Just got our Dragon

[Marc Ruiz]

I was asked about the Anamorphic so I figured I would share it here as well...

My initial effort was to do a Epic v Dragon test but I realized that the Epic had a motion mount on it so that is a test I will be doing later this week.

The match is obviously an inconsistent light source but that said even it's still a fun test. I think previous tests shown by others with candles are a more constant light source.

That said here is another test with a match.

The subject is in a basement with 2 candles lit in the background so the bokeh of the anamorphic can be seen.

RF Lomo 100mm at 3.2 (Wide Open For This Lens)

R3D Link: https://www.dropbox.com/s/zikjdrg4vqy844r/A001_C007_0115XJ.0000432F.R3D

 

[Matt Ryan]

Still looks clean for 2000ISO (and keep in mind it's RAW and doesn't use on board noise reduction). By judging the images without a direct comparison, I'd say from past experience this looks like ISO 800 noise on MX. Obviously a direct comparison would help support or disprove that hypothesis... Thanks for the samples Marc.

I'm curious to see a comparison of Dragon + Neat Video vs. Canon C500 at 3200ISO and see how they compare. That would be interesting to see. Obviously, Dragon would need to be down sampled to 4k to match the C500. I'm very curious to see this...

The only thing I have been disappointed by is the OLPF flare from Dragon but since Jarred said they are already working on a fix, all is good.

 

[Marc Ruiz]

I wonder if that flare would have different qualities if I put on one of our motion mounts. That said, this lens is slow at a 3.2 and the motion mount eats up over a stop.

 

[Matt Ryan]

Interestingly enough, these C500 low light shots don't look as good or as "brightly exposed" as the Dragon shots we've seen online. So much more detail in the Dragon shots and what looks like a ton more dynamic range. What's even more crazy is the Dragon shots are a single match vs. multiple candles for the C500. Skip down to the bottom for some 4K C500 candle light tests at varying ISO's. I'd venture to say that Dragon (down sampled to 4K) is cleaner at 2000ISO than the C500 at 800 ISO. What do you guys think by just looking at the samples? Again, nothing definitive without a side by side, but until then, this will help.

http://prepostconsulting.com/en/english-c500-and-canon-raw-my-somewhat-technical-review/

C500, 800ISO: (right-click and open in a new tab to view full size. The 100% crop gives the best idea)

100% crop:

(Balanced and Noise Reduction)

Dragon match/candle footage curve only, no noise reduction.

Straight Redlogfilm at 2000ISO, no corrections:

 

[Robert Ruffo New]

I'm inclined to agree that the Dragon 2000 ISO looks on par or better than the MX 800.

I have another Epic still waiting for the upgrade I am tempted to do a side by side comparison.

Please, please do this.

 

[Nick Morrison]

Interestingly enough, these C500 low light shots don't look as good or as "brightly exposed" as the Dragon shots we've seen online. So much more detail in the Dragon shots and what looks like a ton more dynamic range. What's even more crazy is the Dragon shots are a single match vs. multiple candles for the C500. Skip down to the bottom for some 4K C500 candle light tests at varying ISO's. I'd venture to say that Dragon (down sampled to 4K) is cleaner at 2000ISO than the C500 at 800 ISO. What do you guys think by just looking at the samples? Again, nothing definitive without a side by side, but until then, this will help.

http://prepostconsulting.com/en/english-c500-and-canon-raw-my-somewhat-technical-review/

C500, 800ISO: (right-click and open in a new tab to view full size. The 100% crop gives the best idea)

100% crop:

(Balanced and Noise Reduction)

Dragon match/candle footage curve only, no noise reduction.

Straight Redlogfilm at 2000ISO, no corrections:

Fantastic break down Matt, thanks. Yeah Dragon's lowlight capability looks very robust. So usable at high ISOs. The c500 footage doesn't appear to have comparable range or detail (but perhaps other tests can show otherwise?).

 

[Matt Ryan]

Fantastic break down Matt, thanks. Yeah Dragon's lowlight capability looks very robust. So usable at high ISOs. The c500 footage doesn't appear to have comparable range or detail (but perhaps other tests can show otherwise?).

Exactly Nick. I'd love to do a side by side if possible between the C500, F55, Alexa, and Dragon.

