Camera Labs

[Gordon Laing]

Hi everyone, this thread is to discuss the video samples shown here:

1080p HD video shootout!

They show 1080p clips from four cameras filmed straight after each other for comparison: two DSLRs and two super-zoom compacts.

Technically speaking, one of the DSLRs is actually not a DSLR at all - it's the Panasonic GH1 which doesn't have an SLR optical viewfinder or mirror. I'm also having some issues converting its 1080p file for hosting at Vimeo, so in the meantime, I've included a version filmed with its 720p mode. I'll upload the 1080p version as soon as I can.

As always, registered members of Vimeo can download the original clips for viewing on their own computers.

So which one do you think looks best?

Gordon

[Bob Andersson]

Hi Gordon,

I'm not sure this will be the answer you were looking for and I'm equally sure that as a video neophyte I'll be showing my ignorance but the answer to which one I like best is "none of the above"

So, just to make sure, I signed up to Vimeo and downloaded your 5D MkII video. My big issue is what I can best describe as a strobing effect but I'll return to that in a moment. By default the clip opened up with Quicktime 7 and it looked pretty washed out (Update: the Quicktime video controls are all at the default neutral settings). Thinking the strobing effect might be an issue with Quicktime I then opened the clip with Vista's Media Player. The difference in image saturation was startling and the result looked 100% better so do I assume that Media Player uses a different codec or does it just make better use of the Apple codec? Does anyone else see this?

But the strobing was still there. I'm not talking about the jello effect but about the feeling of jerkiness during panning. It shows up most noticeably in those vertical white steel posts in the foreground which come into view at about 11 seconds. Pausing playback in Quicktime and stepping through frame by frame shows nothing amiss and as my monitor is running at 60Hz (twice the frame rate of the video) and I use a high end graphics card I'm forced to conclude that the problem lies with the camera selecting too short an exposure time. Persistence of vision then gives the impression of several sharp images imposed on top of each other and, for some reason I don't understand, a flickery effect. Would an exposure time closer to 1/30th of a second help here?

As I say I'm a neophyte in these matters, as using the video capability of my 5D MkII has been pretty low on my list of priorities, but I'd not be happy with any of the results above as they all seem to show the same effect. If adjusting the shutter speed isn't an option (as it now is on the 5D2) then maybe there's some clever software out there to add a little motion blur when the camera is panned rapidly? Maybe that's a topic for the Editing Canon H.264 video files thread or maybe it's already been discussed and I've missed it.

Bob.
_________________
Canon EOS 5D Mk II, EF 16-35mm f/2.8L II USM, EF 24-105mm f/4L IS USM, EF 70-200mm f/4L IS USM, EF 85mm f/1.2L II USM
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head, PowerShot G10 and some other stuff.

[popo]

Just downloaded and viewed all the files. Even updated my system codecs to make sure I'm up to date as I haven't done so since 2007! I use a codec pack, media player classic and qtlite as I really can't stand apple software interfaces.

Anyway, first impressions are:

Both the 5D2 and SX1 files were quite jerky for me. I think this is probably due to a limitation in my playback ability as I notice only a single CPU core was being maxed out while the other three were sitting around twiddling their thumbs. I do note they're higher bitrate than the other two, but it definitely isn't disk transfer limiting as the file sizes are small enough to sit in cache after the first play. I believe if these were transcoded to a nicer format they would play much better.

HX1 and GH1, both played much more smoothly. The HX1 is the only one I'd say looks different from the rest, but hard to put my finger on it. I think its exposure might be slightly lower.

In all 4 cases I was getting some tearing from the old mismatched framerates between source and display.

As a general comment, I found the still frames to have a decent amount of resolution when they are not motion blur limited. Certainly enough for web size shots.

I think pixel peeping at video is possibly going too far even for me...
_________________
Sony ~ Canon ~ Olympus ~ Panasonic
Buying radar activated. Targets detected:
Medium Range - Canon 7D
Long Range - Sony A550

[Bob Andersson]

Hi Gordon,

As an experiment I transferred the 5D2 "mov" file over to my Media PC and tried it there.

That PC has a relatively modest CPU and graphics card and as a result the clip was unwatchable with QuickTime as there were lots of skipped frames. ArcSoft's TotalMedia Theatre did a much better job and I'm pleased to say that the in your face strobing effect I was complaining about is considerably mitigated. That is no doubt due to viewing the clip on a plasma TV where the phosphor persistence time is rather longer than the 8mS or so that's the advertised figure for my computer monitor.

I hope these observations are relevant to the thread as it seems that both the hardware and software used to view any of the clips makes a huge difference in one's perception of quality. Arguably the poor colour saturation I saw with QuickTime can be tuned away but the strobing effect I saw, while primarily due as far as I can tell to the short persistence time of my computer monitor, would at a guess be either more or less objectionable depending on the shutter speed selected by the camera.

