Jump to content

Canon 1DX-II vs. 1DC - Which one would you buy?


Guest Ebrahim Saadawi

What's a better buy?  

49 members have voted

  1. 1. What's a better buy?



Recommended Posts

Also I have been pixel peeping my 4K frames. When the sharpness is dialed all the way down, I do not see any sharpening effects. It looks to me like Blackmagic, not like Canon video. Every video frame is so high quality that I can use them as stills if I want. I definately could not stand stills from the standard Canon video.

1DX2 FullHD video, based on a short clip I did with it, works like on any other Canon DSLR, and I am suspecting that all the abovementioned deficiencies in the resulting image do apply. However, I will never use it. The 4K mode is so nice. I am after all not missing my BMCC and no longer dreaming about UM4.6K. Actually there was a RED day at e dealer that arranged my 1DX2 purchase. I was thinking of going there, but the 1DX2 is too exciting camera to bother to go to RED-demonstration. I think the only real inferiority with the camera is limitation to 8 bit and no 10 bit not to speak of 12-14 bit raw. If the 1DX2 would shoot raw video at 4K, it would make UM4.6K, UM4K and Red Raven at least, obsolete.

 

Link to post
Share on other sites
  • Replies 214
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Popular Posts

You are right. The genre of documentary was invented in 2016. Before that it didn't exist. TV news does not exist. 100% of the people here that shoot events etc use AF 100% of the time. Every document

I'd go for the 1D C. The only thing the 1D X II has in its favour is the AF. The 4K/60fps produces unmanageably large files and will be extremely difficult to edit on any current CPU, beside

Easy choice. I did it already and bought 1DX mark II.  I use Technicolor Cinestyle profile in the 1DX2 and it works just fine. Check my other post about those. Cinestyle flat cat frame is at

Posted Images

Guest Ebrahim Saadawi

Shoot a light bulb with 1- Neutral -4 Contrast & 2- Cinestyle -4 Contrast. Post here.

With the 1DC, both clip at the exact same point. The image is just distributed flatter in between these two fixed hard spots on the top of the highlights and bottom of shadows. 

Maybe something is different about the 1DX-II Cinestyle/Flat Picture styles and it increases DR. Doubt it. 

Link to post
Share on other sites

As Ebrahim explained so well in his post, C-Log has access to the RAW image data coming off the sensor before it has gone through JPEG encoding. This is why C-Log on the 1DC is capable of recovering highlights while retaining shadow detail simultaneously:

It’s an astounding performance here that rivals or exceeds the DR of many cinema cameras including Canon’s own C500. No amount of tinkering with the 1DX II picture profile settings and “cinestyles” can actually affect the recorded gamma curve like C-Log so as to capture this kind of DR.

Link to post
Share on other sites
1 hour ago, Kino said:

As Ebrahim explained so well in his post, C-Log has access to the RAW image data coming off the sensor before it has gone through JPEG encoding. This is why C-Log on the 1DC is capable of recovering highlights while retaining shadow detail simultaneously:

It’s an astounding performance here that rivals or exceeds the DR of many cinema cameras including Canon’s own C500. No amount of tinkering with the 1DX II picture profile settings and “cinestyles” can actually affect the recorded gamma curve like C-Log so as to capture this kind of DR.

C-log on 1DC offers decent dynamic range, but you are exaggerating saying it's better than C500. In fact it's inferior than most cinema cameras.

See this dyanamic range test:

http://www.cinematography.net/edited-pages/NZCS-Contrast.html

Link to post
Share on other sites
3 hours ago, Luke Mason said:

C-log on 1DC offers decent dynamic range, but you are exaggerating saying it's better than C500. In fact it's inferior than most cinema cameras.

See this dyanamic range test:

http://www.cinematography.net/edited-pages/NZCS-Contrast.html

That's not a very good test as the 1DC in that comparison was shot with a different framing that placed more sunlight into the background relative to the other cameras. They should all have the same framing. Despite this problem, you can see that the overexposure of the building looks equal on the 1DC and C500 frame grabs.

