18%) without improving quality. To maintain the aspect ratio it shrinks the vertical resolution (this reduces the size and quality). Some may argue that this is a workaround due to Handbrake only supporting a maximum horizontal resolution of 720.
1. DVD resolution is 720x480
2. For widescreen, the DVD contains anamorphic video that must be stretched horizontally to 853x480 so it displays in the intended 16:9 aspect ratio
The approach that Visualhub makes is that since the intended native resolution of the original video is 853x480 this is what it creates. Is it fair to call this upscaling?
The approach Mediafork makes is not to upscale/stretch the horizontal resolution since this adds data (over 18%) without improving quality. To maintain the aspect ratio it shrinks the vertical resolution (this reduces the size and quality). Some may argue that this is a workaround due to Handbrake only supporting a maximum horizontal resolution of 720.
More like 40%. [edit] - sorry, you're right, I was thinking of PAL video as I'm a Brit.Bog wrote:The approach Mediafork makes is not to upscale/stretch the horizontal resolution since this adds data (over 18%) without improving quality.
No, it's a deliberate design choice. Why add more data to compress?To maintain the aspect ratio it shrinks the vertical resolution (this reduces the size and quality). Some may argue that this is a workaround due to Handbrake only supporting a maximum horizontal resolution of 720.
HandBrake now supports anamorphic video itself (if you turn it on), which is infinitely better than either shrinking or stretching the video. HB encodes the video exactly as it comes, full resolution, and then tells the player to stretch it horizontally. No extra data to compress so smaller file sizes than Visual Hub, but still the full resolution.
The only problem is that the current release of HB/MF creates anamorphic video that is not compatible with QuickTime or AppleTV (at least not without fiddling) - but that's going to be sorted in the next release.
Last edited by hawkman on Tue Mar 27, 2007 5:59 pm, edited 1 time in total.
Exactly, in our current svn there is no need to use qt at all and it goes right into appleTV without a hitch. I have been using this feature extensive with appleTV since sunday and it really works well.hawkman wrote:The only problem is that the current release of HB/MF creates anamorphic video that is not compatible with QuickTime or AppleTV (at least not without fiddling) - but that's going to be sorted in the next release.
Not great. It doesn't choke on the file - plays it fine - but not in 5.1. More like 3.0, apparently.... http://www.thismuchiknow.co.uk/?p=34Nonsanity wrote:How does the AppleTV do with the 5.1 AAC audio?
Because the alternative, as HB does it with anamorphic OFF, is to toss some vertical pixels --- somewhere down the line something has to upscale them back. Tyler chose the other way, to preserve all the pixels + "add a column of 122 interpolated ones."hawkman wrote:No, it's a deliberate design choice. Why add more data to compress?
I'm not claiming his choice is better, just noting that while upscaling isn't ideal neither is tossing part of your source ...
Yes! Thankfully, HB's anamorphic encoding makes this whole thing a moot point!hawkman wrote: HandBrake now supports anamorphic video itself (if you turn it on), which is infinitely better than either shrinking or stretching the video. HB encodes the video exactly as it comes, full resolution, and then tells the player to stretch it horizontally. No extra data to compress so smaller file sizes than Visual Hub, but still the full resolution.
The simple answer is I use both. Sometimes and with the current 0.8.1 beta I find, my audio drops off or right at the end of an encode from DVD MediaFork crashes. This is probably just some bugs in the code and I am sure the guys working hard on the software will sort this over time.Bog wrote:Just to make sure I understand what's happenning...
1. DVD resolution is 720x480
2. For widescreen, the DVD contains anamorphic video that must be stretched horizontally to 853x480 so it displays in the intended 16:9 aspect ratio
The approach that Visualhub makes is that since the intended native resolution of the original video is 853x480 this is what it creates. Is it fair to call this upscaling?
The approach Mediafork makes is not to upscale/stretch the horizontal resolution since this adds data (over 18%) without improving quality. To maintain the aspect ratio it shrinks the vertical resolution (this reduces the size and quality). Some may argue that this is a workaround due to Handbrake only supporting a maximum horizontal resolution of 720.
When I do get a successful rip using Mediafork, I always use Anamorphic. It's sightly quicker to encode, and gives me the best results. There is one extra step though, you have to go through when you encode using this method.
You need to set the Anamorphic bit within the MP4 file after encoding so it looks correct on your TV. To do this make note of the Anamorphic settings when you encoded and then open the MP4 file with Quicktime. You then have to set the aspect ratio to to match the ripped source and export it using video/audio pass through.
There is more information here http://kilofoxtrot.bravehost.com/
The example uses different methods for encoding but the results are the same.
Maybe later on the good chaps here at handbrake/mediafork will add this extra bit in later
I like VisualHub as well though but, if I had a choice would probably stick with Mediafork.
G
gswhite, thanks for that info. I'd read various threads here about anamorphic but don't recall that caveat.gswhite wrote: ... There is one extra step though, you have to go through when you encode using this method.
You need to set the Anamorphic bit within the MP4 file after encoding so it looks correct on your TV. To do this make note of the Anamorphic settings when you encoded and then open the MP4 file with Quicktime. You then have to set the aspect ratio to to match the ripped source and export it using video/audio pass through.
There is more information here http://kilofoxtrot.bravehost.com/
The example uses different methods for encoding but the results are the same.
... which many people probably dojbrjake wrote:That would be because it's entirely unnecessary unless you use QuickTime or AppleTV to display,
... which many people probably do notjbrjake wrote: and because it's outmoded for anyone using the latest source code.
Thanks for the clarification and explanation, all the same.
-
- Posts: 8
- Joined: Thu Mar 29, 2007 10:45 am
Not sure if this will work?
I thought you needed a .m4v file for Apple TV, but this gives you an .mp4 file.There is more information here http://kilofoxtrot.bravehost.com/
Re: Not sure if this will work?
Both m4v and mp4 work, it really does not matter which you use.studio1972 wrote:I thought you needed a .m4v file for Apple TV, but this gives you an .mp4 file.There is more information here http://kilofoxtrot.bravehost.com/
G