Z CAM ZRAW vs. Blackmagic RAW Underexposed – Which One is Better? Our Lab Test

September 6th, 2019
Z CAM ZRAW vs. Blackmagic RAW Underexposed - Which One is Better? Our Lab Test

Z CAM E2 newest firmware (0.89) is already available for downloading (see our article by clicking here), and the big news is, it allows internal RAW recording. That made the Z CAM E2 a perfect candidate for our cinema5D latitude test. How does it fare? How does it compete directly with the Blackmagic Pocket Cinema Camera 4K? Interested in finding more? Please read on…

BMPCC 4K and Z CAM E2. Photo credit: cinema5D

For quite a while I have been using the Z CAM E2, a camera made by the Chinese manufacturer Z CAM. I shot a mini documentary with it in India (see my hands-on review here) and currently I am testing the VAZEN Anamorphic 1.8x 40mm T2 lens with it as it offers this nice almost full sensor 3696×2772 mode in ProRes HQ which desqueezes to a 1:2.4 ratio (review is coming soon).

Z CAM is constantly improving on the feature set of this camera via firmware upgrades, and just recently they made another important move and released an internal RAW recording option by the name ZRAW.

My colleague Jeff elaborated on ZRAW in this article, but in short, ZRAW supports up to 4096×[email protected] recording to the internal CFast card (60fps to come at a later stage).

In theory, the new RAW recording feature should put the Z CAM E2 “on par” codec wise with it’s major Micro Four Third sensor size competitor, the Blackmagic Pocket Cinema Camera 4k ( BMPCC 4K).

So we had to put this new feature to the test against the BMPCC 4K in our cinema5D latitude test.

The cinema5D latitude test

If you are a regular reader of our website, you have already seen some of our latitude tests. Let me quickly recap on what it is:

Latitude basically shows the capability of a camera to retain color and detail while over- or underexposing the image and normalize it back thereafter.

In this test we have a standard scene which shows Johnnie sitting next to a color checker:

Standard studio scene showing the Z CAM E2 at zero exposure level: Johnnie’s face at max. 60% luma, the white sheet of paper at 65% luma level.

From the baseline settings (exposing the face at max. 60% luma, the white sheet of paper at max. 65% luma value) we successively underexpose the shot by 1 stop until we have reached 4 or 5 stops of underexposure. Which we then push back to zero in post, by using the respective camera tools like the ZRAW VideoSuite for the Z CAM E2 and the DaVinci Resolve RAW camera tab for the BMPCC 4K.

Developing the ZRAW Files – the ZRAW VideoSuite

As said above, the ZCAM ZRAW files were developed using the standalone ZRAW VideoSuite which you can download from the Z CAM homepage (Windows 10 only for now).

The standalone ZRAW Video Suite from Z CAM (Windows 10 only).

It provides all the necessary exposure correction and grading tools that you could wish for to (batch-) develop the ZRAW files and convert them into DNxHR HQX files. However, it currently does not interface with any 3rd party software like Premiere Pro or DaVinci Resolve, hence the workflow is quite cumbersome for now.

The color grading and exposure correction tools in the ZRAW Video Suite.

For this latitude test I basically only adjusted the exposure settings (shown in 1/3 stops, so the encircled “15” value above means 5 stops of exposure correction for example), and developed the files using the Rec709 color space to avoid any additional grading steps thereafter.

The file sizes of ZRAW are about 50% bigger than the corresponding BRAW 3:1 constant bitrate files at the same resolution/frame rate of the BMPCC 4K. I don’t want to spoil the result of this test, but I don’t see any advantage using ZRAW vs. ProRes HQ on the Z CAM E2 for now.

But, first thing first…

Latitude Test Results for the Z CAM E2 and Comparison to the BMPCC 4K

Both cams were set up at their native ISO’s (500 for the ZCAM E2, 400 for the BMPCC 4K), at UltraHD 25 fps.

Noise starts to appear already at 2 stops underexposure with the Z CAM E2, and we cannot go much further than that as at 3 stops underexposure horizontal stripes start to appear which cannot be removed by noise reduction (see the NR results further down).

Z CAM E2 at 2 stops underexposure, pushed back to zero. Still usable, but noise is already strongly present in the image.

Z CAM E2 3 stops underexposed, pushed back to zero. Horizontal stripes start to appear which cannot be removed any longer. The point of no return has been reached.

