I was using Avidemux a lot to test encoding times. Basically I was testing one computer's CPU vs another. I found with certain CPU's the times were WAY off.
Avidemux gave me a pathetic encode time of 24 minutes with a core i7 3770k w/ hyper-threading. This was done encoding a 16gb raw video file.
I go and install Handbrake and it encodes the same file in under 4 minutes...
Not only that but Handbrake encoded the file SMALLER by a significant amount. No quality difference at all.
I made sure to check Avidemux settings that it was using all cores. Still no difference.
I reproduced the results on a vast number of CPU's. Handbrake always encoded faster with smaller files - H264 high definition.
Why would I ever touch Avidemux again? For simple editing? It can't even do that. I was trying to combine flash clips and it totally warped the audio and slowed down the whole video. I can use a free editing program instead of goofing with Avidemux.
Someone tell a use for this program. I really liked it at first for it's simplicity until I found out these astonishing results.
+ Reply to Thread
Results 1 to 30 of 104
Both AviDemux and Handbrake use x264. If one is faster or gives a different file size it means you are using different settings in each one. You can't really compare it in that case.
I combine flash videos all the time using AviDemux and never had the problems you described. I suspect there is something wrong with your source files. Many times a flv file is not properly indexed or has corrupted meta data.
and there is no excuse for you being so rude and unappreciative of the people who write these programs and make them available for free - I guess you are from the gimme generation...
A much nicer way of posting whould be "why is Avidemux slower than Handbrake in my experience?" Methinks you may have received some useful information.BeyonWiz T3 PVR ~ Popcorn A-500 ~ Samsung ES8000 65" LED TV ~ Windows 7 64bit ~ Yamaha RX-A1070 ~ QnapTS851-4G
I like Avidemux, and I would use it more if not for one oddity.
I've found Avidemux to be about 33% slower when encoding to x264 than when using Ffmpeg to encode to x264.
That's with no filters in Avidemux and copying the audio.
The identical settings used in both Avidemux and Ffmpeg, and the output file sizes are identical.
Why is Avidemux slower ?
I was just comparing the encoding speed, shouldn't the encoding speed be pretty close if it's just a simple transcode with no filters and copy audio ?
Then I suppose you advise everyone to not use Avidemux as an encoding frontend then right ?
Just to be clear I don't think Avidemux is "bad", it's just slower than FFmpeg.
I also compared Mencoder, it's a lot slower than Avidemux.
I don't usually use frontends to encode, I prefer the command line so I can't say anything about other frontends.
Last edited by gregalan; 9th Aug 2013 at 03:04.
AviDemux's documentation. Do the free gregalanEditor and DUDERANCHEditors claim such speeds?
HandBrake to my wife for her DVD to iPhone encoding because she has no need to edit any of the DVD footage. Is that what you are trying to ask me?
Now if I need to losslessly edit DVD footage(cutting on key frames of course), Avidemux is pretty high on my list along with MPEGVCR. If I need to exchange/replace audio in a video file(MPEG2 mostly) and adjust the sync of the new audio to the video then Avidemux is on TOP of my list of chosen software. I then output those results(with ZERO video quality loss/changes) to a new MPEG2 file with brand new LPCM audio(if the replacement audio came from a CD). Then if I need an iPod/iPhone version to take with me I break out HandBrake because it's usually faster because it's a dedicated transcoder.
Does that about take care of it for you?
x264 encoder settings are the same. Which Handbrake preset did you use? The High Profile preset uses the default x264 settings and will probably be the slowest Handbrake preset. There's many x264 settings which can dramatically effect encoding speed.
x264 has it's own speed presets built in. The default is "Medium". They're designed to automatically change x264's advanced settings to give you an appropriate balance between encoding speed and quality/file size (so you don't need to fiddle with advanced options yourself). Handbrake gives you access to x264's speed presets (at least the latest version does). Try running an encode using the default medium preset, then again using one of the fast presets, then again using one of the slow presets.... you'll get the idea. You'll find the x264 speed presets under Handbrake's encoder configuration.
