Jump to content

OliKMIA

Members
  • Posts

    628
  • Joined

  • Last visited

Posts posted by OliKMIA

  1. 1 hour ago, KnightsFan said:

    Agree with what @OliKMIA said but wanted to be a little more cautious as well.

    While the flu kills a lot more people total, its mortality rate is on the order of 0.05% in the US. That's a difference of 40x. Those numbers are the US flu rate vs. the primarily China coronavirus, so it's not apples to apples, but the coronavirus is an order of magnitude more deadly. The flu is global, the coronavirus has really only hit a single city so far, with a few isolated cases escaping.

    The rate of spread is extraordinary. The SARS outbreak infected 8,000 people total over six months. This has hit 10x that in two months. And this is even with Wuhan being in complete lockdown with extraordinary measures undertaken to limit the spread. If a similar outbreak happens in New York, a similar lockdown could devastate the global economy.

    Also agree with that. On the other hand, the Flu is is a well known virus with available vaccine and organized disease management which also explains the low mortality rate. On the contrary it also justifies the containment measure taken against the Corona virus. If anything, the SARS was not very worrisome. It appears that the progression of the virus is now on the decreasing phase in China, thanks to the vigorous actions taken by the authorities over there... To be confirmed.

    39 minutes ago, IronFilm said:


    40x more deadly than flu.....  but only IF you believe the figures coming from the Chinese government!

    I'd take that with an extremely BIG grain of salt. 

    For all we know, it might be many times worse. 

    https://www.newsshooter.com/2020/03/01/will-nab-2020-be-cancelled-due-to-the-coronavirus/

    The Chinese government is not a model of transparency but they remember how bad the 2003 SARS situation was managed and all the health organizations agree that there is a huge improvement this time (albeit, not perfect). I don't really believe in massive cover up for two reasons:
    1. The information always comes out, especially in an age of social media despite the digital great wall in China
    2. But more importantly, the virus is now hitting strong democracies like South Korea and European nations. These countries won't be able to hide anything if bodies start to pile up in complete contradiction with official statements. So even is China lied, the mortality rate in the other countries will be public knowledge very soon.

    According to public health specialists, the real challenge will come when poor countries will be impacted because they have very little resources to handle that and unlike Ebola, the world is already busy enough with domestic situations. To make matter worse, the limited supply of protective equipment and medicine coming out of China won't be prioritized to Africa or south Asia...

    That's all for me. I'm not going to step any further in this discussion because I'm not a doctor or a virologist and my point of view is pointless beyond this general statement.

  2. The facts behind the panic:

    • Corona virus death toll so far after two months of existence: 3000 persons
    • Mortality rate is about 2-3% and old people are the most vulnerable, about 80% of deaths recorded were from those over the age of 60, and 75% had pre-existing health conditions including cardiovascular diseases and diabetes. (source: WHO)

    What should really be a cause of concern:

    Leading causes of death in the US in 2017:

    • Heat disease: 647,457 deaths
    • Cancer: 599,108 deaths
    • Unintentional injuries / accidents: 169,936 deaths

    So we may want to take a huge chill pill. Eat healthy, double check left and right twice before crossing the street, put the seat-belt on and stop worrying about a small pandemic of a not so dangerous virus. That being said, the public safety measures and large gathering restrictions are totally justified to prevent the spread of the disease, Photokina or the Olympic Games are clearly not a priority at this point.

  3. 15 hours ago, sgreszcz said:

    That looks cool, but very $$$.  I found an open-box Afidus ATL-200 which I'm going to test now.

    I've got one, but not had great day-to-night or night-to-day weather to try it with.  Might attempt through a window.


    Personally, I think that the price is fine considering the technology and the niche market this is. You got a weatherproof ready to go package with 2 MFT sensors, a communication module (Wifi and Sim card slot), a huge battery and a solar panel. There is an app to control all this an upload the images remotely.
    It's not cheap if you just want to have fun but for professional production I would pay this off on the first job I got. You can charge premium for long terms time-lapse.

    But let me review this unit first, right now this is too early to tell and I need to make sure this things works as it's supposed to.

    As for the day to night, I quickly tried the timelapseplus and it works fine most of the time. Nowadays the metering system of the latest Sony camera is really good and it would be unnecessary on those. Personally, I tend to do my holy grail manually and post process with LRTimelapse.