On that basis one could diss the output of a camera not through some inherent fault of the camera but because of the shutter speed in use at the time. But that then begs the question of whether the shutter speed is controllable by the operator and, if not, how bright the scene being shot is. And I won't even start on whether neutral density filters are a good idea in such circumstances...

What do the experienced videographers think?

Bob.
_________________
Canon EOS 5D Mk II, EF 16-35mm f/2.8L II USM, EF 24-105mm f/4L IS USM, EF 70-200mm f/4L IS USM, EF 85mm f/1.2L II USM
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head, PowerShot G10 and some other stuff.

[Bruggeling]

dowloaded your samples from vimeo. Played with VLC they all looked better. The Panasonic is very shaky, the IS is a lot better on the Canons.

What would interest me though is how a "normal" video camera would do. I am still wondering whether to buy a new dedicated video camera.

[Taninpv]

It is hard to answer the question, because frankly. I don't know which video is which.

Here you have four cameras. Panasonic GH1 vs Canon 5D Mark II vs Canon SX1 IS vs Sony HX1

But when you posted the videos. They are not in the correct order of your post, therefore, it is confusing. We don't whether the camera Panasonic is really Canon 5D Mark II, etc.

Example:Panasonic Lumix DMC-GH1 720p / 50fps HD video
But the video is titled : Canon PowerShot SX1

Also, you posted Canon PowerShot SX1(twice)...There is no Canon 5D Mark II video.

Best regards,

[Gordon Laing]

Hi Taninpv, you've got me confused...

The videos on the page are in the correct order, from top to bottom: GH1, 5D2, SX1 IS and HX1. There is also a title above each one to remind you.

I#ve just rechecked, and none of the videos have been posted twice - there are four separate videos on the page, and they are for the correct models listed - unless I am going completely mad.


Hi Bruggeling, a comparison with a normal camcorder is a great idea - I'll include that next time I do something like this.

[HTG]

The confusion is probably caused by this: Once one has seen a video to its end, there is a display of another video of Gordon's video collection - which is not the one one just watched. An invitation to view it, too, so to say. It confused me, too, until I found out.

Btw, I liked the Sony video the best.

regards,
HTG
_________________
Support Cameralabs.com! - Choose an affiliated store or make a PayPal donation.

[Frank Stanton]

Oddly enough, I think the cheapest camera, the Sony, had the most useful video. It seemed less twitchy. I've never been impressed by SLRs which have video capabilities. I see one real use and that is to put the camera on a tripod and point it at a talking head. This could be useful for serious journalists who are interviewing people in the field who are sitting still and either cannot afford to carry a ton of equipment around, or who are afraid of being arrested for interviewing real people (think Iran and China). I have a Nikon D90 which is a very impressive still camera when mated with my 16-85 mm zoom. However, the video part of the camera is amateur hour to me. So, with Sony's legacy as a great videocam company, I would say that the Sony is the most well integrated of the cameras in your shootout.

[Frank Stanton]

Gordon,
Great site. Your site is the most influential one for me in considering what to purchase. Thanks. I am wondering if there may be some confusion about what one is looking at because after viewing a video clip the window shifts to a one shot preview with an arrow which changes the view within the window; if one puts his mouse over the preview shot one can see which camera is being covered by that window. Also, when the window shows the smaller images it only shows three video clips. When I refreshed the page, the windows reset and there was one clip per heading. Sorry if this seems confusing to explain. The bottom line is they are all there. Thanks.

[arvydas007]

I agree with Frank Stanton, that Sony hx1, had the most useful video, but there is some "burst" at the end of zoom. It seems, that SX 1 has more rapid zoom.

[pgtips]

Those videos made me sick

It must have been really cold, as that would explain Gordon's trembling hand. I guess this shows us that for professional video use in the field, nothing beats a shoulder mounted video camera.
_________________
http://philipgoh.smugmug.com/

[MattW]

hehe so this is going to be my first "constructive" post here...

I'm really interested in cameras with video capabilities since I want to make photos to show other people how it looks like wherever I am / was. But sometimes simple photos are just not sufficient to record the particular scene or atmosphere and then I'd like to make a decent video. So whichever camera I'll be going to buy in the future must have a good video function. 1080p is not really necessary in my opinion, I just don't have the right hardware to edit these amounts of data.

When I was looking for one of the best compacts out there in the last year, I chose the Panasonic Fx150 and am quite happy with it (yea, I'd always like to see less noise). Here's a video I just made:

http://vimeo.com/5481390

In my opinion its 720p quality is comparable to the DSLR videos above, altough I have to film with a fixed zoom (would love Panasonic to come up with a software update) and another thin to say: the sky was bright and blue that day...

Regards,
Matt
_________________
Panasonic Fx150

[catn]

or could you upload the raw 1080 material to Viemo for the panasonic?

[catn]

sorry, I found it, terrible indead

you had problem to convert the 1080i gh1 material, according to panasonic should the new adobe premiere been able to open it and export it to other formats, if you could upload the raw 1080i material here would be great so I can do the convertion test myself.