Now the "cinema cameras" I was thinking of are those in the 1DC's general price range such as the FS7 and C300. But, in any case, different tests will produce different results, especially when we have no idea what the exact settings were on each camera. Here is an actual side-by-side video that shows a very good performance for the 1DC in terms of its DR when compared to cinema cameras including the C500:

It's also important to note that the C500 has an expanded DR in HD 12 Bit RGB 4:4:4 relative to what it achieves in 4K 10 bit RAW, where its latitude is reduced. In the HD mode, it is a very impressive camera indeed. In 4K, however, the C500 has a reduced DR of around 12 stops, which is roughly equivalent to the FS7 and the C300 II:

https://***URL not allowed***/canon-c300-mark-ii-review-dynamic-range/

The 1DC itself was measured at 12.5 stops by Hurlbut, who has tested both the C500 and the 1DC very extensively:

http://www.thehurlblog.com/film-education-online-the-next-gen-in-digital-film-capture-canons-4k-1dc/

His exposure and latitude tests for the 1DC and C500 are available on his channel. Here are just a few examples:

 

 

 

 

In both overexposure tests, you can see that the C500 and the 1DC will go about 3 stops over before they clip in a way that is unrecoverable, unlike the Alexa which can go much further. Based on Hurlbut's tests, I would say that the 1DC indeed "rivals" the C500 when it comes to DR but not in any other way as a proper cinema camera.

Link to post
Share on other sites
1 hour ago, Kino said:

That's not a very good test as the 1DC in that comparison was shot with a different framing that placed more sunlight into the background relative to the other cameras. They should all have the same framing. Despite this problem, you can see that the overexposure of the building looks equal on the 1DC and C500 frame grabs.

Now the "cinema cameras" I was thinking of are those in the 1DC's general price range such as the FS7 and C300. But, in any case, different tests will produce different results, especially when we have no idea what the exact settings were on each camera. Here is an actual side-by-side video that shows a very good performance for the 1DC in terms of its DR when compared to cinema cameras including the C500:

It's also important to note that the C500 has an expanded DR in HD 12 Bit RGB 4:4:4 relative to what it achieves in 4K 10 bit RAW, where its latitude is reduced. In the HD mode, it is a very impressive camera indeed. In 4K, however, the C500 has a reduced DR of around 12 stops, which is roughly equivalent to the FS7 and the C300 II:

https://***URL not allowed***/canon-c300-mark-ii-review-dynamic-range/

The 1DC itself was measured at 12.5 stops by Hurlbut, who has tested both the C500 and the 1DC very extensively:

http://www.thehurlblog.com/film-education-online-the-next-gen-in-digital-film-capture-canons-4k-1dc/

His exposure and latitude tests for the 1DC and C500 are available on his channel. Here are just a few examples:

 

 

 

 

In both overexposure tests, you can see that the C500 and the 1DC will go about 3 stops over before they clip in a way that is unrecoverable, unlike the Alexa which can go much further. Based on Hurlbut's tests, I would say that the 1DC indeed "rivals" the C500 when it comes to DR but not in any other way as a proper cinema camera.

 

Your claim that the 1DC and C300 Mk II have equal DR makes your posting history seem specious. Talk about picking and choosing methodologies to forward an inaccurate narrative. The C300 Mk III kills in DR when measured properly.

That said, there are so many awesome options now. It really does boil down to individual need. Or whether you can afford an Alexa or not.

Link to post
Share on other sites

The Canon 1DC may well have better dynamic range in clog but I believe the advantage may only be a stop at best with the 1DXii already having a 1 1/2-2 stop dynamic range advantage at base iso, and a further 1 stop gain shooting in cinestyle -1 ev htp, which adds up to around 3 stops. This compares favourably to the 1DC and its 2-3 stop gain using canon log from a weaker sensor. 

What I'd really like to see is a direct comparison. 

Link to post
Share on other sites
Guest Ebrahim Saadawi

The 1DC image has more dynamic range and much better highlight rolloff, period. It's not a point to be argued. What we want to see is how much that gap in DR actually is. 

