DSLR Videography: Low quality videos?

[HuiXL]

Hello! I have a pretty 'noob' problem here.. tried googling but did not get the solution I needed.
I have been doing photography for a while and I am hoping to venture into videography (nature cinematography). I have shot some footages but noticed that the quality of the videos is extremely poor.

I recorded the same scenes with my cheap phone and have gotten much better results. I know that DSLR videos cannot be compared with the quality of still photos, but the difference was huge. My settings were something like: 1080p 24 FPS 1/50 F8 ISO 200 neutral picture style with post-sharpening. I have tried shooting with my Canon 650d and 70d and obtained similar results. For post-processing, I used Adobe Premiere Pro, Davinci Resolve and even ACR from photoshop as I initially thought something was wrong with my workflow.

However, the original footages were already extremely muddy, unclear and full of artifacts. Videos shot using my UWA lens (10-18mm STM) were exceptionally bad while those shot using my 100mm L macro were better (I suspect it's due to the background blurring from shallow DOF). I even went to shoot some raw footages with Magic Lantern. While the videos were slightly better, they were still a far cry from my phone's videos.

Am I doing something wrong here, or is this to be expected? I mean what's the point of using my DSLR for videos when a cheap phone can outperform it.

I can upload some samples if needed. Hope someone can enlighten me on this! Thank you.

 

[Octarine]

[...]However, the original footages were already extremely muddy, unclear and full of artifacts. [...]

What were your compression settings? Artifacts usually are the result of aggressive image compression.

Secondly, how good was the light? Using f/8, 1/50s and ISO 200 requires bright daylight to work. Did you try taking a picture in Manual mode using the same settings?

 

[HuiXL]

What were your compression settings? Artifacts usually are the result of aggressive image compression.

Secondly, how good was the light? Using f/8, 1/50s and ISO 200 requires bright daylight to work. Did you try taking a picture in Manual mode using the same settings?

Thanks for your reply! Do you mean the render settings from the editing programmes? For davinci resolve I used Quicktime Apple ProRes 422 HQ. I rendered a few with H.264 and did not see much differences. I shot the videos outdoors on a sunny day. No, I did not take a picture but judging from my previous photos with same combination of lens and body the photos should be sharp.

 

[kandinsky]

Am I doing something wrong here, or is this to be expected? I mean what's the point of using my DSLR for videos when a cheap phone can outperform it.

I can upload some samples if needed. Hope someone can enlighten me on this! Thank you.

I would say it's unexpected. What video dimensions/settings are you using? Are you shooting in 1920x1080? There's no reason for 650/70d videos to perform worse than a cellphone, except for user error. Yes, samples would defintely help.

Looks fine enough.

[video=youtube;MBRX5g7YE9Q]https://www.youtube.com/watch?v=MBRX5g7YE9Q[/video]

 

[one eye jack]

Perhaps a careful reading of this will help?
Is your playback monitor progressive scan or interlaced?
Did you panned too quickly? Using class 10 memory?

Edit: Image stabaliser on or off?
ISO - Manual or auto?
http://www.imaging-resource.com/PRODS/canon-t4i/canon-t4iVIDEO.HTM

 

[HuiXL]

Apologies for the late reply. I only have this sample for now: https://youtu.be/IToSkv3XJJY. I applied a lot of sharpening to this video in an attempt to see if it would help. I cannot remember the ISO and aperture, but the FPS was 24 and shutter was 1/50, not shot in raw. I had the camera mounted on a Flycam with lens stabiliser switched on. I think that it is too 'blocky' e.g. the leaves are just green blobs.

 

[HuiXL]

Perhaps a careful reading of this will help?
Is your playback monitor progressive scan or interlaced?
Did you panned too quickly? Using class 10 memory?

Edit: Image stabaliser on or off?
ISO - Manual or auto?
http://www.imaging-resource.com/PRODS/canon-t4i/canon-t4iVIDEO.HTM

Manual ISO, image stabiliser on. I played the videos on my macbook with retina display.. I don't have much knowledge of different monitors sorry. Yes it was a Sandisk Extreme Class 10 card. The read speed is 80MB/s, not sure about write.

 

[Octarine]

Thanks for your reply! Do you mean the render settings from the editing programmes? For davinci resolve I used Quicktime Apple ProRes 422 HQ. I rendered a few with H.264 and did not see much differences. I shot the videos outdoors on a sunny day. No, I did not take a picture but judging from my previous photos with same combination of lens and body the photos should be sharp.

