toggle

AAPL Stock: 570.56 ( + 13.59 )

MacNN News

Apple's QuickTime loses in Video Codec Shootout

updated 01:50 pm EST, Fri March 12, 2004

Video Codec Shootout

ExtremeTech's evaluates four different video codecs for home use, including DivX, Windows Media Video 9, QuickTime 6.5/Sorenson 3, and QuickTime 6.5/MPEG-4: "We give the nod to both DivX 5.1.1 and WMV9 -- it all depends on your priorities. If you're going to stick to computers as playback devices, the faster speed of DivX is welcome. If you want to play your stuff on the PDAs, portable video players, and DVD players coming out later this year and beyond, WMV9 has broader industry support and is worth the extra encoding time. Both codecs delivered quite impressive image quality."

"QuickTime 6.5 faired reasonably well when using the common Sorenson3 codec, only really breaking down in certain high-motion scenes. It's also the fastest-compressing codec of the four tested here. The same cannot be said of Apple's MPEG-4 implementation, however, which is absolutely awful. It shows plenty of artifacts even in relatively easy, low-motion scenes, and turns into a mess of blocks when the action gets even the least bit intense. These blocky artifacts are sometimes exaggerated in still shots, but in the case of our MPEG-4 clips, it looked just as bad in motion. This is not to say that all MPEG-4 will look this bad, but there are precious few free or cheap (sub-$50) programs that offer it at this time."

by MacNN Staff

(33)

TAGS :

software, video editing

Share the Article

Facebook Twitter Delious Slash Dot Digg

toggle

Comments

ghost_flash
Professional Poster

Joined: Feb 2003

0

03/12, 02:09pm | report spam | close Reply |
WTF?

"A great many CE products shipping this year and next will support it. DivX has a reasonable amount of CE adoption as well, and may be the better choice for playback on a PC, as encoding times are much shorter and the quality is comparable."

Hmmm. They tested the Codecs on a PC only.
Whatever. M$ Owns everything.

fizzle.

slboett
Grizzled Veteran

Joined: May 1999

0

03/12, 02:10pm | report spam | close Reply |
WTF?

Can someone explain why this came out so poorly for Apple? Is it really because they used a PC?

Durandalus
Fresh-Faced Recruit

Joined: Jan 2001

0

03/12, 02:11pm | report spam | close Reply |
I observed this as well

Quicktime's MPEG 4 implementation gives horrible image quality. Let's hope Apple's been working on it.

bashar
Fresh-Faced Recruit

Joined: Jan 2000

0

03/12, 02:18pm | report spam | close Reply |
Windowz format....

so, what is it, will we have windows Money format, Windows Credit format, what the h*** is next?? windows picture format, windows c*** format, I am sick of hearing windows c*** format, Now virgin is getting into the game with windowz audio format, DVDs with windowz format.....I am suffocating here, can't they come up with windowz condom format and leave us alone.......

loudgazelle
Fresh-Faced Recruit

Joined: Dec 2001

0

03/12, 02:19pm | report spam | close Reply |
that's it!

I'm switching to windows. I'm sick of this apple bullshit.

hayesk
Professional Poster

Joined: Sep 1999

0

03/12, 02:21pm | report spam | close Reply |
What a joke!

Not only did they do the tests on Windows, when most video production is done on a Mac, they also used low bit rates, and used MPEG2 (a DVD) as the source?

So they're recompressing already lossy-compressed video?!?

An accurate test is for them to go out and shoot some footage, and then capture and compress it directly to the tested format.

Anything else is ridiculous.

Trodo
Fresh-Faced Recruit

Joined: Aug 2000

0

03/12, 02:31pm | report spam | close Reply |
Cleaner MP4 much better

They are right--Apple's MPEG4 compressions aren't very good. I have been using Cleaner and getting great MPEG4. They should not have used MPEG2's as their source, because it throws off both the quality and the compression times. Most MPEG2 video is going to be encoded anyway.

jedi1yoda1
Fresh-Faced Recruit

Joined: Dec 2001

0

03/12, 02:32pm | report spam | close Reply |
benchmarks...