Compared to the BMPCC 4K which we ran side by side to the Z CAM E2, we can clearly see that the BMPCC 4K has about 1 stop better latitude – which is consistent with our Dynamic Range measurements of both cams (Z CAM E2 shows 10.8 stops at SNR = 2, BMPCC 4K shows 11.6 stops at SNR = 2):

Z CAM E2 vs. the BMPCC 4K at 3 stops underexposure, pushed back. The Z CAM E2 (Left) has crossed the limit of usability, whereas the BMPCC 4K (Right) still holds up.

At 4 stops under, also the BMPCC 4K reaches its limits, showing horizontal stripes which cannot be removed (the same result was found in our BMPCC 4K vs. BMPCC 6K comparison here).

Z CAM E2 vs. BMPCC 4K at 4 stops underexposure, pushed back to zero. Both cams are beyond their limits.

At 4 stops underexposure, the BMPCC 4K image starts to become pink, whereas the Z CAM E2 image retains the colors – but a lot of pink noise appears. Both not very desirable. By the way, have a look at our recent BMPCC 6K lab article – you can directly compare the results as the studio settings as well as the exposure values were exactly the same.

Last question – we are dealing with RAW files on both cams, can noise reduction on the RAW files help to recover the 3 or 4 stop underexposed image? So I used a temporal noise reduction of 3 frames, threshold: 30 and spatial noise reduction with a threshold of 10 in DaVinci Resolve on the files from both cams.

Z CAM E2 vs. BMPCC 4K at 3 stops underexposure, pushed back and using noise reduction (temporal and spatial) in Resolve. The Z CAM image cannot be recovered as the stripes don’t go away (have a look at Johnnie’s shirt, for example).

Z CAM E2 vs. BMPCC 4K at 4 stops underexposure, pushed back to zero and using temporal and spatial noise reduction in Resolve. Now also the BMPCC4K image cannot be recovered any longer (more easy to see during playback of the video files, not so easy to see in the scaled static images of this article).

Now you may ask why I haven’t tried the noise reduction features of the ZRAW Video Suite?

ZRAW VideoSuite DeNoise settings

I did various tests with the DeNoise settings on the ZRAW files from the Z CAM E2 while developing them, from “Low” to “Highest”. As there is no documentation from Z CAM, you are left to speculate what the individual settings do to the image.

I found that leaving the noise reduction settings “off” in the VideoSuite and using the DaVinci Resolve settings described earlier was giving me the best results (I used the same settings for both cams).

Conclusion and a Statement from Z CAM

I love the image from the Z CAM E2 at the correct exposure settings, and I applaud the efforts of Z CAM to bring the internal raw recording to the E2 camera.

But at the end of the day, we have to see what it really offers. In essence, you should not underexpose important areas of your image more than 2 stops, and ZRAW shows rather limited possibilities to bring it back.

As things are currently, I do not see an advantage of ZRAW over ProRes for the Z CAM E2, and for the following reasons I cannot recommend using ZRAW:

  • File sizes are about 50% bigger than the corresponding highest quality settings for the BMPCC 4K, BRAW 3:1 constant bitrate (and you have the option to use up to 12:1 for even smaller file sizes in BRAW).
  • I honestly can’t see a difference to the great ProRes options that the Z CAM E2 offers as well.
  • Workflow is cumbersome because you have to use the standalone ZRAW VideoSuite (which is currently only available for Windows 10 computers) and do a conversion step before you can use the files in your standard editing software
  • The ZRAW VideoSuite offers a lot of options to color correct and exposure correct your footage but does not offer a video playback (only frame preview) – which makes it difficult to see the effect of e.g. DeNoise settings on the image

In stark contrast to all the shortfalls above, stands the strong competition from Blackmagic RAW: File sizes are comparatively small (with a lot of options to further reduce them while retaining the 12 bit raw character of the image), the workflow in DaVinci Resolve is super easy and even 6K BRAW files run smoothly on my 2014 Laptop …

A Statement from Z CAM

We want to be fair and give a chance for the camera manufacturers to respond to our findings before publishing such an article, so we approached Kinson Loo (Z CAM’s CEO) and asked for his response.