     

  4. Looks like the guy just wanted to do a stunt and play with rockets in the first place. The flat earth thing seems to be an excuse.
    As dumb as he and his team were, they knew that a steam rocket couldn't go very high. A small Cessna ride at the local airport would have brought him much higher than this rocket. But by playing stupid game, he won stupid prize...

  5. I did a couple of DIY solution back in the days with Gopro and battery pack but I recently got this for review (product sent to me by the manufacturer):

    https://enlaps.io/produit/tikee-pro2/?lang=en

    I'm testing it at the moment but so far I'm pleased with this product and it seems to be a good value. It's been running for 2 weeks on my balcony day and night taking a fair amount of rain already. The battery is still very high thanks to the solar panel. The next step will be to  check the app and marge the pano pictures from the two cameras.

     

     

  6. I reviewed the EVO I for fstoppers. Bottom line: good hardware, some nice features but the gimbal performance was horrible with constant and severe horizon drift (roll axis). I still see some drift on the video posted above. It also suffers from multiple glitches. The colors were a bit "flashy". It had some potential but they need to fix that.

    As for the EVO2, I'm curious about the quality if 8k on such a small sensor with a very limited bitrate (120mps). There is much sensor an h265 can do and the lens won't exactly be Sigma art... Looks like a useless gimmick to me. It's advertise as a digital zoom feature. Perhaps because they can't mount a proper optical zoom or don't have a good gimbal tech to stabilize it?

    I'll try to get my hand on the EVO2 for review. We really need more competition on this field.

    PS: the support chat is very good and responsive with Autel. Big difference compared to DJI...

  7. 48 minutes ago, Michi said:

    So then it‘s a question of heat generated by the sensor?
    Don‘t get my wrong, I‘m no  engineer, but it does‘t seem logic that a processor has more work to do when processing information from a bigger sensor compared to a smaller one. 32mp is 32mp... is it not? 
    I don‘t know about the S1H but on the C200 the fan does only start after 25 or so minutes of recording. And it has two processors inside... 

    With the R5 having a newer processor and probably a recording limit, 8k does not seem entierly impossible...

    Correct, mostly a sensor issue even though processor plays a role but 8k is similar to manage whether it comes from tiny or large sensor.

    Now, I don't have any info and I'm just going to wait like everybody else. Hoping for the best, I'm still deeply invested in the Canon ecosystem so it would be a real win for me.

  8. 46 minutes ago, Michi said:

    What is the physical reality of 8k?
    Consider there‘s now a Samsung Smartphone shooting 8k and working Sharp prototypes with 8k MFT-sensors. Both (I assume) with smaller form factors than the R5. 
    There probably will be a limit for 8k recording: 30min best case, maybe less, if we look at the 1DX3. A crop is not really possible since there aren’t many (if any) pixels left to leave out. 

    The way Canon promoted this Camera only months after the 1DX3 anouncement is a clear indication for more than a „sticker on the box“ feature. 
    Nonetheless: those hoping for 8k RAW/ProRes will have to prepare for disapointment. 
    Maybe 10bit with H.265 compression but I‘m not betting any money on that. C-Log is a given. The deciding question for me: what will the 8k feature mean for the 4k modes...? 

    Doing 8K on FF or small sensor is a whole different story in term of thermal dissipation. Heat is the physical limit at the moment with the current technology (CPU, memory, etc.). See the fan on the SH1 and most pro camera.
    Agree with the crop, 8k video shouldn't crop much unless they go way beyond the "native" 8k resolution (need a minimum of 40mpx fora  3:2 ratio sensor to output 16/9 8k).

     

  9. 7 minutes ago, Django said:

    It's funny Canon always seems to be on the "Guilty until proven innocent" side. I mean I know it's hard to swallow for some because Canon has had a poor rep in the past, but please take a look at the 1DX3 specs. 

    To me it's a clear indicator Canon have changed strategies, and seriously upped the ante when it comes to video specs.

    Pretty sure they're on a mission to sweep up the hybrid floor and take back their lead position.

    In any case, this seems a solid camera on both photo & video side even with the unknowns .