Just one thing to say: Since Philip blooms Video review of the 1DC where he showed footage on Brighton Beach at winter and landscapes of Dubai, then seeing every single 1DC footage out there graded by anyone, the image is absolutely gorgeous, with a very unique filmic signature, that can only be compared to f35/Alexa. While all the 1Dx-II footage I've seen so far have a different look. A different feel. It's a very high-end 4K image with lovely colours, but just non of the 1DC mojo I've seen. 

I always say wait for side by sides but also watching genertal footage gives an accurate idea of how a camera looks. It's like watching different film stocks. The 1DC is a different stock from the 1DX. And the 1DX has those amazing DPAF/60p tools that alleviate production value. 

The mojo is most definitely in the C-LOG gamma dynamic range/roll-off and don't forget C-Log has a different colour science. So it's C-Log most likely that forces the posted footage on the internet to have a certain look. 

Link to post
Share on other sites
10 hours ago, Policar said:

 

Your claim that the 1DC and C300 Mk II have equal DR makes your posting history seem specious. Talk about picking and choosing methodologies to forward an inaccurate narrative. The C300 Mk III kills in DR when measured properly.

That said, there are so many awesome options now. It really does boil down to individual need. Or whether you can afford an Alexa or not.

No need to argue with me about the virtues of the C300 II, as it is one of my dream cameras. I just wish the price were about $5-6K lower so that I could justify such a purchase. At $10K, it would be very much worth it.

Now the only "inaccurate narrative" is found in Canon's claims about the C300 II's 15-stop DR, which would place it beyond the Alexa! Cinema5d's Xyla test of the C300 II shows only 12 stops of DR. The FS7 was also found to have lower DR than Sony's claims of 14 stops:

https://***URL not allowed***/canon-c300-mark-ii-review-dynamic-range/

With the C300 II's internal 10 bit and external 12 bit RAW output, however, you should be able to recover much more of its highlights and shadows than the 1DC, so perhaps we could say that the C300 is a 14-stop camera with RAW recovery. There is no debating that at 2X the price ($16K) it is a better cinema camera than the 1DC and that it produces stunning results. Just because the 1DC can reach similar DR numbers as the FS7, C300 and C500 in a latitude test does not mean they are equal in other ways as I have stated. The same could be said for the A7SII DR, which comes in at just below 12 stops, but which has the worst codec of the bunch:

https://***URL not allowed***/ultimate-sony-a7s-ii-vs-a7s-test-difference/

Link to post
Share on other sites
11 hours ago, PabloB said:

The Canon 1DC may well have better dynamic range in clog but I believe the advantage may only be a stop at best with the 1DXii already having a 1 1/2-2 stop dynamic range advantage at base iso, and a further 1 stop gain shooting in cinestyle -1 ev htp, which adds up to around 3 stops. This compares favourably to the 1DC and its 2-3 stop gain using canon log from a weaker sensor. 

What I'd really like to see is a direct comparison. 

Where are you getting these DR numbers for the 1DX II and why would you assume that the 1DC has a "weaker sensor"? In fact, most of the tests we have seen show equal performance, if not an edge for the older sensor as in the DPreview article already cited on this forum:

http://***URL removed***/news/8090146652/canon-eos-1d-x-mark-ii-studio-tests

After all, the 1DX/1DC sensor has larger photosites than the second generation (6.95 vs. 6.65 microns) and can theoretically achieve a greater signal-to-noise ratio. Perhaps processing advancements and the gapless micro-lens sensor technology can make up for the smaller photosites of the 1DX II so as to even out the noise performance between the cameras, but it is doubtful that it could be much better.

From what I have seen, the 1DX II video footage performs exactly the same as the 1DC in neutral no matter what you do with those superficial picture profile settings, which means that it has 9 stops of DR and a lot of crushed blacks and nearly-clipped highlights:

The 1DC on the other hand will give you a very different look with better DR, smoother roll-off, and much lower contrast, not to mention the wider and richer color gamut that goes with C-Log when grading the footage in post:

 

Link to post
Share on other sites
4 hours ago, Kino said:

No need to argue with me about the virtues of the C300 II, as it is one of my dream cameras. I just wish the price were about $5-6K lower so that I could justify such a purchase. At $10K, it would be very much worth it.