He and Jason Zhang from the technical team were very cooperative and responded immediately, giving us some insights on their motivation for implementing ZRAW and de-noising:

“The key point for RAW is the details of the image. … The RAW test should always include detail test. Our ZRAW is designed to achieve a visually lossless image of a complex scene. We suggest that you shoot complex scenes to check the resolution of the ZRAW. … This would explain why you need larger ZRAW files. … If you don’t care about extreme details but care more about noise then ProRes is a much better choice. …ZRAW itself should have a lot more noise than a YUV based format as the noise came from the sensor. YUV based files including ProRes and H265 have the image signal processor involved so it will have a different level of noise reduction which cannot be performed at Bayer level. … In camera noise reduction / curves / color correction are some major components of the image signal processor … We don’t add 2D noise reduction and chroma noise reduction in ZRAW VideoSuite and our Z CAM  DeNoiser adopts full optical flow based 3D noise reduction which does not hurt details. We would consider adding more denoise related features to our image processing pipeline in the future. … Since your test is noise based so we recommend using our denoise software Z CAM DeNoiser to process the video and test again. We could reduce more noise in the camera but some professionals don’t like it and that’s why we introduce ZRAW”

What do you think? Have you used the Z CAM E2, do you have experience with ZRAW? Please share your comments in the section below.

49
Leave a Reply

guest
1
1
- 2
Filter:
all
Sort by:
latest
GS Ford
GS Ford
Guest
April 27th, 2020

Thanks for these tests! This is exactly what I have been wanting between these two cameras!
Looking forward to your testing of the ZCam S6, as well as the ZCam F6, S1H and Mavo LF comparison.

GS Ford
GS Ford
Guest
April 27th, 2020

Was Johnnie wearing a black shirt? Does the Pocket4k not have an IR filter?

Cristian B
Cristian B
Guest
December 8th, 2019

Blackmagic is so vastly superior in this test that I feel that Zcam better do their homework if they dont want to go under soon. In fact I think quality/dollar there is nothing remotely close to Blackmagic this days. For indy cinematographers such a dream come true.

Alex Mand
Alex Mand
Guest
January 12th, 2020
Reply to  Cristian B

Not so sure. Zcam has interesting features like wi-fi, so you can use a tablet or cellphone. Or even better, use it as director’s monitor. Prores quality is insane and as mentioned before, the way zcam stores the stops is different than the blackmagic pocket. So this test is not a precise one. I got both cameras and to me at least, the zcam is better for everyday use. The lowlight performance is really good and on par with the pocket 4k. But is so small and light that I’m thinking on selling my pocket 4k and just keep the zcam.

Roberto Mettifogo
Roberto Mettifogo
Guest
October 3rd, 2019

the color checker is reflecting the softbox, cannot be a good reference in this way, it should tilted down a bit. then you can use it to expose properly, the way it is now will give underexposed picture.

 Alex Raaen
Alex Raaen
Member
September 13th, 2019

I feel bad for anyone who bought the zcam.

 Leigh Bloomfield
Leigh Bloomfield
Member
September 10th, 2019

Hello, any chance you will be able to make a download available to the video you shot for us to download and test out?

 Radu Vlad
Radu Vlad
Member
September 8th, 2019

“provides all the necessary exposure correction and grading tools that you could wish for to (batch-) develop the ZRAW files and convert them into DNxHR HQX files”…
So it’s using an Avid MC Workflow?

Member
September 8th, 2019

Actually Im glad to hear a 2014 laptop deals with 6K BRAW so well. Mac or PC? Oh and by the way, arguing over a title ?

 David Belmonte
David Belmonte
Member
September 7th, 2019

FYI someone wrote this in ZCam FB wall:

“I tested there are 6,5-7 stops above middle grey. That leaves 5-5,5 stops below middle grey. It is fixed on E2. So exposing correctly is the secret.
The BMPCC4k at 400 ISO has 9 stops below middle grey (and is adjustable because of Cine EI), but only 4 stops above. So comparing 9 stops against 5 stops Blackmagic has nearly double the latitude in the shadows: but at same native ISO I think Blackmagic is way less sensitive than E2, so you need a lot of light BMPCC4K to compensate. If you want absolute clean blacks that are recoverable you’d need ISO 100 and more light because sensor is less sensitive because the digital gain is reduced.
They should taken the same lens at same t-stop, same lightning, same shutter and resolution and should have tried to match the P4K by changing ISO to get the middle grey at 40% to match the sensitivity of E2 at 500 ISO.
I had a test with my GH5s: ISO 2500 on E2 equals ISO 5000 on GH5s. So numbers don’t matter, the sensitivity is changed by CineEI. And at 5000 ISO GH5S is really worse than E2.”

