DxOMark gives RED Helium 8K sensor a record-breaking 108 score

ROFL
RED=Camcorder....
What planet are you on?
And it's an APS-H sensor.
Almost FF.
Have you seen the Arri dual pixel technology?
:-)

The RED uses temporal processing to reduce noise. It looks at more multiple frames to gather information, so it's not exactly a fair comparison against a ff stills camera.

Well if it gets the job done.
It gets the job done.
If it looks a several frames and merges them into one which is better who cares?
People don't scream about pixel shift cameras being fraud.
It's all about technology and the end product.
:-)

All RAWs are cooked.
Every signal off every sensor chip is processed.
1.3 is a lot closer to FF than 1.5 or 1.7.
Arri is cooking their RAWs based on your logic.
And they are anyway, since they use dual pixel technology.
Time to move on and enjoy the high DR that these cameras offer.
And then you try to change the subject to a camera that definitely cooks the output and say everyone likes it.
So one camera that is cooked is Ok for you and another that is cooked isn't Ok.
Sounds like a fanboy kind of comment.
Go buy an Arri and enjoy shooting!
:-)

you as well.....from wiki

"A revenant is a visible ghost or animated corpse that is believed to have returned from the grave to terrorize the living."

cheers

Then again these scores are probably for stills.

I agree with both of you, but think transparency, not subjectivity, is the issue. Compare DXO and Lens Rentals blog. Lens Rentals utilises an arguably more objective procedure, AND, are extremely transparent about their testing methodology in regards to the inclusion of subjective elements and weighting.

I think the opposite. DR an noise are good enough these days while color depth still has ways to improve. Especially on bayer sensors. Color can often end up looking fake or lackingwhen manipulating images.

samtan i know what you are saying but i dont think its so much color depth ...my 5D mk 1 has not got the best color depth ..it is 12 year old tech ..i compared lots raw files from a sony A6000 to the 5D ..DXO say the sony has a color depth of 24.1 bit vs 22.9 for the 5D, the ISO performance is the same for both but the sony has 2 stops more DR...so the sony must make better photos.. err no on landscapes if i try to emulate the nice blue sky from the 5D the sony will go noisy before that happens to the extant NR is used ..i never had to use NR at base iso on the 5D..i have found with many newish camera i cannot get the nice 5D colors with it looking over worked ....apparently the 5D has a very strong color filter array which newra cameras don't have so maybe thats it

"using for stills" ... really?

It can be a good price if you go cheap on the lenses. If you don't, it's another $40,000-$100,000 for a lens set.

@Olypent - Why would you spend that much on such a body and go cheap on the lenses? That makes no sense. If you're spending that much, you need it for cinema and you need cinema lenses. It's a bargain because of what it does, not because it's particularly within the range of a hobbyist.

lol, "within the range of a hobbyist"
I agree with stevo23.

it is ridiculous thinking that a hobbyist would by one to shoot it like a regular DSC...

@Zdman

The simple answer is that temporal noise reduction is being done in camera and applied to the RAW file (this makes sense, it's best to reduce noise before compression), thus the RAW converter makes no difference.

Even if you tell the camera to shoot a single frame, there's no way of knowing whether or not it takes multiple frames and averages them.

The quality of your electronics has a strong influence on shadow noise, but in the midtones it's 99.99% photon noise, which is determined by pixel size and quantum efficiency (proportion of photons converted into charge). DXO are reporting that for pixels of that size the SNR figures are beyond what's physically possible (i.e. they seem to be achieving more than 100% QE). This means they *must* be performing noise reduction somehow, it's more than just an assumption.

Oh, and DXO have published several articles demonstrating RAW cooking with Sony and Canon previously.

How do you know that? The camera has no mechanical shutter, so how would you tell how the exposure was achieved? It could be multiple frames and a clever algorithm that suppresses the noise floor.

Actually, I have a paper to appear at Electronic Imaging 2017 on a program called TIK, which can convert ordinary video into a time domain continuous format and from that render virtual exposures. A virtual exposure rendered for the exact same time interval as an actual frame generally has notably better DR and SNR. It wouldn't surprise me at all if RED is doing some level of temporal interpolation to enhance their raw frames....

I agree that the E-M1 II is really bad value but compared to the RED, it is affordable and justified for people using 43 lenses.

People need to stop talking about MFT having a "small sensor". Has anyone actually ever shot some footage on the GH4? Tried out the amazing autofocus and IS system from the EM-1 II? Have you given any thought to the fact that the upcoming GH5 will shoot 4k60p or 10-bit 4:2:2 4K internally for $2000? That's some amazing value for money. I'd very much rather buy a GH5 than a Sony A7 II for the same money.

Bgpgraebner. who are you to tell me i can't complain about a camera system i would never want to buy, nor have invested time learning the best ways for my type of shooting. and all the lenses i have invested in that system. why do you think i have an account here, to learn something? i can't. i know everything. and if i don't know it, nobody needs to know it.

