Mokara

On 6/10/2020 at 3:14 PM, Rinad Amir said:

How do you know this have you tried R5 yet?

Have you compaired them both in action?

I bought my 1DXmiii last year before R5 was announced and i do 90% video main selling point for me was no fans best af on market and internal 5.5k raw 

As for calling people stupid.... knowing your reputation on this forum says alot.

The fact that the video specs are a significant jump from the 1DXIII does not give you a clue?

Lets have this conversation again in a month or two and see who is right. I will be expecting some groveling from you

36 minutes ago, Andrew Reid said:

4K/120p almost definitely won't be a regular recording mode for people, more an occasional shot.

4K/60p has to hold up well though, if they is also pixel binning, it depends on the standard of binning and methods used.

4K/24p from 8K readout will have rolling shutter issues, or else will have pixel binning like the higher frame rates, detrimental to image quality.

Optimal sensor for today is in the 24 megapixel region / 6K.

We are not yet at 8K without heavy compromises.

Don't get me wrong, I want this camera to be great.

At $4000 and with Sony A7S 3 on the horizon it has to be.

Assuming that 8K30p is the limit of what the camera can handle, then it is unlikely that 4K60p will just be an oversampled version of the same image since that would entail twice as much data. So that mode will almost certainly have significant compromises relative to 8K30p itself or 4K30p unless it is some sort of crop.

4K24p will probably not have RS issues. We can be pretty sure about that since the camera supposedly can handle 8K30p with DPAF, that means it should be a lot more capable than the 1DXIII in terms of processing power and bandwidth.

My guess is that a7SIII will be 4K only, no 8K and probably not a big oversample either. High spec for that camera will likely be 4K60p.

55 minutes ago, TheBoogieKnight said:

I'm sure it's on the way and just got held up in the post... 😄

I'm more worried about how useful the 120 is going to be given (unless it's a Quad Bayer sensor) you're probably gonna lose 4 stops of light over the 8k/30 with the skipping/binning and faster shutter speed. If this thing has a usable ceiling of 8,000 ISO (for example), you're down to 500 at 120p. Same as the S1 in 1080/120 I guess which needs lots of light and is often unusable indoors. I hope they add a crop mode. At least then you can get the same DOF without losing a ton of light.

 

No one will be shooting at 120 unless they are doing slow motion, and if that is the case you don't have a choice.

Why would adding a crop mode increase the amount of light? 

1 hour ago, Andrew Reid said:

I can confirm Canon did not give me one

In my opinion they are hiding something and probably the usual shills will gloss over the shortcomings.

Rolling shutter, pixel binning, over heating, the lot probably.

They probably only handed them out to folk who have a youtube presence for reviews, and to trusted pros for beta testing.

Your lack of a youtube presence probably bites you the butt in that regard. If you had an active channel with significant view you might get included in new release marketing activities more.

On 6/8/2020 at 2:32 PM, Andrew Reid said:

Are those stills frames or video frames?

I still think I'd prefer not to have to workaround 32ms rolling shutter on a £6000 pro DSLR but that's just me

But it can. Dual Pixel AF in 4K/60p in 1.3x crop mode.

The issue may be that the sensor goes into a different sampling mode in full frame 4K/60p and cannot remove the impact of Dual Pixel AF from the image (see 70D and Magic Lantern initial builds)

Crop mode uses a lot less data though. It is the amount of data that needs to be processed that is the issue. You only have so much processing headroom and when it runs out some sort of compromise has to be made.

On 6/9/2020 at 7:02 AM, Django said:

Clearly not stupid if you are a pro sports action hybrid shooter like the above. Probably the best tool on the market for that.

Let me repeat, "If you are shooting video IMO you would have to be pretty stupid to buy a 1DXIII over the R5."

Even if you are shooting pro sports with a hybrid, the R5 is still going to be the better option, because it IS a dedicated hybrid whereas the 1DXIII is a stills camera with video functionality.

1 minute ago, Video Hummus said:

I think for people that care, it is important if you have a YouTube channel to also take the time and energy to turn your videos into blog posts with links to resources and to your videos on YouTube. This has many benefits such as improved content quality over time and extended SEO to and from your video a and website.