 Paul Corneille
Paul Corneille
Member
September 7th, 2019

Braw workflow is just a dream, is just perfect for almost any production in 2019-2020.
Not shooting in raw (at least 12 bit) for a professional production at the end of 2019 is just a nonsense.
Leave all the under 12bit not raw formats to v-bloggers and amateurs the real golden Era of digital film making just started.

 David Belmonte
David Belmonte
Member
September 7th, 2019
Reply to  Paul Corneille

hahaah boom!

sam broggs
sam broggs
Member
September 7th, 2019
Reply to  Paul Corneille

All these amateurs out there who don’t own a bmpcc4k. ;))

Camera Guy
Member
September 14th, 2019
Reply to  Paul Corneille

That’s a bold claim Paul. That only applies to a Professional Scripted TV/Film arena that already have the money to go Alexa/Red and would only use one of these for a crash cam. The majority of broadcasted TV (Reality, News, TV Doc) doesn’t shoot in Raw. It costs far too much for Post to cut in RAW for many productions (It would crash most Avids, especially BRAW) for a final product in 1080i. The 422 (444 if possible) 10 bit codec is a necessary option for all New Professional Cameras.

 Wayne S
Wayne S
Member
September 7th, 2019

The file sizes for high level of detail retention fit my estimations. I have been trying to get BM to do a special mode and or options, exactly like this, and to move to new JPEG XS to get file sizes down. Braw is not much Bayer with the noise correction affecting the Bayer values, it is like an in between.

Gunther, are there any latitude step chart and waveform testing coming? It is a nice way to see just home useful stops are at the end of ranges.

 David Belmonte
David Belmonte
Member
September 7th, 2019
Reply to  Gunther Machu

FYI someone wrote this in ZCam FB wall:

“I tested there are 6,5-7 stops above middle grey. That leaves 5-5,5 stops below middle grey. It is fixed on E2. So exposing correctly is the secret.
The BMPCC4k at 400 ISO has 9 stops below middle grey (and is adjustable because of Cine EI), but only 4 stops above. So comparing 9 stops against 5 stops Blackmagic has nearly double the latitude in the shadows: but at same native ISO I think Blackmagic is way less sensitive than E2, so you need a lot of light BMPCC4K to compensate. If you want absolute clean blacks that are recoverable you’d need ISO 100 and more light because sensor is less sensitive because the digital gain is reduced.
They should taken the same lens at same t-stop, same lightning, same shutter and resolution and should have tried to match the P4K by changing ISO to get the middle grey at 40% to match the sensitivity of E2 at 500 ISO.
I had a test with my GH5s: ISO 2500 on E2 equals ISO 5000 on GH5s. So numbers don’t matter, the sensitivity is changed by CineEI. And at 5000 ISO GH5S is really worse than E2.”

Elijah
Elijah
Member
September 7th, 2019

I don’t see ZRAW being compatible with any of the main NLEs Davinci Resolve, Final Cut ect. any time soon. I think the ZRAW workflow will be cumbersome and clunky for quite a while. You are right that ZRAW really doesn’t offer anything more than ProRes does. Didn’t think a cheap chinese knock off of BRAW was going to be better then the real deal anyway. With ZRAW having 50% bigger file sizes like how ProResRAW is also huge the BRAW workflow is simply blissfully easy. The BRAW workflow just works. You can only shoot so much in ZRAW whereas you can shoot double the amount of footage on BRAW just because BRAW files are half the size.

 Christopher Scharf
Christopher Scharf
Member
September 7th, 2019
Reply to  Elijah

Knock it off, Elijah. You sound like a cheap fanboy.

 David Belmonte
David Belmonte
Member
September 7th, 2019
Reply to  Elijah

Which means more compression?

Anonymous
Anonymous
Guest
September 20th, 2019
Reply to  David Belmonte

Pretty sure ZRAW will beat fake RAW of BRAW. Larger filesize isn’t leper, it’s bless. You are NOT going to get free lunch. Basically all else being equal the bigger your files the better image quality you get. There is no magic that can save you space without losing quality. That’s just a delusion of fanboys. I’d rather get ZCAM instead of fake RAW. (See: Noise reduction and half debayering)

 Leigh Bloomfield
Leigh Bloomfield
Member
September 6th, 2019

Do you have a link to your test footage so I could download and test it?

 Leigh Bloomfield
Leigh Bloomfield
Member
September 13th, 2019
Reply to  Gunther Machu

You could reduce the file size by making the clip shorter, 5 secs would do.