 

[Matthew Scott]

Interestingly enough, these C500 low light shots don't look as good or as "brightly exposed" as the Dragon shots we've seen online. So much more detail in the Dragon shots and what looks like a ton more dynamic range. What's even more crazy is the Dragon shots are a single match vs. multiple candles for the C500. Skip down to the bottom for some 4K C500 candle light tests at varying ISO's. I'd venture to say that Dragon (down sampled to 4K) is cleaner at 2000ISO than the C500 at 800 ISO. What do you guys think by just looking at the samples? Again, nothing definitive without a side by side, but until then, this will help.

Interesting that no-one has mentioned my test I did a few months back on the Scarlet...

https://vimeo.com/79066426

https://vimeo.com/79196056

 

[Nick Morrison]

Hey Matthew did you denoise the Scarlet low-light clips in anyway? Love the clean blacks.

 

[Eric Haase]

Dragon match/candle footage curve only, no noise reduction.

Straight Redlogfilm at 2000ISO, no corrections:

Holy moly these are a perfect case study of the OLPF flare problem. I hope a fix is imminent.

 

[Matt Ryan]

Interesting that no-one has mentioned my test I did a few months back on the Scarlet...
]

Saw that in the other thread and thought it looked good. This thread is specifically about Dragon and it's performance in low light at 2000ISO. Some seem to think that Dragon isn't a leap in noise and sensitivity over MX, and I disagree. Thus the samples.

 

[Matt Ryan]

Holy moly these are a perfect case study of the OLPF flare problem. I hope a fix is imminent.

Jarred stated in another thread that they were already working on it and would be taken care of soon.

 

[Matthew Scott]

Hey Matthew did you denoise the Scarlet low-light clips in anyway? Love the clean blacks.

Hey Nick, thanks for commenting. There's no denoising here. I'll upload the R3D's if anyone is interested. It was very interesting for me comparing my 4k Scarlet R3D's with Phil's 6K dragon ones. To me, it's not a massive world of difference like everyone is saying. Yes, dragon is better (it bloody should be!) and probably the best on the market, but compared to an MX, is it mind blowingly amazing or game changing at 24p? Despite what the forum gods say, I don't see it, but then again, I'm not a hollywood DP (not yet anyway hehe) and I don't have 10K+ posts behind my name. I guess my point is, the MX still has plenty of life left in it and when compared to the Dragon, the difference isn't as big as the hype around here suggests. Will I upgrade? Probably...and I'll love it.

 

[Matthew Scott]

Saw that in the other thread and thought it looked good. This thread is specifically about Dragon and it's performance in low light at 2000ISO. Some seem to think that Dragon isn't a leap in noise and sensitivity over MX, and I disagree. Thus the samples.

Wait....then why the c500 shots? How sepcific do you want to be? My low light Scarlet test, with a match, (which Phil seemed to light the same way) is a great comparison of MX vs dragon in similar light. That's why I thought it was interesting for you not to include it in your comparisons. A "leap" in performance is one thing....for sure it's better in low light once downsampled, and for sure it has better DR. I personally found it great to compare the R3D's. It is very telling.

 

[Matt Ryan]

Wait....then why the c500 shots? How sepcific do you want to be? My low light Scarlet test, with a match, (which Phil seemed to light the same way) is a great comparison of MX vs dragon in similar light. That's why I thought it was interesting for you not to include it in your comparisons. A "leap" in performance is one thing....for sure it's better in low light once downsampled, and for sure it has better DR. I personally found it great to compare the R3D's. It is very telling.

Definitely, but you also used more light and a faster lens than Phil's test (his blue backlight was from a bounced flashlight and shot at t2). So it's kind of hard to compare apples to apples that way. I've been looking through my old Epic MX footage and I see more noise at 800 in some of the lit stuff than the Dragon at 2000ISO.

I'm not looking through threads on Reduser to compare MX to Dragon, that's for the doubters to do. I'll test myself when I get my hands on a Dragon, which I suggest everyone do. Plus Mark at OffHollywood is supposed to do an MX/Dragon side by side.

 

[Noah Yuan-Vogel]

Fantastic break down Matt, thanks. Yeah Dragon's lowlight capability looks very robust. So usable at high ISOs. The c500 footage doesn't appear to have comparable range or detail (but perhaps other tests can show otherwise?).