I agree. If you are doing this for a living or for earning a substantial portion of your income you need to be taking a holistic approach to reach your audience. Some media are better suited for some types of messaging, while something different may work for others.

9 hours ago, Andrew Reid said:

I will have to push on with my YouTube channel double-quick pace. Really enjoyed doing the first video.

However, I fear the written word is dead and here's why.

Look how much space YouTube takes up on a typical Google search and how far it pushes the first result down the page for EOSHD.

The first result for "Iscorama GH5" is this forum. But it's barely noticeable with all the YouTube videos.

YouTube is a Google service and this is an abuse of a monopoly plain and simple and they should be taken to court.

It is not the result of a monopoly, but simply reflects that people are using video both for presenting stuff and researching material. It is much easier to get an impression of how something is done or what it looks like from video than from text.

It is sort of like the transition from silent movies to talking ones. The old way simply can't compete with the added breadth that video can offer.

On 6/7/2020 at 6:44 AM, Andrew Reid said:

Slashcam I would trust. Others not so much. DPReview don't even say how they arrived at the 28ms for GFX 100 or whether it was a scientifically measured test or just real-world impressions.

So it is not easy to compare rolling shutter readings from different source as they all use different methods.

In real-world tests people even all use different focal lengths and panning speeds.

The S1H in full frame 4K/24p is definitely much less than 1D X III in same mode.

The frustrating thing about the 1DX3 for me for the £6k, is that it clearly is capable of 16ms rolling shutter in 60p but cripples 24p with doubly worse rolling shutter.

Why doesn't it use 16ms all the time?

Because it has to focus as well, and it apparently can't do both at the same time.

9 hours ago, Django said:

Personally, my interest for 1DX3 completely died down after R5 was rumoured  shortly after.

As for GFX100 vs 1DX3 vs XT4.. well 3 different sensor size cameras at 3 completely different budgets.

Again personally, above the $6K line I'd rather be looking at dedicated cine cams.. especially if RS, NDs, XLR etc are of importance to you.

XT4 definitely seems like best bang for buck at the moment, kinda like the new GH5.

But if you're into FF (even if just for stills) you still may wanna invest in another hybrid system.

If you are shooting video IMO you would have to be pretty stupid to buy a 1DXIII over the R5.

4 hours ago, Kisaha said:

No useful info over there for some time now, ages; still watching some of Chris and Jordan though, these guys are just likeable!

They may be entertaining, but I have serious doubts that they are likeable, or at least what most people would consider to be likeable. Both of them come across as having a strong potential to be d*cks in person  

On 6/6/2020 at 4:06 AM, gt3rs said:

Please read a sensor specification sheet.

In sensor’s specification there is the scan time / fps for every mode, windowed and not. Some they mention in ms some they mention in fps with comma precision eg. 128.21 fps....  go check some sensor specs.

Even in single shoot mode the RS stay the same and you have all the time to process the photo buffer.

So A9 sensor is more than twice as fast to scan ca 6-7ms than the 1Dx iii ca 16ms. But Canon can process 60 fps almost full res 17:9 where Sony not so an educated guess the processing power of the 1dx iii is > A9 but sensor readout is exactly the inverse.

That is just the mode they set the sensor at. If they wanted to read at higher fps they could make it do that. The limitations come from the ability to process the data generated while at the same time doing whatever else the camera needs to do. Obviously you don't create a spec for the sensor that your processor can't meet. The two are designed around each other, and if the sensor is repackaged to be sold to third parties, those third parties get whatever specs were imposed on the design by the processors Sony had available.

The absolute speed the sensor can be read is kind of irrelevant if the processor has to do other things at the same time, which is what is happening with the 1DXIII. The sensor clearly can be read faster, since it does exactly that at 60fps, but not when the processor is doing AF at the same time (AF has a heavy computational load associated with it). AF processing has to be done constantly, with other data being read in-between, it can't just stop while the sensor is being read, if it did then AF response times would become much longer and you would get sluggish AF performance. So, a sensor frame read would go something like data...AF....data...AF.....data....AF etc, they are interleaved. Those spaces in-between during which AF functions are going on is why rolling shutter is 30+ ms. The frame read data is being spread out to allow AF functions to happen in timely matter instead of coming all at once. Basically AF gets priority over the sensor read. if you had fewer AF sites that needed to be polled and data processed, then you would have lower RS with the same sensor. The problem for Canon is that DPAF has a bucket load of AF sites that require attention, hence the RS, while other manufacturers are using conventional PD sites which are far fewer in number. Fewer PD sites = less AF processing overhead = less time needed to complete the frame read = lower RS. 

