Messages

Perhaps the difference is that it turns global draw on or off when you've plugged in the monitor?

If I have HDMI plugged in and global draw off, I get 50mbs

If I unplug HDMI, I get 54-55

If I turn global draw on, I'm banging on 60-61.

Not for me...

When going to external monitor at a higher resolution than my card can allow for, I get 224 frames with external monitor vs 337 frames without.

Without it I'm getting max of 60.1 / 60.2mbps, when I plug in the HDMI cable I get 54-55mb/s. I tested a few times to be sure.

In saying this, I might have different camera settings to you currently.

Can you please give full details like: photo mode or video mode, live view on/off, fps you are recording at, raw or jpg setting, etc etc?

If you can tell me the exact settings, I'll try to repeat my test using those too.

If you're concerned with aliasing then maybe give the 5D2 a test run before committing to buying one.

The available resolutions for the 50D in 16:9 aspect ratio are:

640x480     - 9.2mb/s
720x406     - 11.6mb/s
960x540     - 20.7mb/s
1280x720   - 36.9mb/s
1329x742   - 39.2mb/s
1440x810   - 46.7mb/s
1592x896   - 57.1mb/s   (max res unless using 5x zoom)
1600x900   - 57.6mb/s   (5x zoom only)
1720x960   - 66.0mb/s   (5x zoom only)
1880x1056 - 79.5mb/s   (5x zoom only)
1920x1056 - 81.2mb/s   (5x zoom only)
1992x1056 - 84.2mb/s   (max res for 5x zoom mode)

So it doesnt go as high as the 5D3, and doesnt have options like 60fps or whatever.

However given that it's 1/5th of the price or less, it depends on what that extra res, lack of aliasing etc is worth to you.

In saying all of this, it's still early days.
Unless you've got a burning desire to film in raw ASAP, it'd be wise to wait a little before committing $$$$$$ either way.

If money isnt much an issue, there's no disputing the 5D3 is king.

If you want to take a punt, then there's not much to lose with buying a 50D, even if as soon as you buy it the value drops in half, you've still lost less than the 'not brand new in the box anymore' depreciation of a new 5D3.

Yep I've updated so far with the numbers that Alex has provided - Is there any way I can find out buffer size etc for the cameras I own? As I could add 50D and 600D buffer sizes if so.

Most of the information left which needs filling in is ML related, so cant find it by looking at camera specs alone.

Wouldnt hold your breath for a 7D port at this stage...

As per usual the dual processors are the thorn in it's side when it comes to trying to make anything work with ML.

I think I remember reading in one of the threads that ML can be loaded onto one processor, but the raw stuff gets loaded onto the other, and the transfer speed between the two isnt good enough for a reasonable raw output.

(or along those lines)

It's a real pity the 60D has an SD card slot, it would be awesome if it could get CF speeds.

50D is alright but I really wish it had a flip out screen, and a nicer LCD resolution and a few other things I've become accustomed to with the 600D.

Thanks for the information - Yes I'll work on a better format that's updatable, later tonight.

However in the meantime if anyone's able to fill in any of the blanks as per above, appreciated.

Interesting to know about the 60D buffer - If only it had the CF slot.

Interesting about the 6D record rate too.

Okay thanks... Yep if people can specify exactly which resolution each camera is capable of in 16:9 @ 24fps then I'll update it to show the resolutions instead.

Nah, for me writing is at 57/58mb/s max while recording.

Higher in read/write tests, but there's a bottlneck elsewhere currently by the looks.

[Here's a google docs version that you can edit:]

Hey, I thought I would put together a quick comparison of where things are currently at with raw video for each camera type, and what the comparative upsides/downsides/limitations/strengths/etc of each are.

I've listed them top to bottom, of which to my understanding are currently the best options.

Here's a google docs version that you can edit:
https://docs.google.com/spreadsheet/ccc?key=0AgQ2MOkAZTFHdFFIcFp1d0R5TzVPTVJXOEVyUndteGc#gid=5

Note:  The image below is simply a screenshot of the chart, and may not contain updated accurate information.  Please visit the link above for the latest updated chart.

You just press the normal liveview zoom button, Same as zoom for normal live view without ML.

It's near where your right thumb is, top right hand side of the camera.

I'll try it when I get home, and report back.

It's not a winrar file - It's just that R00 R01 etc are also file extensions that spanned .RAR files use.

In this case instead its first .raw, then .r00, then .r01, etc.

Basically you need to merge the files back into one, using a dos command or similar in before you can decode them to dng files.

If they changed the 600D to what you want, it would be called a 60D and you wouldnt have bought it because something else was cheaper...

I've used my 600D on the end of a camera crane, and having something LIGHT makes a big difference to how much counterweight you need, makes the whole setup a lot less cumbersome.

No doubting that other things feel built a bit more solidly, but unless you're planning on throwing it down a set of stairs or something it's irrelevant in terms of reliability or longevity. Being light is one of the best things about the 600D, plastic ftw.

My 50D feels a lot more solidly built, but to what end? If I drop it, i'm more mindful that the extra weight is going to make it more likely to smash a lense. If I graze it it's still going to scrape the paint off... And not having a flip out screen is a pain.

If only the 60D had a CF card slot, it would be the best!

Does this mean that in all modes for 50D ML, there are no manual controls?

No, but that's what I initially thought as I had the settings wrong...

When you are shooting a regular H264 movie with the 50D, there are no manual controls.