Now the only "inaccurate narrative" is found in Canon's claims about the C300 II's 15-stop DR, which would place it beyond the Alexa! Cinema5d's Xyla test of the C300 II shows only 12 stops of DR. The FS7 was also found to have lower DR than Sony's claims of 14 stops:

https://***URL not allowed***/canon-c300-mark-ii-review-dynamic-range/

With the C300 II's internal 10 bit and external 12 bit RAW output, however, you should be able to recover much more of its highlights and shadows than the 1DC, so perhaps we could say that the C300 is a 14-stop camera with RAW recovery. There is no debating that at 2X the price ($16K) it is a better cinema camera than the 1DC and that it produces stunning results. Just because the 1DC can reach similar DR numbers as the FS7, C300 and C500 in a latitude test does not mean they are equal in other ways as I have stated. The same could be said for the A7SII DR, which comes in at just below 12 stops, but which has the worst codec of the bunch:

https://***URL not allowed***/ultimate-sony-a7s-ii-vs-a7s-test-difference/

Cinema5D's testing methodology is a joke. Every reputable source has rated the C300 Mk II as 14-15 stops in Canon Log 2 but with the caveat of noisy shadows. Apples-to-inaccurate-apples, what's that site's rating on the 1DC? I'd guess 10-11 stops. What's their rating of the C300 (a legit 12)? Or the 5D (8-9)? I'm guessing something inaccurate. Granted, the Alexa's current generation sensor and firmware has well surpassed 15 stops and Arri reps will privately confirm it, so the Alexa does have better DR and cleaner shadows than anything else. A friend just picked the Alexa Mini over the C300 Mk II and his DR tests just blow my mind. I dropped out of my math major so all the bits and stuff don't really mean anything to me, just not my strength. What does 14 stop with RAW recovery mean? How does a RAW recovery relate to a stop of DR? I never studied image formats.

Accurate comparisons of the C300 Mk II and Alexa show that the Alexa has a worse SNR in the highlights, slightly better in the mids, and far better in the shadows, but that despite very noisy shadows relative to the Alexa, the C300 Mk II (which is also far far cleaner at high ISOs than the Alexa, a common trade off) technically captures nearly as wide a spread of information. It's just not as pretty in the shadows, much more noise there. But a properly exposed image wouldn't emphasize this part of the curve in either camera, so the difference in noise would be minor, but the detail still present in most properly shot and graded material.

Of course, you're right, this is all irrelevant because what matters are your given needs. If you find noise in the shadows (even if there's still detail there) loathsome like the above test does, then you'd make the same subjective call, but call it what it is–subjective. But his cut off point is totally arbitrary and the test is pretty meaningless. Fwiw, the A7S (haven't used the A7SII) does have great DR (but also not up to Sony's claim of 15.3 stops or whatever), while the the F5 (haven't used the FS7, but same sensor) I remember having more than the C300, Red Epic, or any dSLR. Sony's 14 stop claims on its cinema line are fairly accurate, it's just a shame the image doesn't look better than it does. But then again, that's also a subjective call.

I also think dynamic range needs are overrated on lower end productions, and that the above video looks awful, but that's another story. 

Link to post
Share on other sites

What exactly does it matter what the DR is of a dedicated video camera when compared to a hybrid? C300, FS7 etc... are all permit cameras. Try taking one into your local neighborhood coffee shop to shoot a quick interview or scene and see how quickly you are ejected. Hybrid = stealth. That's kinda the point. 

Link to post
Share on other sites
4 minutes ago, DBounce said:

What exactly does it matter what the DR is of a dedicated video camera when compared to a hybrid? C300, FS7 etc... are all permit cameras. Try taking one into your local neighborhood coffee shop to shoot a quick interview or scene and see how quickly you are ejected. Hybrid = stealth. That's kinda the point. 

Not all of us are trying to make spy videos of our neighborhood coffee shop.