    True, it's all about this company track record. The 1DX3 is a huge body compared to this R5, easier to manage thermal dissipation on this, and I see it more as a B-cam option for the C500mk2.

    Also, I'm not anti-canon. All my still cameras are Canon and I recently purchased a C200 (great cam, the 8 bits footage is solid). If Canon delivers good video specs on the R5, I'll get one or two for sure.

  10. 16 minutes ago, KnightsFan said:

    well canon officially announced that it will shoot 8k, so that's reliable at least. Hopefully it will be at least 24p. As far as i can see, there is no official word about whether it has 4k60 or bit depth or chroma subsampling, but i think we should assume 8 bit 420 for 8k

     

    14 minutes ago, Django said:

    @OliKMIA Dude, you didn't get the memo? It's official, Canon is bringing the first 8K MILC to the market. Brace yourself: it's happening.

    I get that some sort of 8K is coming, I also see the track record of this company and the physical reality of 8K. Plus, I note that more advanced companies are having issues to do high res video on FF sensors. Finally I can see the size of the body in relation to a 24-105 f/4 lens.

    Now, if we get some sort of "useless" 8k (eg. no AF, limited burst, etc), I don't think that it will massively change the game for video people like us. I'd rather have good 4k60 with all the nice codec and features (peaking, zebra, etc.) than limited 8k for the sake of it. I have this bad feeling that we are falling for a marketing trap when the real stuffs that matter are completely unknown. That being said, it seems to be a solid camera on the still side.
     

  11. 5 hours ago, Django said:

    A7S3 Dev Team: Ok, back to the drawing boards folks.

    Seriously, I don't see how Sony or anyone is going to compete with this in the near future.

    I'm just glad I stuck with them throughout the years (which will be regarded as the 'dark ages' lol).

     

    4 hours ago, Andrew Reid said:

    Seems Canon is keeping some stuff close to their chest to avoid Sony being able to outgun with A7S III.

     

    3 hours ago, Andrew Reid said:

    X-T4 "dead to me"

    A7S III has to "step over FX line"

    Yep, it's not looking good for the competition.

    The brand name Canon and lure of 8K will be too much for many people, but the best thing for me about Canon's great awakening is that hopefully they will get all the basics right... Like full frame 10bit 4K with no crop, great colour science, the classic cinematic 1D C look in C-LOG, Dual Pixel AF and very good IBIS, all in a nice mirrorless body which has a much better and more "pro" control layout compared to the silly toy one that came before it.

     

    I think that we are headed for a massive disappointment. There is absolutely no reliable info on this camera and we already expect that a small canon body will leap frog the entire competition? It makes zero sense, even the S1H with its brick size and embedded fan can't do crop-less 4k60. How would Canon manage to do 8k24 in a Rebel/R type body without heating issues after behind being for so many years? As for ProRes, this is wishful thinking as someone mentioned earlier. It looks like some sort of limited 8K "for the sticker" on the box and then massive 4k crop or pixel binning/line skipping output. But I hope to be proven wrong!!!

    Finally, I don't think that Sony needs to wait for the official Canon announcement to figure out what they plan to do. Look how all the major brands (Pana, Canon, Nikon) came out with FF mirror-less camera last year at the exact same time. They all have similar supply chain and providers, they do business intelligence and know what the competition is going to do (more or less). The Z and R release were both underwhelming. They all came out with slow lenses and a few useless super wide and overpriced options for the show before releasing f/2.8 glasses later on.
     

     

     

  12. 1 minute ago, Trankilstef said:

    The 8k feature might just be a burst ofna few seconds where you can extract stills or shoot an 8k material then processed into 4k video file as it implies in this screenshot I made from one of the press release. Do we understand the same thing? 

    Screenshot_20200213_073311_com.brave.browser~2.jpg

    Yes, see what I said before. Basically it will be 8K for the headlines and to have the "8K ready" sticker on the box...