As with most benchmarks on Apple/M$ stuff, I always find the test to be somewhat bias to the PC side. You hear all the time that in video in Avid on a PC beats Avid on a mac. Avid is geared for PC, not mac. So you should always take into account how the tests are being performed

edtekker
Fresh-Faced Recruit

Joined: Mar 2003

0

03/12, 02:39pm | report spam | close Reply |
Lowered Priority

I think Quicktime isn't at the same level of priority for Apple as it once was. Have you seen Apple out there pushing it as a standard? Have they been promoting QT the same way that Microsoft has been WM9 for film distribution, HDD, and so on?

It doesn't matter whether the test was on a Mac or a PC. Apple wouldn't allow iTunes to be a crappy experience on PCs, and they shouldn't let Quicktime be a poor PC option, either. It doesn't matter that most video professionals use Macs; most people viewing compressed video files are (or soon will be) the general computer user. If Apple intends for Quicktime to be in the race, they need to make it the best product in all platforms - Windows, Mac, and Linux.

I just don't think Apple cares that much about it right now. If they do care, they aren't following through very effectively.

piracy
Mac Elite

Joined: Mar 2001

0

03/12, 02:40pm | report spam | close Reply |
3ivx D4 MPEG-4 codec

3ivx's (www.3ivx.com) D4 MPEG-4 codec for QuickTime is much, much better than Apple's MPEG-4 codec. It should be made clear that MPEG-4 is NOT a bad format, or "worse" than DiVX or VC9. It's just that Apple's impementation of MPEG-4 for encoding in the default codec sucks.

chrisale
Fresh-Faced Recruit

Joined: Mar 2000

0

03/12, 02:44pm | report spam | close Reply |
Pictures don't lie

It makes absolutely no difference that it was done on a PC or Mac. The fact of the matter is that the Quicktime MPEG4 and even the Sorensen movies... looked like c***.

Period. End of story. It's undeniable.

You want the best quality for the least money, you use DIVX or WMV9.

Now, the flipside is that most Movie trailers on the Internet are in Quicktime Sorensen3 format... and of course they look awesome.

What's the difference? Mooooooooooola. You have to pay for the tools that Apple and Dreamworks and the rest use to encode their video.

The writers of the article said this right up front, so don't b**** and whine about the results because they are absolutely true using the tools they did. Apple obviously has some serious catching up to do if they want to compete in the "consumer" priced encoding space.

EMC
Fresh-Faced Recruit

Joined: Jun 2002

0

03/12, 02:47pm | report spam | close Reply |
well

MPEG-4 is a standard for compressing video, not a codec. It specifies a method of encoding image patches and vectors to patches, but leaves all of the actual decisions up to the software designer.

Things as basic as the number of key frames (frames without temporal compression), I frames, and S frames can be chosen at will. How many images patches and vectors can be similarly chosen. Basically, you create your own compression scheme using the set of allowable things to do under the standard. It appears that while Apple has adhered to the standard, they have a ways to go to make their codec best compress video while maintaining image quality. This tweaking would be cross platfrom so the windows implementation is unfortunately a good indication of a similar test on an apple machine.

The good news is that Apple can mess with the settings and build a better codec.

Rufo
Fresh-Faced Recruit

Joined: Oct 2000

0

03/12, 02:52pm | report spam | close Reply |
Shouldn't be a PC

The fact that they used a PC shouldn't affect quality. Encoding time, yes. Quality, no. Unless something is very wrong and Apple actually has different codecs on a PC vs. a Mac, the quality should be the same on both platforms. Perhaps Apple has different codecs on a PC vs. a Mac, but if they really want to push QuickTime they really should have the same quality on both platforms.