Seems a little crazy to draw any conclusion from a comparison of those frames. Completely different kinds of light/fire at fairly different distances from the subject at totally different camera settings processed (or not processed) very differently...

 

[Matt Ryan]

Seems a little crazy to draw any conclusion from a comparison of those frames. Completely different kinds of light/fire at fairly different distances from the subject at totally different camera settings processed (or not processed) very differently...

Very different indeed. But interesting none the less. You can't deny that there is more light in the C500 shot....

 

[Noah Yuan-Vogel]

Very different indeed. But interesting none the less. You can't deny that there is more light in the C500 shot....

Sure I can deny that. Yes there is more fill light (the Dragon ones mostly have none) but that just further complicates any direct comparison. The key side, however, could easily be 4-5 stops lower since most of the Dragon shots are unobstructed matchsticks within 3-8in from subject's face whereas the C500 shot the candle looks closer to 1ft away (could be 2-4 stops right there), and the candle flame may be partially obstructed by the candle's edge and who knows how bright or large the flame is and how it compares to the flames in the other shots. Looks like a short wick and the wick is probably partially drowned in wax... I'm not saying it is or isn't brighter or darker but with sources like that you really don't know and most likely they are not even close comparisons. Lights so dependent on what and how much is combusting at distances so small that a couple inches drops exposure several stops, it just seems a little silly to even put them next to each other.

 

[Matthew Scott]

Definitely, but you also used more light and a faster lens than Phil's test (his blue backlight was from a bounced flashlight and shot at t2). So it's kind of hard to compare apples to apples that way. I've been looking through my old Epic MX footage and I see more noise at 800 in some of the lit stuff than the Dragon at 2000ISO.

I'm not looking through threads on Reduser to compare MX to Dragon, that's for the doubters to do. I'll test myself when I get my hands on a Dragon, which I suggest everyone do. Plus Mark at OffHollywood is supposed to do an MX/Dragon side by side.

Yeah man, that's why I said that it's similar lighting and it's still a great comparison, probably better than looking through your old MX footage since they are so similar. But as for this:

"I'm not looking through threads on Reduser to compare MX to Dragon, that's for the doubters to do."

That grinds my gears....I look through threads on Reduser to compare MX / Dragon all the time! And I'm sure thousands of others do to....not because we're doubters, because we're explorers. Instead of basing big purchase decisions on whatever REDUSER tells us is fact, it's wise to ask questions, question the facts and make comparisons with RAW clips. I'm sure your test will conclude what you already believe and I'm sure many will support you. I don't mean to sound like a jaded mean head, sorry if I come across that way. I just don't see such massive gains in low light sensitivity...I know the dragon upgrade offers much more than that, just pointing out my findings.

 

[Matt Ryan]

Yeah man, that's why I said that it's similar lighting and it's still a great comparison, probably better than looking through your old MX footage since they are so similar. But as for this:

"I'm not looking through threads on Reduser to compare MX to Dragon, that's for the doubters to do."

That grinds my gears....I look through threads on Reduser to compare MX / Dragon all the time! And I'm sure thousands of others do to....not because we're doubters, because we're explorers. Instead of basing big purchase decisions on whatever REDUSER tells us is fact, it's wise to ask questions, question the facts and make comparisons with RAW clips. I'm sure your test will conclude what you already believe and I'm sure many will support you. I don't mean to sound like a jaded mean head, sorry if I come across that way. I just don't see such massive gains in low light sensitivity...I know the dragon upgrade offers much more than that, just pointing out my findings.

I like you Matt, but you're missing my point. You can compare all you want with other images, but without physically testing them side by side, same settings, same lens, at the same time, no samples will really tell the whole tale. Thus my comment about not searching Reduser for stuff to compare it with... I don't listen to what others here on this site recommend, I test for myself. With Dragon I'll continue that trend. My beef is with everyone who see's somewhat similar shots from MX footage here and say Dragon doesn't look like a huge improvement... That's like seeing a shot of a Summilux C at a 2.8 and saying, I don't think it's any better than a Rokinon from image samples I've seen at a 2.8. It just doesn't make any sense to me at all. If you still have a problem feel free to PM me but let's not derail the thread anymore.