Link to post
Share on other sites
3 hours ago, Kino said:

Where are you getting these DR numbers for the 1DX II and why would you assume that the 1DC has a "weaker sensor"? In fact, most of the tests we have seen show equal performance, if not an edge for the older sensor as in the DPreview article already cited on this forum:

http://***URL removed***/news/8090146652/canon-eos-1d-x-mark-ii-studio-tests

After all, the 1DX/1DC sensor has larger photosites than the second generation (6.95 vs. 6.65 microns) and can theoretically achieve a greater signal-to-noise ratio. Perhaps processing advancements and the gapless micro-lens sensor technology can make up for the smaller photosites of the 1DX II so as to even out the noise performance between the cameras, but it is doubtful that it could be much better.

From what I have seen, the 1DX II video footage performs exactly the same as the 1DC in neutral no matter what you do with those superficial picture profile settings, which means that it has 9 stops of DR and a lot of crushed blacks and nearly-clipped highlights:

The 1DC on the other hand will give you a very different look with better DR, smoother roll-off, and much lower contrast, not to mention the wider and richer color gamut that goes with C-Log when grading the footage in post:

 

http://www.fredmiranda.com/forum/topic/1421137

1DX2_1DX_80D_7D2__DR.png

1DX2_D5__DR.png

"Weaker" in the sense that the new sensor is ADC design, and having less noise in the shadows.

Anyway, until we see some direct comparisons and see some footage that hasn't been sharpened either in camera or post we won't really know for sure. Numbers are in some cases meaningless.

 

Link to post
Share on other sites
39 minutes ago, PabloB said:

http://www.fredmiranda.com/forum/topic/1421137

1DX2_1DX_80D_7D2__DR.png

1DX2_D5__DR.png

"Weaker" in the sense that the new sensor is ADC design, and having less noise in the shadows.

Anyway, until we see some direct comparisons and see some footage that hasn't been sharpened either in camera or post we won't really know for sure. Numbers are in some cases meaningless.

 

This so-called "source" is a member on a fredmiranda.com forum who posted his own chart to the forum with the following caveat:

"Note: all the RAWs (except for the 1DX2) come from Imaging Resource ISO series (the 1DX2 have been received from another source which I won't disclose; please do not ask me for these RAWs)."

This is not a like a publication by DPreview with transparent examples for everyone to see. Moreover, I wouldn't put too much value in that chart until I could see the 1DX II examples for myself. DPreview's test of the 1DX II is fairly damning and is there for everyone to see.

Another test with examples showed that the 1DX II and 1DX are actually identical in performance:

 

Just like DPreview, he provides concrete examples to prove his point: the 1DX II offers no improvements on DR or noise performance. I also believe they are identical for reasons I stated above in that the 1DX II is starting out with a signal-to-noise disadvantage with its smaller photosites that it has to overcome through more advanced sensor tech and processing (Digic 6+).

Link to post
Share on other sites

From dpreview: Like the EOS 80D there's a big improvement in the camera's dynamic range. Canon's move to a design using on-chip analog-to-digital conversion ensures less noise is added before the signal is converted into digital values, meaning it's easier to distinguish between captured information and background noise. In turn, this means more malleable Raw files with more useful information available when you try to process them.

http://***URL removed***/news/8090146652/canon-eos-1d-x-mark-ii-studio-tests

Same chip ???

 

Link to post
Share on other sites
3 hours ago, Policar said:

Cinema5D's testing methodology is a joke. Every reputable source has rated the C300 Mk II as 14-15 stops in Canon Log 2 but with the caveat of noisy shadows. Apples-to-inaccurate-apples, what's that site's rating on the 1DC? I'd guess 10-11 stops. What's their rating of the C300 (a legit 12)? Or the 5D (8-9)? I'm guessing something inaccurate. Granted, the Alexa's current generation sensor and firmware has well surpassed 15 stops and Arri reps will privately confirm it, so the Alexa does have better DR and cleaner shadows than anything else. A friend just picked the Alexa Mini over the C300 Mk II and his DR tests just blow my mind. I dropped out of my math major so all the bits and stuff don't really mean anything to me, just not my strength. What does 14 stop with RAW recovery mean? How does a RAW recovery relate to a stop of DR? I never studied image formats.