I have no idea what x264 settings Avidemux uses by default. If it's the default x264 settings and you used one of Handbrake's own presets which doesn't, then Handbrake will probably be faster, or much faster..... depending on the Handbrake preset you use.
The basic rule of thumb would be to start with x264's default settings (High Profile for Handbrake) and pick a quality you're happy with. Then adjust the x264 speed preset to the slowest you can stand (the Medium speed preset is fine to use). At the same CRF value (quality) slower presets will probably reduce the output file size for the same quality but there's no guarantee.... sometimes the output file size may increase.... that's a whole other story but it depends on the video being encoded. If you use 2 pass encoding (preselect an output file size) slower speed presets should (theoretically) increase the quality a little for a given CRF value.
Keep in mind the above can be "theoretical" to a certain extent. For example if you run an encode using CRF 18 and the medium speed preset, then the same encode again using CRF 18 and a slower preset, the encoding speed will be different, the output file size will be different, but the quality may appear identical. At a high enough quality setting any differences in quality might be there in theory, but nothing you'll physically see etc. Or the quality may be the same but the slower speed preset allowed the encoder to compress the video more efficiently.
PS A little while ago I ran some test encodes using Handbrake vs the commandline version of x264 (MeGUI) because someone said they thought the latest version of Handbrake was slower than the previous version. There were some small differences when running 2 pass encodes (slightly different first pass speed) but for CRF (quality) encoding using identical x264 settings the encoding speed was virtually the same.
Last edited by hello_hello; 9th Aug 2013 at 07:43.
"No. Just because you used the same quality setting (CRF value) that doesn't mean all the other x264 encoder settings are the same. Which Handbrake preset did you use? The High Profile preset uses the default x264 settings and will probably be the slowest Handbrake preset. There's many x264 settings which can dramatically effect encoding speed."
All settings were identical, I made sure of that.
I set all the x264 options in Avidemux to be equal to the "veryfast" preset.
I double checked with media info to make sure the x264 settings were the same in the output files, also the encoded file sizes were identical.
Last edited by gregalan; 9th Aug 2013 at 19:31.
there just so many wrong things said in this thread, by both sides i might add, that it's very difficult to know where to begin.
1) the OP said that handbrake produce the destination file 5 times faster with a much smaller file size to boot but he also claims that he used the exact same settings between avidemux and handbrake. the much smaller file size produced by handbrake at the same crf should be a dead giveaway that the same bit rate wasn't used for both encodes and that the settings differed substantially between applications.
BUT, the position that a video editor should not be expected to encode video at the same speed as a transcoder, when both applications are using the same underlying encoder for the encoding duties is ridiculous beyond belief because, wait for it, in both cases neither app is actually doing the encoding, they are both acting as front ends for the underlying encoder.
of course there another big but because I LIKE BIG BUTTS AND I CAN NOT LIE, sorry i thought i was Sir Mix Alot for a second, even if both apps use the same decoder and underlying encoder there are still differences in how they handle I/O, what version of a particular decoder/encoder they are using and more importantly what compiler options they used in the case of open source software where the developer builds the underlying encoder from source. using a different compiler build with different compiler optimization options can and does have big difference in final app performance.
having said that, in my experience, avidemux does seem to finish encodes slower than other apps, including video editing apps, like virtual dub. i strongly believe they use less than optimal build options when they compile their app or more likely that cheesy gcc compiler they use sucks balls (it really is worth every penny you pay for it).
In every way?
Do you know what the odds of that are?
I'd LOVE to see a screenshot of that.
Of course we will never see the screenshot.
We MAY see a MediaInfo text file but I'm SURE you won't edit it to make the two encodes identical....RIGHT?