Secondly, I don't think compressing from a high-quality MPEG-2 source on a DVD should produce results *that* bad. Maybe Sorenson has an excuse, but DivX is based on MPEG-4 and doesn't produce anything anywhere *near* the artifacts that QT does.

hayesk
Professional Poster

Joined: Sep 1999

0

03/12, 03:00pm | report spam | close Reply |
Re: Pictures don't lie

Pictures can easily lie. The author referred to the DVD source as "uncompressed video" then they re-compressed it into Indeo 5.1 before testing the video. The final products have been compressed three times!!! Who would do this?

All the test proves is that Apple's MPEG4 is not as good as compressing already twice compressed video at sub-DVD level bitrates.

It in no way reflect's original content which would never be compressed in such a fashion.

Todd Madson
Mac Elite

Joined: Apr 2000

0

03/12, 03:17pm | report spam | close Reply |
Thoughts....

Read Ben Waggoner's comments at the thread at slashdot.org about this topic. Ben is an expert
on video codecs having written a book as well as
a series of articles over the past 5 years about
this subject (or his website at benwaggoner.com).

Summation:
(1) "these guys couldn't compress themselves out of a wet paper bag."

(2) They're using MPEG2 sources which are highly compressd in the first place.

(3) They never used a lossless codec to begin with.

(4) They used Indeo 5.1 as their intermediate codec. Ben says this is "terrible" due to Indeo's YUV-9 sampling which only has one measurement of color per 4x4 block of pixels, thus throwing away 75% of the color information from the DVD (whic uses 4:2:0 sapling with 2x2 blocks).

(5) They don't know the difference between the freebie Sorenson Video 3 (a freebie with quicktime) and Video 3.3 Professional (a pay for version with multiple pass variable bitrate encoding available that Apple uses to compress its movie trailers)

(6) Ben also talks about how Apples MPEG4 encoder simply isn't very good.

(7) Ben notes that they did not use 2-pass encoding.... "No quality conscious encoder would ever put content on a spinning disc without using 2-pass."

This is paraphrased from his comments in the thread at Slashdot but check it out.

They didn't even consider using 3ivx or xVid which are arguably better than the formats listed above.

ghost_flash
Professional Poster

Joined: Feb 2003

0

03/12, 03:26pm | report spam | close Reply |
Todd: Great comments

I started a thread on this topic in the "Lounge" at this link for you and all to join in a discussion:

http://forums.macnn.com/showthread.php?s=&postid=1879021#post1879021

rtamesis
Dedicated MacNNer

Joined: Jan 2000

0

03/12, 03:37pm | report spam | close Reply |
Pixar to the rescue!

Maybe Apple should use Pixar's Pixlet codec more extensively. If it's good enough for Pixar, it should be good enough for the rest of us.

jasonbw
Fresh-Faced Recruit

Joined: May 2000

0

03/12, 04:12pm | report spam | close Reply |
test; limited but flawed

As stated before, the test wouldn't have been much different if it would have run on apple hardware (does there even exist a wm9 encoder for mac?). It was supposed to be based on 4 common cheap formats, so using the $300 pro version was just out of the question.

the one test that the apple mpg4 codec did win repeatedly was the speed test, although this was barely mentioned in the story. the apple encoder was approaching twice as fast as the windows media, but the image quality was apparently all they were concerned about.

if they went back and redid the test using the two-pass method for the apple encoder, i wonder who would have won? i'm wondering if speed would be an issue then?

klinux
Senior User

Joined: Jul 2002

0

03/12, 04:20pm | report spam | close Reply |
Jokester

"used MPEG2 (a DVD) as the source?

So they're recompressing already lossy-compressed video?!?"

Oh yeah? I don't see you crying foul when recordstoreview or whatever random web site annoucing AAC (or whatever codec of the month) to be superior than MP3 when CD too has its problem! Why not do it from a master recording??