When you are shooting a raw video however, you dont even need to be in movie mode... it can sample the frames from the live view screen in photo mode.

So you can set the exposure time and ISO etc as you would for a picture (with exposure compensation turned on) and it operates as per the settings you've put in there.

However I've noticed that if I choose an exposure of say 1/10th of a second when I'm recording at 24fps, it still brightens the screen towards that... which it shouldnt, I dont know what it does to the raw file when you do this.
So just be mindful of setting the exposure length to some amount below the indicated frame rate.

I'm not sure how fps override affects the indicated exposure length though, whether it's accurate for 30fps only.

So just to be clear though, it hasnt been tested connected to an actual camera?

As that's half the battle by the sounds, making sure it's all a format etc that the camera is willing or capable of interacting with.

I wouldnt be getting carried away with getting 100s manufactured etc at this stage, if it's not been proven to work with a camera yet.

On Another note, any word of get file spanning functional?

It seems to work for me already?

I accidentally recorded a file over 4gb yesterday, and I ended up with .raw and .r00

Then just recombined the files (in this case dv.raw and dv.r00) in Dos prompt with the command:

Copy/b dv.raw+dv.r00 output.raw

to end up with output.raw file which was over the 4gb limit.

Then I could output all of the files to DNG fine.

aahhh I see... I thought I needed movie mode turned on, for the raw video to work. (which is part of where I got stuck)

Makes sense, thanks guys!

I'll give it another try tomorrow.

Aha! I think I know where I went wrong.

I had exposure simulation OFF from when I was playing around with the camera before... But when I was out tonight I already had liveview/movie mode turned on, so when I went to the exposure menu to check, it just shows exposure simulation mode as MOVIE rather than on or off.

Would it be possible to update the exposure simulation setting title to be 'movie(off)' or 'movie(on)' perhaps?
Or just force it on, when in movie mode?

I cant think why you'd have it off while recording.

Ahh that's good news about the sticky DOF. 

Perhaps I had some weird settings there, prior to starting movie mode which have translated across. I'll try some more to replicate it, and find out where I went wrong.

Been doing some more tests tonight. I'm not convinced that there is any control over ISO or shutter speed at this stage. (using a manual lense, samyang 35mm)

For example if I'm recording at 30fps, I can set it to higher than 1/30th of a second which shouldnt be possible. Setting it any lower appears to make no difference I can discern either.

Also selecting through different ISOs, and the raw histogram stays exactly the same.

When I scroll the camera past a dark scene, to a light scene, and back to a dark scene, the camera is automatically adjusting the exposure to suit.

I havent checked the footage itself yet though... But I would assume it corresponds with the live view image, given that's where the raw file comes from.

As best I can tell there's no way to adjust the "exposure override" settings when it's in movie mode.

EDIT: Here are two frames from a panning shot with no settings adjusted by me:







Another update:

Okay so doing some more checks back home , it looks as though changing exposure works... up to a certain point. I think it just needs a 'lock' at the upper limit relating to FPS.

And it looks like ISO adjustment works now too... super weird.  Working in both recording and not.

I definitely had the camera set to 'M' and tried adjusting settings with no luck while out earlier.

Depending on which picture option you select, the camera reserves more (or less) memory for that, which then becomes unavailable for raw recording.

So basically we're trying to select the option that hogs the least amount of memory, to give a bigger buffer available.

Just thinking about this some more, it seems obvious that the gopro would have some onboard sound hardware to interpret the signal etc. So a straight mic jack to USB plug wouldnt help if there's nowhere to send the analogue signal on the 50D.

However I'd imagine a regular USB mic has this onboard, and just uses the USB port to stream digital data and send/recieve requests like stop/start.

Something like this perhaps, with an adaptor to mini usb:

http://www.planetheadset.com/usb-microphone.php

Although it needs drivers installed to work with windows, so I've no idea how/if there would be any useful way to transfer data to/from a camera with it where things are a bit more limited.

Or alternatively, the Blue Yeti mic doesnt need any drivers installed to make it run... And I've got one of these already.

http://bluemic.com/yeti/

Might see what else I can find out about it.

Yeah good point... I guess there might be more variance on account of that 50Ds are about 5yrs old now, compared to others. I think mine might be due for a clean.

Aahhh yeah true! I forgot about neat video heh.

Perhaps that could be the way to go, then upload Neat video presets for the pattern noise removal.

@Roman: thanks for testing. Looking forward to the results

Hey sorry I've been busy but when I had a look... some footage I filmed at 10x zoom didnt seem to be affected by this (although it was probably 'stretching' the image rather than downsizing it to achieve 10x zoom?)

I havent had a good look at the 5x ones yet.

But since it's not present in the still picture raw files that I've taken, I've got a few theories...

1. It's something to do with how raw2dng debayers the image perhaps, or this needs to work slightly different for 50D compared to other cameras? Probably not if it doesnt affect 10x zoom images.

2. When it downsamples the image for livewview perhaps it takes 'clusters' of pixels rather than 1? So you get an aliased/moired 'edge' on the blocks of 4 or 6 pixels wide which is giving the square pattern. (I'll try count the 'block' size tonight)

Might be a plan to try shoot some 'dark' frames and subtract the difference from an image to find the exact pattern and see if that helps figure out what might be causing it. Or, I wonder if part of the in camera iso noise reduction could remove this in real time if it's caused by predictable sensor noise. Or maybe worst case scenario batch subtracting 'dark' frames from shots that are affected might help if it's uniform in appearance.