Accurate comparisons of the C300 Mk II and Alexa show that the Alexa has a worse SNR in the highlights, slightly better in the mids, and far better in the shadows, but that despite very noisy shadows relative to the Alexa, the C300 Mk II (which is also far far cleaner at high ISOs than the Alexa, a common trade off) technically captures nearly as wide a spread of information. It's just not as pretty in the shadows, much more noise there. But a properly exposed image wouldn't emphasize this part of the curve in either camera, so the difference in noise would be minor, but the detail still present in most properly shot and graded material.

Of course, you're right, this is all irrelevant because what matters are your given needs. If you find noise in the shadows (even if there's still detail there) loathsome like the above test does, then you'd make the same subjective call, but call it what it is–subjective. But his cut off point is totally arbitrary and the test is pretty meaningless. Fwiw, the A7S (haven't used the A7SII) does have great DR (but also not up to Sony's claim of 15.3 stops or whatever), while the the F5 (haven't used the FS7, but same sensor) I remember having more than the C300, Red Epic, or any dSLR. Sony's 14 stop claims on its cinema line are fairly accurate, it's just a shame the image doesn't look better than it does. But then again, that's also a subjective call.

I also think dynamic range needs are overrated on lower end productions, and that the above video looks awful, but that's another story. 

The C300 II Trick Shot video is actually a pretty amazing demonstration of DR and color. It just shows you what you can do with C-Log2 even in a compressed 10 bit 4:2:2 codec. This is why it is so important that the 1DC has C-Log. It makes an 8 bit camera act like 10 bit, while a 10 bit camera acts like 12 bit. That's the magic of a good Log implementation.

Cinema5D's methods certainly leave something to be desired as they switch between talking about "visible DR" and "usable DR" in the same review or across reviews. They never posted a 1DC Xyla, but claim that both the C300 I and 1DC achieve around 10-11 stops of "usable dynamic range," without noting whether they employed C-Log (I would imagine they did):

https://***URL not allowed***/dynamic-range-sony-a7s-vs-arri-amira-canon-c300-5d-mark-iii-1dc-panasonic-gh4/

Personally, I find Hurlbut's over-and-under tests of the 1DC and C500 to be a far more valuable indication of DR. His series of tests with the C500, the Alexa, and RED Dragon are also very impressive in terms of what they reveal about the C500 as a camera that surprises with great color and low-light performance:

 

What I mean by "RAW recovery" is simply what happens when you go into Resolve and use highlight and shadow recovery to extract more DR in the RAW file that you otherwise would not see if you are just looking at the ungraded image. Xyla tests such as the one used by Cinema5D do no represent that extended range, so we could add at least an estimated 2 stops to their results for RAW-capable cameras like the C300 II. Note that there is nothing scientific here about how many stops you can recover. It's just an estimate, but that is certainly the way it works with the CinemaDNG files from my BMPC-4K when I take them into Resolve. I recall that Cinema5D rated the BMPC-4K at 9 stops whereas using Resolve you can lift the shadows at least one stop and recover highlights by at least a stop. Thus, the recovered DR is closer to 11 stops.

As for the Sony F5/F55, it outputs true 16 bit linear RAW . . .  no contest there. It's pretty much the king when it comes to RAW bit depth. Even RED R3D files offer only "16-bit precision" while Alexa and Ursa 4.6k use Log to achieve greater a bit depth that they don't actually record.

Link to post
Share on other sites
46 minutes ago, Kino said:

The C300 II Trick Shot video is actually a pretty amazing demonstration of DR and color. It just shows you what you can do with C-Log2 even in a compressed 10 bit 4:2:2 codec. This is why it is so important that the 1DC has C-Log. It makes an 8 bit camera act like 10 bit, while a 10 bit camera acts like 12 bit. That's the magic of a good Log implementation.