The fact is most consumers - and the keyword is consumers - do use DVD as a source. I rip my DVD to DivX all the time so I do not have to mess around with the discs. I use DivX because it has the best quality (2-pass, of course) and I encode it on a PC because my cheap-a** 2.6 ghz PC is much faster than any Mac I have or available to me!

Makosuke
Fresh-Faced Recruit

Joined: Aug 2001

0

03/12, 04:27pm | report spam | close Reply |
Hrm...

I don't much care for the methodology in this article. I did some extensive testing with DivX, Xvid, 3ivx, Apple's MPEG4 encoder, as well as an open-source MPEG4 encoder (I think it was based on MS's MPEG4 hack codec, but I don't remember for sure now), and their results aren't quite in line with what I saw.

First of all, using DVD-quality MPEG2 source (which they openly admitted actually had problems of its own with the extreme action sequences in the Matrix), then recompressing it (even at very high bitrates) with a codec like Indeo 5, is bad form.

People doing what the article claimed it was aming at--compressing some video of the kids for grandma--would never be using anything of the sort; they'd have a chunk of DV video from their video camera, which isn't lossless, but has less compression artifacts than MPEG2 most of the time. Why, exactly, didn't they just use DV video for their tests? (They also then could've supplied test clips without worry of copyright infringement.)

They said it was because the "high quality" of hollywood productions, but what does that have to do with anything--just take some nice, well lit DV footage with a decent camera. Odd.

The other big issue I have with their tests is that those aren't the results I saw; my tests showed XviD being the clear winner, with Apple's MP4 codec beating DivX in some scenes, and loosing in others--MP4 was more prone to streaking during pans with light grain in a flat area of color, and at lower bitrates does indeed get quite blocky, but DivX produced weaker colors and tended to look blurry in my tests. MP4 generally beat DivX and was close to XviD when used with Open Shiiva's encoder, though that makes the exact resulting bitrate impossible to set (not a problem if you're careful). 3ivx didn't look too bad, and the alternate MP4 codec looked awful.

Another thing that affects tests a lot is single versus 2-pass encoding; QT's encoder is single pass, and so if you set it to a hard bitrate limit it does badly in scenes with a lot of motion; with a more flexible encoder, though (again, Open Shiiva), it does EXTREMELY well to my eyes, though you might get files bigger than you were expecting if there's a lot of motion--it seems to up the bitrate as necessary to maintain quality.

Anyway, XviD has my money among the "Divx"-style, MP4-compatible codecs, though I have absolutely no problem using Apple's MP4 encoder with the right settings, either.

sakusha
Fresh-Faced Recruit

Joined: Aug 2001

0

03/12, 04:41pm | report spam | close Reply |
Flawed Benchmark

As usual, the benchmark was hideously biased against QuickTime. They used the standard Sorenson 3 codec, not Sorenson 3 Pro. The Pro version has considerably better quality than the standard version, but is still playable on any QT system. I use Sorenson Pro and the improvement in quality is dramatic.

chrisale
Fresh-Faced Recruit

Joined: Mar 2000

0

03/12, 05:11pm | report spam | close Reply |
You're missing the point

The point of the experiment was what people should expect when ripping and reencoding DVDs on their computer with free (or cheap like QT Pro) tools and using encoding schemes that are widely accepted throughout the computer world

That means Sorensen Pro is irrelevant.

The fact is that Quicktimes built in MPEG 4 encoder is far inferior to other codecs. This is confirmed by the "expert" on Slashdot.

Maybe instead of being so defensive, some people here should assk Apple why they're giving us (yes I've used a Mac for the past 8 years) inferior software?

schlie
Registered User

Joined: Jan 2004

0

03/12, 05:24pm | report spam | close Reply |
take a look at the images

to my eyes the quicktime images seem to be one of the
two best in most of the side-by-side comparisons, wonder
how it ends up ranking third?

hayesk
Professional Poster

Joined: Sep 1999

0

03/12, 05:26pm | report spam | close Reply |
Re: Jokester

"Oh yeah? I don't see you crying foul when recordstoreview or whatever random web site annoucing AAC (or whatever codec of the month) to be superior than MP3 when CD too has its problem! Why not do it from a master recording??"

iTunes *does* make their AACs from the master recording. Home users would do it from the CD because that is the best they have. But they wouldn't first encode it to MP3 or OGG or anything else and then convert to AAC. That's the audio equivalent (if there is an equivalent) of what the Extreme Tech guy did.