Well you've made me curious.... so sometime tomorrow I'll download and install AviDemux and give it a test drive while comparing the speed with Handbrake. I wouldn't say you're imaging things until I've tested it for myself..... because when it comes to computers I've learned even the impossible is possible.... even likely, so I'll take your word for it until then.
MeGUI a while back, when running 2 pass encoding with different x264 speed presets, the chosen x264 speed preset effected the speed of the first pass when using one program, but not the other (I think the first pass speed stayed pretty much the same when using MeGUI but slower speed presets slowed the first pass using Handbrake and faster ones sped it up... I think). I was a little curious as to why the two programs behaved differently, but as I virtually never run 2 pass encoding myself I didn't find the motivation to investigate.
So AviDemux doesn't give you access to x264's speed presets? I didn't know that. I wonder if it'll be added at some stage. Or does AVIDemux let you manually add the speed presets to the command line? Maybe I should stop asking questions and just try it for myself.
WhyTF does Avidemux use the x264 settings it does and whytf doesn't it have a button to load the x264 defaults and why are quite a few of the settings greyed out??
I just ran a couple of quick encodes (2 minutes of the same 720p video). I clicked the "load defaults" button in MeGUI's encoder setup to load the x264 defaults as you'd expect to be able to do because not having a button to load the defaults is seriously mental....
Then I fiddled with the advanced settings in AVIDemux's encoder setup until I thought I had most of them at their defaults, but chances are I missed a lot so I'll need to do go through the x264 wiki to work out what the defaults are, given Avidemux wasn't endowed with a button to load them automatically.....
Anyway.... I wanted to run a couple of quick encodes tonight but I'll probably try to test it properly tomorrow. MeGUI took 35 seconds longer to encode the same 2 minutes of video. So for me.... Avidemux was quicker but no doubt that'll change when I get the x264 settings the same for each program.
How did you work out what the x264 veryfast preset settings are? I can probably work it out using MeGUI's encoder setup but unfortunately many other encoder GUIs don't change x264's advanced options in their configuration when you change the speed preset (ie Handbrake/Vidcoder) so you can't tell what the new defaults are.... which as I explained to the Vidcoder developer in a thread over at doom9 is completely mental.....
Anyway.... I'm going to try to get AVIdemux to use the x264 defaults, but my money's on x264's settings being the cause of any encoding speed differences.... at least until I run some more tests.
Last edited by hello_hello; 10th Aug 2013 at 10:37.
furthermore the line between video editor and transcoder is kind of blurred and has been for at least 10 years. most apps that you would consider a transcoder have had rudimentary editing capabilities for year, media coder, xvid4psp, xmedia recode, procoder, the venerable TMPGEnc, even handbrake has filtering and cutting capabilities, you simply set the start and end frame/time/chapter of what you want to encode, virtual dub, the list is endless.
even people that use x264 from the command line often frame serve video to it via avisynth, unless you have a raw uncompressed source that is already edited to your liking chances are you're not feeding video straight to x264 via a cli.
and as for family pictures, yes, i would spend the cash and buy photoshop elements rather than some free app because you never know when you're going to want to do more than just cropping and resizing.
do you also buy your cars with the the most basic necessities like brakes, a manual transmission, base engine, steering wheel and base interior?
or how about your meals, do you only eat that most basic of concoction needed to keep you alive or do you go for a few frills?
you're on the wrong side of the argument on this one, sorry.
The OP has gone rather quiet.
I seem to recall a request to provide reports of both encodes just to give an idea of what, other than bitrate, is different.
Maybe I have been living under a stone but I would have thought that 24 minutes is reasonable but then again what, in this example, is a 'RAW' file ?.
By the same token, 4 minutes appears too quick.
avidemux is "bad" for being slower than handbrake somehow brought this "very busy" forum to a crawl.
now that you had the chance to go all batty over his statement we can all return to our "very busy" schedules of explaining to various users how to transcode and edit their porn.
seriously, what are you smoking?