Not rocket science.

3 hours ago, gt3rs said:

This is why 1Dx III is not a mirrorless they don't have a sensor with that fast read out but the A9 sensor has other problem like less DR also no fast mechanical shutter (led panels are problem in stadium) and also have really limited video modes. Crippled or tech who knows?

Personally for sports I prefer OVF, I tried quickly a colleague A9 on a pro hockey match and I did not like the EVF. I had more problem following the action but maybe using it for a couple of games I would get used.... 

 

Probably a mix.... finally trying to do their best and not cripple but also they don't have a quantum leap on sensors tech that people are dreaming of..... thinking that Canon would come out with a FF 8k 30fps RAW with great RS DPAF with super sampled 4k 120fps DPAF at 4000 USD is  a day dream when their 6 months old 6K FF cinema camera cannot do 4k 120fps..... and no other competitor has similar spec even on the most expensive cinema line

I'm really curious to see what R5 will bring the good and bad...
 

The limitations are not due to the sensors, it is a result of what the processors involved can handle. And that in part is affected by the form factor of different models due to the different thermal envelopes involved.

I am pretty sure the R5 will outperform the 1DXIII when it comes to video. My expectation is that it will come close to C500M2 and C300M3 general performance, less the dynamic range those camera have.

3 hours ago, Andrew Reid said:

I think your guess will turn out to be right.

EOS R5

8K = very bad rolling shutter and unusable file sizes, impossible to edit

4K = pixel binned mush.

Have Canon really turned a corner or are they playing the same old games with new, higher numbers?!

If they are competing with Sony A9 then they really need to change the rolling shutter in stills mode too, it's unusable compared to the A9 in that respect.

The stacked DRAM on the A9 sensor allows for the fastest sensor readout on the market.

It's also why the 4K rolling shutter is so low on the RX100 IV / RX10 IV

RX100 IV? Are you sure you don't mean one of the later ones?

Sony cameras poll fewer data points for AF, so that is probably why they can get away with less RS. DPAF in theory has millions of AF points, but in practice no processor is going to be able to cope with anything near that number. Canon don't normally say how many AF points they are actually using for most of their cameras, but iirc in the R series they did mention that those cameras are using around 5k points. Or was it one of the M cameras, I forget exactly which one. I would guess the quality of data coming from DPAF is less than that from dedicated PD points and that is why they need more. More does impose a larger processing demand though, and that is probably what is eating into the sensor read headroom (the processor has to stop reading/processing data during AF polling, hence the RS - the more polling you have to do, the less flexibility you have when it comes to reading image data).

2 minutes ago, gt3rs said:

How could they get 1.5 crop with a 1-1 pixel read out with a 20mpix sensosr....

Extrapolation. Remember, the image you see is a processed image, not what the sensor sees.

12 hours ago, Snowbro said:

But doesn't the 1DX traditionally have dual processors, one entirely dedicated to AF? Even with a new generation processor in the R5, you're talking 8K instead of 6 and one less CPU, it would have to be a pretty huge leap in generational technology. It's always possible that Canon held back though & they rarely rectify it after release c200 ii anyone? Lol

The way Canon have usually segmented the FF line is as follows:

1D cameras: 2 current processors to handle image processing and 1 older processor to handle DSLR focusing/metering (three processors in total)

5D cameras: 1 current processor to handle image processing and 1 older processor to handle DSLR focusing/metering (two processors in total)

6D cameras: 1 current processor to handle image processing as well as DSLR focusing/metering (one processor in total)

The processing capability in the three models accounts for their relative capabilities. The 1D has more complex electronics so it costs the most, while the 6D is a cut down camera to reduce manufacturing cost, allowing it to be offered at a lower price. Along with all sorts of materials and mechanical compromises as well of course.

They deviated from that formula with the 1DXIII by having a single processor (not clear if there is a separate processor for DSLR focusing/imaging since they don't usually mention that one in their promotional material). My guess is that the Digic X in it is basically two Digic 8 processing units with some additional encoding unit all mashed into a SoC. 