Not the render settings but the in camera video compression. Here a link will a few pointers. Do check the settings (use your manual) and adjust to lowest compression / highest quality.
http://www.imaging-resource.com/PRODS/canon-70d/canon-70dVIDEO.HTM

 

[one eye jack]

Manual ISO, image stabiliser on. I played the videos on my macbook with retina display.. I don't have much knowledge of different monitors sorry. Yes it was a Sandisk Extreme Class 10 card. The read speed is 80MB/s, not sure about write.

From your video the flycam is working perfectly so is the exposure.In the link I post there are sample vids with camera's stabaliser on or off.From the stills it looks like stabaliser off is sharper?

Here is something you need to know about electronic stabaliser technology used in cameras IBIS (in camera body) vs in lens stabaliser.They are designed for static subjects handheld or with tripod.
Many don't perform well if subject or camera is moving on entry/mid level cams.

If you mount the camera on a tripod (or similar stable platform) without cutting the IS, you risk creating what’s called a feedback loop, in which the camera’s IS system essentially detects its own vibrations and starts moving around, even when the rest of the camera is completely still. This introduces motion objects to your camera system and brings with it blurriness. This is one of the key reasons to turn off image stabilization.

On a final note: it is well worth mentioning that, for the sharpest results when photographing still subjects, nothing beats a camera mounted on a sturdy tripod with the image stabilization turned off. This is because image stabilization, by its very nature, using motion along one axis to counter motion in the opposite axis, often creates varying degrees of image degradation of its own, whereas a camera firmly coupled to a stable tripod and tripped with a cable or remote release with the mirror locked in the up position will, in almost every instance, take a sharper picture.

Above quotes from here: https://www.bhphotovideo.com/explor...ge-stabilization-when-use-it-and-when-turn-it


Many assume electronic stabaliser can correct in all directions (unless you have those that are 5 axis) so on most dslr it will default to side to side correction/compensation (panning).So by using a flycam stabaliser is as good as saying you are using a tripod which the above link advise to not turn on stabaliser for still subjects for sharp photo. Since only some parts of your video is blurry like center of frame,tree and top of pavilion my guess is that the default stabaliser setting is not suitable.Here is a trick you can do..try to select a camera stabalising setting that corrects in the up/down movement or combo (side/up/down).Failing that then try without stabalisation.Don't forget you are moving forwards too.

Lastly about editing... because you are editing jpegs try not to oversharpen too many times as each
edit will increase the blocking effects seen on each video frame.

EDIT: After some online research your problem is "weird" no mention of such thing.
Although you used settings 24p , 1/50Sec.for the cinematic look by right you should be ok. Octarine's link is informative where 70D offers 2 choices for compression rate namely All-I or IPB which differs in data rates.It may have some connection but there are some variables like frame rate choice.It seems Canon uses progressive scan to record video.

So I wonder what's the Macbook Retina display specs. is..( I guess all computer displays are progressive scan.) While all these specs. have to do with screen flicker problem the cause for your problem could be in the editing software settings and compatibility of 70D video file format. I don't really know.It could be the rendering video file settings.

It could be also the autofocus thing although 70D is designed for autofocus video...hmmm.
Is 24p the best optimum framerate? ( Our electrical power frequency is 50hz.)

 

[Bamboopictures]

The video looks like thumbnail resolution. Definitely not 1080. Did you accidentally import .THM files instead of the actual files?
https://www.youtube.com/watch?v=ws-40He6jK4

 

[kandinsky]

The video looks like thumbnail resolution. Definitely not 1080. Did you accidentally import .THM files instead of the actual files?
https://www.youtube.com/watch?v=ws-40He6jK4

It would be quite obvious cos .thm would not have video data? Upon import it would have behaved like a still?

HuiXL, that really looks terrible, are you able to share copies of some original out of camera files, with the accompanying .thm files so we can see the metadata as well? Like Bamboopictures, I'm wondering if you didn't shoot in 1080p, it looks severely up-rezzed from a low bitrate/reso file. You can use something like wetransfer.com. Just place your email address in both fields. Once uploaded, it will send you the link which you can share with us to download the file.

 

[Bamboopictures]

Not familiar with 70D but my Yi 4K's .THM file is actually a video file. Which often gives me a heart attack when I accidentally put them on my timeline! They look like the crappy quality TS is experiencing.

 

[kandinsky]

Not familiar with 70D but my Yi 4K's .THM file is actually a video file. Which often gives me a heart attack when I accidentally put them on my timeline! They look like the crappy quality TS is experiencing.

Wow, that's scary. If that's the case, then you might have hit the nail on the head :thumbsup: