Epic vs 5D3 Raw Side by Side

[Emanuel A.]

Erm.. I don't see how.

Also, here's some fun at 4K:
(...)

I agree with Phil... before this thread to have opened and already after this video to have been available on youtube first, later gone, now available again but on vimeo:

I did some testing by my own (5D upsampled to match 4K/5K) more or less the same parts of Phil and here is only a sample of my findings very much discussed on skype among partners between Europe and NZ, you can bet once me and other members of the gang, we've spent hours to analyse route options and investment(s) (where) to make

In order to match both crops once they have no natively the same framing -- therefore, they give a different size, I had to upsample EPIC's 9%, so 18% in both H & V directions (larger equivalent than UHD to 4K [BMPC vs 1DC or Scarlet-X] or 1080p to 2K):

 

[Nick Gardner]

I now own your old Lomo Ekran Super Speeds.

I am glad they found a good home ;-) Totally under rated lenses. Those were a good set too.

For a 1080 deliverable, if you don't need any of the things that a professional camera offers, then ......I guess you don't need those things. I really think once you cross the threshhold of say an iphone 4, you get a decent picture. I shoot a terrible TV series with the c300. I hate that camera on a thousand different levels, but mainly I have to devote so much more energy towards technical bull shit that would otherwise go towards thinking about the shot. But again, at the level of camera 5d, epic, c300, sony f whatever the fuck it is now, they all make very nice pictures. When you start looking past the immediate aesthetics, to the codec, the resolution, how easy is it to do your job, thats where the differences lie.

At the end of the day, graded, all cable compressed, I think there is not going to be any solid difference between the two, but, I also think that's not the point ;-)

Nick

 

[Bryan Moore]

OK, and these punch ins looks identical to me...

Then you are absolutly blind.

 

[Peter Majtan]

Michael, with all due respect, you say that you have both Scarlet and the 5DmkIII and I take your word for it.
What I do not understand is why are you holding onto this example you posted a link to?
Not only the images are clearly not the same (according to everyone but yourself), but also that particular scene does not push the cameras to show their best (in terms of image quality - which seems to be your "only" concern) and also the scene was not shot with identical settings (such as the focal plane differs between the two)...

Since You have both the Scarlet and the 5DmkIII and are pushing so hard your POV (that everyone else here disagree with) - I challenge you to do the following:

- setup a scene that will challenge both cameras in terms of image quality, with main concern in that case being resolution and dynamic range
(good and simple example would be a shot through a window with bright daylight scenery outside showing some clouds in the sky and lots of foliage with tiny detail, while the edge of the frame shows inside of a dark interior of the room with both some detailed props (small plants, complex fabric patterns, etc) and also some smooth darker surfaces (to show grain detail))

- chose a lens that you will use on both cameras along with the same filter (use good quality lens - such as the EF 24mm f/1.4L II - not some $50 plastic EF lens from the 70's...)
(make sure to stop the lens to f/8 which will guarantee max IQ (your main concern) and adjust the filtration accordingly)

- with the lens set as above shoot 4x 3-sec clips with the Canon (set to 24P with 1/48 shutter) as follows:
1 - with the lens set to f/8 shoot the scene without any filtration (which will overexpose the outdoors) - focus on the interior
2 - with the lens set to f/8 shoot the scene without any filtration (which will overexpose the outdoors) - focus on the exterior
3 - with the lens set to f/8 shoot the scene with filtration for the outdoors (which will underexpose the indoors) - focus on the interior
4 - with the lens set to f/8 shoot the scene with filtration for the outdoors (which will underexpose the indoors) - focus on the exterior

- now repeat the same with the Scarlet, but this time shoot all 4 clips in three different settings:
A - set the Scarlet to 4K 24P with 1/48 shutter - for a direct comparison with the Canon (all 4 clips)
B - set the Scarlet to 5K 12P with 1/48 shutter - to get an idea of the resolution of the Epic (all 4 clips)
C - set the Scarlet to 5K 6P HDRx+6 1/48 shutter - to get an idea of the max DR achievable with the Epic (all 4 clips)

So you will have 4 sets of 4 clips each - please upload them somewhere so we all can download the RAW data (you can reduce the time of each clip to 1 sec if file size is a concern). Do let me know if you need a server to host these - I can help...

Then process these clips for the "best image quality" - which seems to be your main concern.
We will all do the same.

Then let's see if you still stand by your opinion that the 5DmkIII produces identical results (in terms of IQ) as the RED's DSMC - even when scaled down to 1080P...

I look forward to seeing this...

:sifone: Peter

 

[Christopher Barrett]

I shot one project with my 5d2 and the moiré rendered several shots unusable. In my work, with fine detail I have to shoot on the RED. There is no moiré at all. That aside, the resolution is also huge... Punch-ins, stabilization, reframing... all impossible with an HD file.

For me it comes down to ultimate quality and versatility (which directly affects ultimate quality). I paid 50k for my M and I still feel it's worth that money.

 

[Kemalettin Sert]

I agree with Phil... before this thread to have opened and already after this video to have been available on youtube first, later gone, now available again but on vimeo:

I did some testing by my own (5D upsampled to match 4K/5K) more or less the same parts of Phil and here is only a sample of my findings very much discussed on skype among partners between Europe and NZ, you can bet once me and other members of the gang, we've spent hours to analyse route options and investment(s) (where) to make

In order to match both crops once they have no natively the same framing -- therefore, they give a different size, I had to upsample EPIC's 9%, so 18% in both H & V directions (larger equivalent than UHD to 4K [BMPC vs 1DC or Scarlet-X] or 1080p to 2K):

those tests are so unrealistic..http://www.eoshd.com/content/10475/canon-1d-c-vs-5d-mark-iii-raw-and-c300-gh2-resolution-comparison you can see 5d MK3 looks like pal video next to 1DC /C300 which is under Red IQ.im gonna do my 5D mk3 vs Epic test when i get the 1000x card

 

[Mike Quill]

Hey Peter,

Once I'm up and running with the ML hack on my 5D3, I'd definitely like to do a comparison and post my findings. Just to clarify, though, I absolutely recognize that the 5D3 unhacked is inferior to RED's dynamic range and ability to grade the footage. Grading raw Canon photos, on the other hand, is a different story, which is why I think there's so much potential with the hack.

Mike

Michael, with all due respect, you say that you have both Scarlet and the 5DmkIII and I take your word for it.
What I do not understand is why are you holding onto this example you posted a link to?
Not only the images are clearly not the same (according to everyone but yourself), but also that particular scene does not push the cameras to show their best (in terms of image quality - which seems to be your "only" concern) and also the scene was not shot with identical settings (such as the focal plane differs between the two)...

Since You have both the Scarlet and the 5DmkIII and are pushing so hard your POV (that everyone else here disagree with) - I challenge you to do the following:

- setup a scene that will challenge both cameras in terms of image quality, with main concern in that case being resolution and dynamic range
(good and simple example would be a shot through a window with bright daylight scenery outside showing some clouds in the sky and lots of foliage with tiny detail, while the edge of the frame shows inside of a dark interior of the room with both some detailed props (small plants, complex fabric patterns, etc) and also some smooth darker surfaces (to show grain detail))

- chose a lens that you will use on both cameras along with the same filter (use good quality lens - such as the EF 24mm f/1.4L II - not some $50 plastic EF lens from the 70's...)
(make sure to stop the lens to f/8 which will guarantee max IQ (your main concern) and adjust the filtration accordingly)

- with the lens set as above shoot 4x 3-sec clips with the Canon (set to 24P with 1/48 shutter) as follows:
1 - with the lens set to f/8 shoot the scene without any filtration (which will overexpose the outdoors) - focus on the interior
2 - with the lens set to f/8 shoot the scene without any filtration (which will overexpose the outdoors) - focus on the exterior
3 - with the lens set to f/8 shoot the scene with filtration for the outdoors (which will underexpose the indoors) - focus on the interior
4 - with the lens set to f/8 shoot the scene with filtration for the outdoors (which will underexpose the indoors) - focus on the exterior

- now repeat the same with the Scarlet, but this time shoot all 4 clips in three different settings:
A - set the Scarlet to 4K 24P with 1/48 shutter - for a direct comparison with the Canon (all 4 clips)
B - set the Scarlet to 5K 12P with 1/48 shutter - to get an idea of the resolution of the Epic (all 4 clips)
C - set the Scarlet to 5K 6P HDRx+6 1/48 shutter - to get an idea of the max DR achievable with the Epic (all 4 clips)

So you will have 4 sets of 4 clips each - please upload them somewhere so we all can download the RAW data (you can reduce the time of each clip to 1 sec if file size is a concern). Do let me know if you need a server to host these - I can help...

Then process these clips for the "best image quality" - which seems to be your main concern.
We will all do the same.

Then let's see if you still stand by your opinion that the 5DmkIII produces identical results (in terms of IQ) as the RED's DSMC - even when scaled down to 1080P...

I look forward to seeing this...

:sifone: Peter

 

[Paramon Ivanovich]

I don't know where this guy got this video but i shot it, and i took it down from youtube because it's not really accurate, i rushed to make my opinion, but a few hours later i shot another test paying more attention to the details and settings , and then another one, an so on,
it took me a week (yeah, i am slow) to form my opinion,
ML is amazing, if you don't have an access to a real camera.

 

[Peter Majtan]

Hey Peter,

Once I'm up and running with the ML hack on my 5D3, I'd definitely like to do a comparison and post my findings. Just to clarify, though, I absolutely recognize that the 5D3 unhacked is inferior to RED's dynamic range and ability to grade the footage. Grading raw Canon photos, on the other hand, is a different story, which is why I think there's so much potential with the hack.

Mike

Even "hacked" 5DmkIII is inferior to RED DSMC (yes - even the current Scarlet, never mind Epic or the soon-to-arrive Dragon) in terms of IQ - your main concern, as well as many other aspects. But this is not to say the 5DmkIII is a "bad" camera. Not at all. Since you like car analogies - this is simply saying that Golf GTi is inferior to Nissan GTR. While the GTR is a must-have for those desiring its performance and practicality (and those who don't mind the much higher price-tag) - it does not mean that the GTi is a trash...

I have said myself several times that the 5DmkIII is IMHO the best all-round (H)DSLR, especially for still images. And yes - it shoots decent video (which now has an edge on the H.264 with the RAW hack). But comparing it (hacked or otherwise) to the RED DSMC and trying to question the extra $10K for Scarlet / $30K for Epic is simply ridiculous...

:sifone: Peter

 

[paul schefz]

Hey Peter,

Once I'm up and running with the ML hack on my 5D3, I'd definitely like to do a comparison and post my findings. Just to clarify, though, I absolutely recognize that the 5D3 unhacked is inferior to RED's dynamic range and ability to grade the footage. Grading raw Canon photos, on the other hand, is a different story, which is why I think there's so much potential with the hack.

Mike

in my experience even the canon raw still files (even at 100 iso) have less DR then red raw files....and at 800iso that is no comparison....the canon loses maybe a stop DR per stop....detail is a different story of course....and in general the still file comparison really also depends on the red compression....i am really interested in the dragon for exactly that reason....

 

[jake blackstone]

I agree with Phil... before this thread to have opened and already after this video to have been available on youtube first, later gone, now available again but on vimeo:

There is something wacky going on. According to that video, 5D has better DR, than Scarlet. It just can't be. Clearly, whoever was doing the grading goofed up.

 

[Peter Majtan]

in my experience even the canon raw still files (even at 100 iso) have less DR then red raw files....and at 800iso that is no comparison....the canon loses maybe a stop DR per stop...

That is expected as the sensor simply has less DR:

DR 11.7 - Canon 5DmkIII
DR 13.5 (18 with HDRx) - current Scarlet / Epic MX
DR 16.5+ (22.5+ with HDRx) - Scarlet / Epic Dragon

:sifone: Peter

 

[Peter Majtan]

There is something wacky going on. According to that video, 5D has better DR, than Scarlet. It just can't be. Clearly, whoever was doing the grading goofed up.

He just posted above you:

I don't know where this guy got this video but i shot it, and i took it down from youtube because it's not really accurate, i rushed to make my opinion, but a few hours later i shot another test paying more attention to the details and settings , and then another one, an so on,
it took me a week (yeah, i am slow) to form my opinion,
ML is amazing, if you don't have an access to a real camera.

 

[jake blackstone]

He just posted above you:

No, I see the problem not with Canon images but with Scarlet. I could be wrong, but I think he's talking about the ML images. For example, the highlights on epic are just blown out on motorcycle, where Canon's are fine.

 

[Michael Dalton]

Nick,

Also, will your opinion change when the 5D3 becomes rock solid, reliability wise? Can you give us links to the videos on the jobs you've done for $20k where you've delivered in 4K? Impressing the client with a big camera is sometimes an important factor, but for sake of conversation, I'm only interested in final image quality (not pecker waving considerations).

Mike

Three years ago this would have been awesome. Shooting 1080 now is like shooting SD when 1080 was making solid inroads. I'd gladly pay what I originally paid for my epic if I had to do it over again. I have no clients to impress, just my eyes at the end of the day. It's cool, but too little too late.

 

[Emanuel A.]

I agree with Phil... before this thread to have opened and already after this video to have been available on youtube first, later gone, now available again but on vimeo:

(...)

those tests are so unrealistic..http://www.eoshd.com/content/10475/canon-1d-c-vs-5d-mark-iii-raw-and-c300-gh2-resolution-comparison you can see 5d MK3 looks like pal video next to 1DC /C300 which is under Red IQ.im gonna do my 5D mk3 vs Epic test when i get the 1000x card

There is something wacky going on. According to that video, 5D has better DR, than Scarlet. It just can't be. Clearly, whoever was doing the grading goofed up.

The provenience of this comparison test is a mystery indeed. Some hours after its first post on youtube, the video had been removed.

 

[David W. Jones]

These vs threads tend to get a little silly do they not?

The thoughts of a commercial whore... I use what ever camera makes my clients happy, so I can make lots of money!

All the best!

Dave

 

[Jeremiah Patton]

The provenience of this comparison test is a mystery indeed. Some hours after the first post had been removed.

Looks like some trolling is going on! Did someone switch those logos around? :thumbdown: Not to mention, Scarlet can be much sharper than either of those shots. In the end it comes down to the CC. You either think it's the camera's fault and give up or you've found that the problem is having a horrid grasp on how to CC RED footage. I've gone through a whole bunch of those stages, and settled on I was a total RED noob.

Down scaled to 1080p RED footage should be pin sharp and I had to learn that the hard way. Sharpening it's an amazing thing.

 

[jimhare]

For me it's simple, the Epic/Scarlet give me more control so I can put much more confidence into what I'm bringing home.

Toia is probably the best example. He comes home from a shoot and there is probably $100k riding on the fact that he got the goods. Everything's sharp, no late surprises with wardrobe moire, rolling shutter or other things you might not be able to predict but can ruin a shoot.

Basically no compromise.

For me, I shoot live concerts that cannot be repeated so if I come home with anything less than perfection there would be a world of hurt and my reputation would be in the toilet.

Also, when working on higher budget projects you have the luxury of being able to afford the best tool for the job.

Otherwise I would definitely be using lower cost gear, working around the limitations and accepting that once in a while it WILL let me down completely.


So I don't think the real question is whether the cameras can ultimately deliver the same image, it's what is it worth to you to make sure they do?


And I agree with those who say you should be comparing only to the Scarlet. There are plenty of Scarlet/Epic "why pay the difference" discussions around to cover that area.


I hope this doesn't come across as condescending because it isn't meant to be, just trying to come from a different direction to address your concerns.

I have an Epic, Scarlet and a 550D (Ti2.)

Finally, amortize it over 3 or 5 years and decide if the cost difference is as large as it seems.

 

[Karim D. Ghantous]

Well, that was an interesting discussion (yes, I do read all posts before I give my opinion). I'm surprised at some of the poor reasoning employed by one or two members here. Camera choice is one's own business, but be honest about it, if only with oneself.

Here's another angle: if you had to shoot stills, would you choose the 5D3 or the Scarlet? Budget aside, I'd choose the Scarlet. BTW I don't own a DSLR or a RED and my main digital camera is a Sony NEX. For what I shoot the NEX is superior to a DSLR. When I feel like spending some real money, I'll keep an open mind and hopefully make the right compromises.