5 hours ago, Alt Shoo said:

You’re stupid and are racist. 
 

Andrew you can ban me

Ok, don't let reality interfere with your fantasies then. Call me names if that makes you feel better but you are still wrong. Instead of listening to what other people tell you to think, why not go and look up the statistics yourself? It is all online if you care to look.

To help you out, here are some sources:

Arrests based on crime categories and demographics: https://ucr.fbi.gov/crime-in-the-u.s/2017/crime-in-the-u.s.-2017/tables/table-43

Deaths during arrests: https://www.bjs.gov/content/pub/pdf/ard0309st.pdf (keep in mind that 60% of arrest related deaths involve an underlying violent crime, so keep that in mind when looking at the first table and considering demographics)

It is pretty clear from the arrest statistics that deaths during arrests closely track arrests overall for all racial groups, with some deviations that can be accounted for by disproportionate involvement in violent crime, where most deaths stem from.

African Americans are disproportionately represented numerically overall because those global statistics don't account for socioeconomic differences. Most street arrests (these are the ones people get killed in) involve communities at the bottom rung of the socioeconomic ladder, and African Americans are twice as likely to be in that category (20% classified as living in poverty versus 10% for other groups). That is why arrests are 2X their share of the population, and, since deaths during arrests closely track arrests overall, why they appear to be being disproportionately being killed during these arrests. The reality is that the chance of this happening to them is no different than for anyone else being arrested. 

Btw, if you think that crime and arrests don't track socioeconomic status, just go to any prison and look at the sorts of people in there - you are not going to find a whole lot of professionals and white collar types in those places.

The reason why you are wrong is that you are looking at the entire population rather than just the lower end of the socioeconomic spectrum where most arrests are actually happening. 

People getting killed during arrests reflects a socioeconomic problem, not a racial problem.

14 hours ago, BenEricson said:

Yep. Exactly.

The cop is being charged with 2nd degree murder because he killed a man and you’re still trying to defend him, solely for the reason that he’s a white cop! Get the hell out of here with that. He had 18 prior complaints filed against him. 

I am not defending anything, I am pointing out the reality of policing. It would appear that this guy had no business being a cop, but on that day and under the same circumstances I doubt the victim being white or black would have made any difference to the outcome. You want it to be have made a difference because that fits your narrative, but reality is different. If the victim had been white, and all else about the event had been the same, the outcome would have been most likely a wrongful death against the city, it would not have made the news and the cop would have been fired but not charged. All the rest of what is going on right now is because people are trying to fit the incident to a narrative that conforms to their world view.

2 hours ago, Video Hummus said:

Eh, still sounds like a square peg with its edges roughly trimmed to be jammed in a round hole to fit. The shiny things on the 1DXIII still seem tarnished at closer inspection.

Im hoping the R5 will change that. I really hope that camera can deliver on the expectation it is building.

I still hoping Panny will announce a GH6 camera in the coming months. I see a lot of value in MFT and am sticking with the format regardless if I have a FF camera.

Based on the public specs for the R5 I would guess that it has a new processor, I am guessing a Digic 9, and it must be capable of better performance than the 1DXIII. They claim that it can do 8K30p with AF, that means there is enough computational power to handle it, quite a bit more than the 1DXIII. Worst case scenario would be 8K with RS of about 33 ms. Lower resolutions than 8K should be able to do higher frame rates with less RS.

On 6/2/2020 at 3:11 AM, Andrew Reid said:

I am pretty shocked by the rolling shutter behaviour reported by Slashcam.

16ms in 4K/60p

32ms in 4K/24p

What a mess Canon!

32ms is completely unacceptable. It's worse than an A6500.

What it means is that the camera is operating at the very limits of it's processing capability. That is why you see AF going at 60p. DPAF is computationally heavy, so in order to process a 4K60p stream the camera has to give up something, which would be the AF headroom. At 24 fps the camera can process the feed and do AF, but because it is being stretched to the limits the rolling shutter is bad. AF polling has to be continuous and that means that frame reads also have to be stretched out to accommodate that, hence the rolling shutter. 

3 hours ago, Andrew Reid said:

Photo by Fibonacci Blue

Will Smith said recently that racism isn't increasing - the difference is that it's being photographed and filmed - now we realise just how much of it there is. If it wasn't for the filming of George Floyd's murder by a policeman, it may never have come to light in public. All of us who have a camera, smartphone or online presence should keep this flame alive and shine a light on the truth wherever it confronts us. Important photos and videos must never be censored.

However when a long-time DPReview forum contributor tried to share his photos of the George Floyd protests in Minneapolis, he was immediately censored.

I have also been subject to censorship on the DPReview forum (my URL is banned completely) so I know what it's like to have your voice silenced from a community.

New blog post:

https://www.eoshd.com/photo/dpreview-censors-frontline-riot-photographer-photos-of-george-floyd-protests/

There was nothing to suggest that this particular incident had anything to do with racism, other than the police officer being white and the victim black. That in itself is not enough. Keep in mind that these two guys had worked together providing security at a night club for about three years, so they must have known each other fairly well. So there is more likely a personal issue between them if anything. The video footage I have seen of the arrest doesn't suggest that the officers were doing anything unexpected initially, nor were they angry or anything like that. It came across as just another day in the office. I know someone who used to be married to a police officer, and she would tell us that her husband spent most of his time essentially being a baby sitter for drunk people, and after a while the officers would get cynical and blasé about it. You get to see the bad ugly side of people and you see it all the time and I think that changes you to be less empathetic in time. Looking at that footage of those guys, it seemed to me like they were in that mode. Then there is a gap in coverage until we see the footage of the officer kneeling on the victim. He looked very calm about it, it is not like he was out of control. He was just parked there with his hands in his pockets, gazing outwards, like nothing much was going on. My guess is that something happened in that gap that provoked the final situation, what exactly isn't clear, but I suppose it will come out in the trial.

That said the racial demographics of people who die during arrest matches very closely with the number being arrested in total. Basically the people doing the crime or appear to be doing the crime and attract police attention in the process have a chance of being killed, and the color of your skin does not significantly affect those odds. Black males account for a disproportionate number of being arrested and/or dieing in the process compared to the overall population (~30% of arrests and of deaths), but when you consider that most street policing arrests happens in poorer areas and African Americans are twice as likely to be living in those areas compared to other groups, it is clear that neither arrests nor deaths are skewed significantly by race. It is more of a poverty thing than a race thing, and even then that is because there is more crime in impoverished areas. There are going to be individual cases that are driven by racism, but it is not the norm. The vast majority of police officers do not behave that way.

If you look up the statistics males account for about 75% of all arrests, and 95% of deaths during arrests. So this is something that mostly affects men, primarily those between 18 and 44. Also, about 60% of the deaths happen during arrests for violent crime, even though such crimes account for ~14% of all arrests. This is also the arrest group that is dominated by young/middle aged black men, which probably accounts for the higher death rate in that demographic. Interestingly, since both black and white deaths correlate with overall arrest rates, it suggests that whites are more likely to be killed than african americans when being arrested for non violent crimes since they dominate that category (in some cases well in excess of their representation in the general population). They have to be dieing somewhere for the overall stats to balance out. But you don't hear about that because if you are white then these events go mostly unreported and certainly no one stages any protests about it. It doesn't fall into the narrative of conventional wisdom.

I know this is going off-track a bit, but it is wise to take a lot of what is going on with a big pinch of salt. In stories like this you can be certain that some (and likely a lot) people are trying to manipulate you in the interests of their own agenda. I think most of the current protests are driven by frustration stemming from the lockdown due to covid-19, and this incident is acting as a surrogate to unleash those feelings.

22 hours ago, OliKMIA said:

They are just screwing their contributors, like they did many times in the past without any consequence. Their portfolio is soaring. Why would they stop milking the digital slaves?  Between Q1 2019 and Q1 2020:

"Image collection expanded 27% to approximately 330 million images."
"Video collection expanded 29% to approximately 18 million clips."

30% portfolio increase in a year! It's going to take more than a few grumpy folks on forums and quickly forgotten headlines to change their policy. As long as shareholders are happy, it's all good. After the 2020 Q1 release, the CFO explains that "we expect to produce significant earnings and free cash flow in 2020, allowing us to continue paying a quarterly dividend and reinvesting in our business." So most likely, the precious shareholders position is secured, perhaps with the shares taken from the contributors to offset the effects of the Covid-19.

And by the way, this company is financially healthy.  They have a good amount of cash and no debt "We are well positioned financially with $296 million in cash and no debt" according to their CFO. As mentioned before, net income alone is not a good way to assess a company situation. Their EBITDA is solid and increasing over the years, especially considering that they have no debt (the I part of EBITDA). The ratio are not amazing but not bad. See their annual reports.

There is no surge in operating expenses over the year despite massive portfolio increase which clearly contradicts the possibility of booming hosting and content review cost. The 2019 report says " Depreciation and amortization expense increased by $4.7 million as compared to 2018, to $40.4 million in 2019, driven primarily by the depreciation of our capitalized internal-use software. We expect that our cost of revenue will increase in absolute dollars in the foreseeable future to the extent our revenue grows". In other word, hosting fee is no hurting the EBITDA. Which explains why Shutterstock never tried to seriously limit or remove unsold assets as they don't represent a big cost.

 

The decision was made not on the basis of contributions, but because demand for images was falling while at the same time retained earnings was plummeting.

That is just the nature of the business nowdays, as more and more competitors come in they have to reduce costs in order to remain viable, and the only way to do that is to decrease the royalty paid to contributors. If they don't do that then competition will drive image price down and that will hammer their margins since their slice of the pie after operational costs is quite small (you can see this effect in their quarterly report where a 1% reduction in gross revenue resulted in a 47% reduction in retained earnings). All the competing companies will have to do the same thing in the end as well, or they will eventually go out of business.

Btw, they are attempting to limit contributions, it is being done by reducing the royalty to low sales contributors. That will drive them away. The contributors they want to keep are those whose images sell well, they don't want those who submit stuff that sells poorly.

38 minutes ago, MeanRevert said:

I thought the ZV-1 was based off the VII design since the VI doesn't have a mic port or Active Steadyshot?

It is a mark V with mark VII electronics and the latest firmware designs, but in a redesigned body. Some elements of the a6600 are included in the design.

VI might not have active steadyshot, but the mark V does.

56 minutes ago, ntblowz said:

The RX100 Va is only $100 more, and used one is cheaper.  I recently bought a used RX100VI for half price mainly as travel cam, and occasional slowmo.

 

Guess for vlogger it wouldn't matter much, but 1s is way too short for capturing slow mo.

The ZV-1 uses the latest AF hardware though, so it should be more capable than the VA when it comes to focusing. Not that it is a big issue though, all of the RX100 models from mark V onward focus good enough. It is only the earlier marks that have problems.

I guess the latest eye AF is good for a blogging camera though. Probably less useful for normal shooting.

5 hours ago, Emanuel said:

Of course, this camera can be as much cinematic as any other with similar specs and from talented handling ; ) However, I've seen very promising combo this time.

The same for their new smartphone BTW, except the 4/60p you can find on the Xperia 1 II sales they're somehow protecting to couple the rest of their cameras range, I guess.

Or overheating may be sorta issue from such 1" sensor and compact package, I wonder. And how can we ever see some similar sensor size going along a smartphone with heat dissipation?

Cell phones use variable frame rates, so in portions where a lot of work is needed to encode, they simply drop the frame rate. You can't do that with a fixed frame rate, the camera has to always be able to complete the frame no matter how fussy it is. The processors themselves could handle higher frame rates but they would overheat without external cooling of some sort, so there is a limit. That is determined by processor cooling, not by anything else. If a cell phone processor starts to overheat it just drops the number of frames being done. While that may be fine for cell phones, it doesn't cut it with regular video.

A camera with a variable frame rate will always be capable of a higher frame rate than one with a fixed frame rate for that reason.

17 hours ago, JurijTurnsek said:

Ok, let me rephrase my statement: Is anyone expecting a cinematic image from a fixed zoom lens 1" camera combinatoin that has been around for ages and we all know exactly how it performs even before we see a sample video?

Did you see anything in the promotional material suggesting that Sony is expecting these cameras to be used to produce cinematic masterpieces?

It seemed to me that they were promoting the camera as a blogger camera. I did not see anything about cinema production capability there. Or did I miss what you saw?