If you can win awards for your film shot with an iPhone 5 (can't remember the name of the movie) then you've got bigger problems than the equipment you're using if you can't make decent footage with a GH4.

The GH4 makes glorious looking 'film', don't knock it, if you've not made a movie people care about, it's you, stop finding excuses.

Also, look at the total system cost, a GH4 + many decent lenses is $15K, at the most.
A RED with a complete set up is $100K+

People here talking about how expensive these cameras are.. meanwhile I was just handed a RED 6K with gaffa tape instead of a lens mount.

Why would someone not spend $2000 on a micro four thirds camera? I recently spent more than that amount on a 5D Mk III, for no better reason than to entertain myself (not a protog). I could have as well spent the same money on the Olympus, had I been so inclined.

And they gave their own camera two scores, one for the regular raw files and one for the "SuperRAW" files. They were fully transparent regarding the fact that the latter received such a high score because of temporal noise reduction.

I have to mostly agree. The DxoMark general measurement graphs are slightly useful, but score not so much. So if Nikon and DxoMark partnered up and made a Nikon D3400 DxoMark edition, it would out score this $40k+ Camera with SuperRAW.

@panther fan, please read the article before trying to explain it to other people.

@Zdman what they said right after is also relevant.

"Whatever noise reduction system RED employs creating the RAW images from the Helium sensor, its presence means that we aren’t measuring just the RED sensor, so its results aren’t directly comparable to those from camera sensors we have tested from other vendors, whose RAW results come straight from the sensor with no prior noise reduction processing."

Also:

"Since the Helium has a much higher SNR than the Sony sensor despite its smaller pixels, it would need a significantly higher QE to achieve those results – SOMETHING BEYOND THE COMMONLY UNDERSTOOD LIMITS OF PHYSICS. This in turn led us to look at noise reduction processing as an explanation for the Helium sensor’s very-high QE."

Yes a wonderful camera, the absolute top in image quality :)

I disagree. My Mamiya 7 loaded with 220mm Velvia 50 is the absolute top in image quality. Also, my Montblanc cigar size 18 carat gold nib founain pen is the best fountain pen in the world. I am a little old-school.

I cannot agree or disagree with your claim of the Mamiya 7 since I never shot with it. What I can tell you however is that your claim of the "cigar size" Montblanc being the best fountain pen in the world is clearly exagerated. The Montblanc that your are referring to is probably the 149 which although a high end pen is not at the top of their offerings. I've enclosed a photo with a large number of high end pens (including Montblanc) which illustrates some of their really top end limited edition pens such as the Agatha Christie, Wilde, Dumas and Hemingway. The 149 is akin to a Rolex Submariner not a Patek Philippe. https://www.dpreview.com/galleries/2106290526/photos/3415393/_dsf3286

Pronounced with a properly snappy je nes sais quoi, Patek Philippe sounds like a rail car going on down the line.

But that score does not exist in DxOMark. They deleted their own score without any reasons.

this pen discussion has to be one of the most pretentious d**k measuring contests I've ever read on a forum, and I've been on forums for 20 years.

@jose garcia - My comments about the Mamiya 7 and fountain pen were not meant to be taken serious. I was trying to make the point that was is "best" is sometimes very subective.
I agree with your comment about the fountain pen. I have a number of high-end fountain pens from Germany, France and Italy. The Montblanc is not the best, but it is worth twice what I paid for it in 1982. Can anyone say that for a digital camera?
Good light and good writing to you.

I don't think the multi-shot or sensor shift results should be used for general performance scores.

General performance scores should reflect performance available in any condition. Blending multiple images is not suitable in all conditions.

It would be interesting to see the performance of the Pentax K-1 with sensor shift used but it should be an addendum or footnote to the general score.

Why are you making stuff up? In Hollywood Canon is used more than red? Put down the crack pipe.

http://www.red.com/shot-on-red

Perhaps DxOMark made a mistake in measuring the 645Z? Very sensible explanation of its non-release.

Yeah for a long time. DxO deleted Pentax 645Z's score and never came back.

Yeah, I'm sure it was probably just a mistake. Makes perfect sense.

Since when has a7RII started shooting 8k video?

I'm not sure why they have a prototype. The helium sensor was launched by red about 4 months ago and they have shipped plenty of units to plenty of customers including myself.

DXOmark explains that the Red Weapon results are likely due to temporal noise reduction. If not, then I say the results show something has gone wrong with the measurement. I don't think Red has repealed the laws of physics.

@GabrielFFontes

We're actually far closer than most people realise. High ISO noise is pretty much entirely photon noise these days, and many cameras are achieving greater than 50% quantum efficiency. The RED appears to be getting greater than 100% QE, hence breaking the laws of physics - you can't convert more than all of the incoming photons to charge!