The fact is most consumers - and the keyword is consumers - do use DVD as a source. I rip my DVD to DivX all the time so I do not have to mess around with the discs. I use DivX because it has the best quality (2-pass, of course) and I encode it on a PC because my cheap-a** 2.6 ghz PC is much faster than any Mac I have or available to me!"

So what? Just how often do you watch the same movie? You have to mess around with the discs to rip it, then you compress it and store it to a CD or fill your hard disk. Isn't it easier just to pop the DVD in and watch it? Regardless, you wouldn't compress it to Indeo and then recompress it to DiVX or WMV, do you?

The test they did was pathetic. I have no problem with them testing cheap consumer quality codecs, but they claimed it was for "sending home movies to gramma." If that was true they should have compressed DV stream shot from a camera - plain and simple. But they chose a convoluted triple compression because that's where DiVX and WMV did better. Not only that, if grandma can play video on a CD you send her, chances are she has a DVD player either in her computer or connected to a DVD. Why wouldn't you use a consumer DVD editing package instead? Or maybe they wanted to test pirating DVDs but didn't want to say that's what they were testing? Still, the triple-compression was stupid.

hayesk
Professional Poster

Joined: Sep 1999

0

03/12, 05:31pm | report spam | close Reply |
Re: You're missing the..

" The point of the experiment was what people should expect when ripping and reencoding DVDs on their computer with free (or cheap like QT Pro) tools and using encoding schemes that are widely accepted throughout the computer world"

Actually, their point was to send home movies to grandma. In which case they should have made a DVD from DV camera with a consumer DVD package. If grandma can play video ultra-compressed onto a CD-ROM, I'll bet she has DVD playing capabilities too.

Even if they admitted the article was to test copying movies DVDs, the average consumer wouldn't compress to Indeo and then to a codec. They'd rip it and compress it only one more time, not two.

Love Calm Quiet
Mac Elite

Joined: Mar 2001

0

03/12, 05:33pm | report spam | close Reply |
Let's just hope...

this is a WAKE-UP call for Apple. Or better yet -- that they've already smelled the coffee and have something awesome cookin!

Eriamjh
Addicted to MacNN

Joined: Oct 2001

0

03/12, 06:05pm | report spam | close Reply |
I agree

MPEG4 sux. Apple needs to do better.

pdot
Senior User

Joined: Aug 2000

0

03/12, 06:36pm | report spam | close Reply |
Re: Let's just hope...

Did someone mention coffee? Darn, gotta get a cup now.

MPEG4 is supposed to be designed for streaming, wasn't it? I know it can be used to reencode DVDs, but do you think grandma really wants to see videos of grandchildren on her computer? She probably won't even like those digital pics you sent her through email. Even if QT's MPEG4 stinks, Apple is better in this grandma area b/c they know well enough to focus on the areas that matter- iDVD and iMovie. If you want to rip DVDs for cheap, then go with Divx. If you want to make movies for grandma, go with an Apple computer with a DVD burner.

tye1138
Fresh-Faced Recruit

Joined: Dec 2003

0

03/12, 07:09pm | report spam | close Reply |
Encoding

It was never mentioned but Apple's automatic encoding is designed for SPEED over QUALITY. WMV and DIVX is based on QUALITY. If the tester had bothered to do 2 pass VBR on the MPEG4 or Soreson 3 encode, the time it took to encode would have tippled, but the quality of the quicktime movies would have far impressed the viewers. The tester did a good job otherwise, but I encode Quicktimes for a living and have NEVER been able to get as poor quality as this tester!

Just as reference, I encode broadcast quality MPEG4 videos all the time at around 1Mbite per second which far succeeds the quality of the testing.

Apple codec's on PC's! PC users are dam lucky to have them at all! I think apple should stop the Quicktime on PC program due to mistakes like this article.

Thats my story, I'm stickin' to it!

Simon
Posting Junkie

Joined: Nov 2000

0

03/13, 10:39am | report spam | close Reply |
test

testing

Simon
Posting Junkie

Joined: Nov 2000

0

03/13, 10:39am | report spam | close Reply |
a test

tesing

Drakino
Grizzled Veteran

Joined: Apr 2001

0

03/13, 04:58pm | report spam | close Reply |
DVD->xvid encoding

My Powerbook 1.25 goes just as fast as my P4 2.6gHz machine here doing DVD->xvid encoding.

On the Mac side, I use Handbreak, set to 2 pass. PC side, I installed Auto Gordian Knot. Quality wise, they appeared to be the same (I would hope so with the same codec), but I still prefer the process on the Mac. Handbreak gives me more options, and also a single program. AutoGK, I still need to do something to get the VOBs off, and have no choice doing direct from DVD.

Regarding the Quicktime MPEG4, the Quicktime Pro encoder does need work. I have encoded a few thing with it, but gave up after some initial disappointing results and lack of flexability.

klinux
Senior User

Joined: Jul 2002

0

03/14, 04:34am | report spam | close Reply |
Hey

"So what? Just how often do you watch the same movie? "

More often then watching the S***** things recorded on your DV camera. What a jerk. You need to chill out.

Show More Comments

Login Here

toggle

Network Headlines

05/23	Google launches redesigned Search app...
05/23	IBM blocking iPhone's Siri on internal...
05/22	Leaked next-gen iPod touch part hints...
05/22	Hulu Plus updates with iPad Retina...
05/22	TiVo Stream to push TV to iPhones,...

Show All

05/23	Google launches redesigned Search app...
05/23	IBM blocking iPhone's Siri on internal...
05/22	Leaked next-gen iPod touch part hints...
05/22	TiVo Stream to push TV to iPhones,...
05/22	With iPad included, Apple leads mobile...

Show All

05/23	Google launches redesigned Search app...
05/22	Leaked next-gen iPod touch part hints...
05/22	Hulu Plus updates with iPad Retina...
05/22	TiVo Stream to push TV to iPhones,...
05/22	Hands on: TouchType case for iPad,...

Show All

03/02	Save $330 on refurb. 15.4-inch, 2.4GHz...
03/01	Pre-owned 8GB iPod touch 2G, drops...
02/29	$150 off Nikon Coolpix S80 Compact...
02/28	Plantronics BackBeat 903 Bluetooth...
02/27	Refurb. Seagate 1.5TB GoFlex Portable...

Show All

toggle

Apple's QuickTime loses in Video Codec Shootout

updated 01:50 pm EST, Fri March 12, 2004

Video Codec Shootout

Related Articles

Recent Articles

WTF?

WTF?

I observed this as well

Windowz format....

that's it!

What a joke!

Cleaner MP4 much better

benchmarks...

Lowered Priority

3ivx D4 MPEG-4 codec

Pictures don't lie

well

Shouldn't be a PC

Re: Pictures don't lie

Thoughts....

Todd: Great comments

Pixar to the rescue!

test; limited but flawed

Jokester

Hrm...

Flawed Benchmark

You're missing the point

take a look at the images

Re: Jokester

Re: You're missing the..

Let's just hope...

I agree

Re: Let's just hope...

Encoding

test

a test

DVD->xvid encoding

Hey

Login Here

Register for our weekly email newsletter:

Connect tools:

Subscribe

to our RSS

Follow us

on Twitter

10 Most Read

10 Most Discussed

MacNN Marketplace