Cinema5D's methods certainly leave something to be desired as they switch between talking about "visible DR" and "usable DR" in the same review or across reviews. They never posted a 1DC Xyla, but claim that both the C300 I and 1DC achieve around 10-11 stops of "usable dynamic range," without noting whether they employed C-Log (I would imagine they did):

https://***URL not allowed***/dynamic-range-sony-a7s-vs-arri-amira-canon-c300-5d-mark-iii-1dc-panasonic-gh4/

Personally, I find Hurlbut's over-and-under tests of the 1DC and C500 to be a far more valuable indication of DR. His series of tests with the C500, the Alexa, and RED Dragon are also very impressive in terms of what they reveal about the C500 as a camera that surprises with great color and low-light performance:

 

What I mean by "RAW recovery" is simply what happens when you go into Resolve and use highlight and shadow recovery to extract more DR in the RAW file that you otherwise would not see if you are just looking at the ungraded image. Xyla tests such as the one used by Cinema5D do no represent that extended range, so we could add at least an estimated 2 stops to their results for RAW-capable cameras like the C300 II. Note that there is nothing scientific here about how many stops you can recover. It's just an estimate, but that is certainly the way it works with the CinemaDNG files from my BMPC-4K when I take them into Resolve. I recall that Cinema5D rated the BMPC-4K at 9 stops whereas using Resolve you can lift the shadows at least one stop and recover highlights by at least a stop. Thus, the recovered DR is closer to 11 stops.

As for the Sony F5/F55, it outputs true 16 bit linear RAW . . .  no contest there. It's pretty much the king when it comes to RAW bit depth. Even RED R3D files offer only "16-bit precision" while Alexa and Ursa 4.6k use Log to achieve greater a bit depth that they don't actually record.

highlight recovery in Resolve only works with DNG RAW, C300/500 RAW is logarithmic (Alexa's Log-C is also logarithmic), meaning it's just uncompressed C-log, there's no "extended range", Xyla tests reveals the entire available dynamic range the camera can produce. As for the BMPC-4K example, they tested with BMDFilm curve applied and also highlight recovery, so again there was no missing DR. Your eye ball certainly were not as precise as IMATEST software.

Link to post
Share on other sites
1 hour ago, Luke Mason said:

highlight recovery in Resolve only works with DNG RAW, C300/500 RAW is logarithmic (Alexa's Log-C is also logarithmic), meaning it's just uncompressed C-log, there's no "extended range", Xyla tests reveals the entire available dynamic range the camera can produce. As for the BMPC-4K example, they tested with BMDFilm curve applied and also highlight recovery, so again there was no missing DR. Your eye ball certainly were not as precise as IMATEST software.

You will notice that I made the distinction between visible and usable DR. The Xyla test reveals the entire "visible range," but Cinema5D's conclusions on DR are generally based on the "usable range," which can be reduced by 1-2 stops (especially with RAW). That's what I'm referring to as the recoverable DR that Cinema5D does not include in its DR ratings. You can see this on their Xyla tests as the difference between all the steps the camera records and the red line that marks what they accept as "usable" DR. They even lift the gamma to expose this expanded range that is accessible in post. On the C300 II, they counted only 12 steps (13 bars) of usable DR, whereas the visible DR range goes beyond that by several stops.

C300 II Xyla Cinema5D.jpg

Also, the C300 II was tested in its internal 10 bit 4:2:2 codec, where it achieved 12.3 stops of DR on IMATEST. However, using 12 bit RAW one would assume a slightly greater DR and a more usable range. They did not test the camera using its 4K 12 bit RAW output. They claim to have tried it out on 2K 12 Bit RAW but do not post the results as, according to them, it was no different from the 10 bit file.

Ultimately, a Xyla test is just another test with variables that can affect the outcome, including the acceptable level of noise or clipping that one interprets when counting steps. Using IMATEST just guarantees that all camera results are treated equally and can be compared against one another. IMATEST is not some higher standard of DR "precision."  It still has to make a quantitative assessment of acceptable noise and clipping levels. Yet there is no universal or scientific standard for noise performance. What is acceptable to one person may not be to another.

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...