Crimson Son
Crimson Son
Member
September 6th, 2019

The headline is highly inaccurate. The test is just really about RAW in an underexposed situation. Where is the overexposure, aberration correction, detail recovery, etc?

Extremely poor journalism here.

 Guy McLoughlin
Guy McLoughlin
Member
September 6th, 2019
Reply to  Gunther Machu

The headline should have been something like…

Z CAM ZRAW vs. Blackmagic RAW : Under-exposure Latitude Test

or maybe…

Z CAM ZRAW vs. Blackmagic RAW : Part 1 of Our Lab Test

Under-exposure Latitude Test

Crimson Son
Crimson Son
Member
September 6th, 2019
Reply to  Gunther Machu

Thanks for the reply but you did not address anything I said other than the apology.
You can correct the headline you know. Being accurate is part of your job and it takes less than 5 min to do it. In fact, it will take less time than your reply to me.

NO WAY you made a judgment about the RAW formats with this very narrow comparison. Or are you?

Thanks for the survey but no thanks. Your reply leaves a lot to be desired.

 Wayne S
Wayne S
Member
September 7th, 2019
Reply to  Crimson Son

Dude. It’s OK. He’s doing a lot for people around here. Sure the title might be inaccurate, and more could be put in, like to do with latitude step chart and waveform which would show just how extensive exposure is, but it’s more than we have done. So, a bit more circumspectly, we should talk.

One thing that has to change is this forum system. It’s one of the worse, clunky, stopping in between letters and editing changes. And getting pointer relocation wrong etc. I had to resort to writing in a email program and copy over. It’s acting like it’s relaying every stroke back to base, which is pervey. I’m using the phone at the moment, and is a nightmare.

I have not noticed a survey?

 Wayne S
Wayne S
Member
September 7th, 2019
Reply to  Gunther Machu

Thank you again Gunther.

 Gentry Jonathan
Gentry Jonathan
Member
September 6th, 2019

Question. In the P4K vs P6K we saw the image go pink at 4 stops under. In this test I notice far less pink bias in the underexposure of the P4K. Did you learn anything there?

 新 村上
新 村上
Member
September 6th, 2019

I want over exposure test.

sam broggs
sam broggs
Member
September 7th, 2019
Reply to  新 村上

Looking at your Thumbnail, I think that’s a good idea. :)

Member
September 6th, 2019

I rarely use RAW, if ever. Pro res is a great format for me because the time spent in post is zilch. I welcome these tests but I am leaning towards the ZCAM, i like the way it looks (color/Res) and its a sensible pro alternative that fits my Niche. Thanks

 Markus Magnon
Markus Magnon
Member
September 6th, 2019
Reply to  Vlad Box Rojas

Why do people still believe RAW color correction and/or grading takes more time than ProRes? You can handle RAW the same way you handle ProRes. You don´t have to change the ISO or whitebalance or shadows, highlight. You can also use a REC709 lut or whatever. It takes the exact the same time to grade RAW.

 Paul Corneille
Paul Corneille
Member
September 7th, 2019
Reply to  Markus Magnon

Because most people don’t know how to grade, they just use some pre build LUT that “look nice”
Grading and color correction are task the require a certain knowledge and time do be done in the ‘pro’ way. And often is faster to grade a RAW file instead a ProRes file because you have much more important information in the raw file that is easier to work with it.
If people need to do fast post production they just have to shoot on rec709 and do a proper WB and exposure. But most of people don’t even know how to.expose properly a skin tone… Then don’t be surprised if is long to do a grade

 David Belmonte
David Belmonte
Member
September 6th, 2019

Ok, so BMPCC is better to shoot with moonlight. Noted guys thanks.
But still ZCam is a beast, that is why you are here doing this comparison :)
6K & 8K coming soon too.

Crimson Son
Crimson Son
Member
September 6th, 2019
Reply to  David Belmonte

Moonlight? I think you misread the article. This is the underexposure noise level test. You would use this on bright scenes like outdoor midday. Overexposure is what you normally want for moonlight or low light situations.

 Christopher Scharf
Christopher Scharf
Member
September 7th, 2019
Reply to  Crimson Son

Technically, underexposure and overexposure are both bad baselines for a good image. Waiting for your overexposure tests to compare.

Crimson Son
Crimson Son
Member
September 9th, 2019

I guess, but that was not what I was addressing.

1
1
- 2
Filter:
all
Sort by:
latest

Take part in the CineD community experience