  13. A few remarks:

    • The body looks awfully small for an 8K FF monster. In the video, it is paired with a 24-105 f/4 lens and the body seems more like a Rebel/RP series than a 5D in relation to the lens size.
    • But the R5 terminology might indicate that this camera fights in the 5D category
    • There is absolutely zero details in this announcement, except for FPS and usual language (new sensor and processor).
    • I wouldn't be surprise if the 8K capture is only available in short burst. Fuji might enlarge its next X-T4 camera for video capture, Panasonic can't event do 4k60 without crop on its fan equipped brick size S1H camera and all the sudden Canon shows up miles ahead of the competition with a small body and 8K recording?
    • Not only this camera would mark the end of the cripple hammer but it means that Canon secretly worked for months/years to make a giant technical leap with camera design, and they just introduced this out of the blue after a decade of sub-par release.

    Anyway, if this thing is real and hit the market with clean crop-less 4k60p, I'll probably get one. I'm not going to complain!

     

  14. Option A: Canon has a secret sauce and manages to release 8k on FF mirrorless camera. Not only Canon is going to abandon years of crippling policy but they are going make a massive technological come back by beating the competition.

    Option B: they gonna fuck it up one way or the other with weird limitations, exorbitant pricing, marketing specs for the stickers on the box but without value for users (8K is actually time-lapse mode or upscaled 4K).

    Normally, I would be inclined to pick Option B but since their sub-par product line and pressure from the competition starts having an effect on their sales, we may not completely rule out Option A. Finger crossed.

     

  15. 1 hour ago, scotchtape said:

    Wtf, "everyone" shoots 60p and drops it into 23.976 timeline for slomo and it works fine... I do it all the time, and I'm pretty sure everyone else does too...

    +1 for 60P dropped in at 23.98 timeline.

    Other than that, it's true that you should ideally have 3 disks (1 for OS/app, 1 for editing files, 1 for app cache file). It's easy to do with the price of SSD these days. SATA SSD are more than enough for file storage when latency is not an issue.

    In any case I always work with proxies and my relative old computer still works fine. Generating proxy is such a game changers and it makes the entire process much more efficient even with the initial encoding phase. Unless you need a monster editing machine for time sensitive projects, using proxy is the solution.
     

  16. 48 minutes ago, Video Hummus said:

    Uh, most large, older companies are filled with this kind of crap...

    Exactly. Have you guys ever worked in a company? Jealousy, self-protection, courting the directors, powerless HR, burnt employee and tyrannic managers, division between services, regions and department, etc. That's business as usual for most company. And the more competition, less revenue there is, the worst it gets.

     

  17. 3 hours ago, no_connection said:

    Linus did a high clock speed test for their encoder server compared to core count recently and things changed sine the "don't use more than 8 cores" thing was relevant. I would not trust that based on years old info at this point.

    Again, I'm not saying "don't use more than 8 cores". I said, in this order of priority:

    #1. Performance depends on the App optimization so check the benchmarks
    #2. Based on benchmarks and your user profile (want to improve scrubbing performance, rendering h265 or ProRes? etc.), use your money wisely and build a balance computer as a whole.
    #3. High core count CPU might improve performance in certain tasks in terms of absolute performance but it might not make sense economically.

    All that is not contradicting Linus video above:

    • Using a 14 cores I9 9990xe costing $3000 gives modest performance gain over very fast frequency CPU.
    • Overage benchmark for Puget on Premiere Pro shows a performance difference of 20% between the $3000 I9 9990xe (14 cores) and the $500 I9 9990xe(8 cores) CPU. So 80% core increase  doesn't translate in 80% performance gain, however the price is multiplied by a factor 6.
    • Linus only compares rendering performance based on cineform source footage. Might be a different story with Pre-Render, h265, heavy effects. So we must check specific benchmark. There is right or wrong answer but if you spend 90% of your time dealing with h265, rendering performance benchmark on intermediate codec might not be relevant. Back to my point #2: check your user profile, make a diagnostic, see which tasks you want to address first and check the bench on this specific point.

    I'm not against high core counts but I would personally save the $2500 difference between the I9 9990xe and I9 9990xe in order to put that money into more DDR (after effects), better GPU (resolve) and a shit load of SSD (mix of SATA and NVMe).

     

  18. 1 hour ago, EphraimP said:

    Good article. I'd love it if you could expand on why 8 cores is the point of diminishing returns. Puget rates some monster 18+ core processors pretty highly and puts an 18 core unit at the top of their list of recs for Premiere. Thanks for link to Puget, btw. I was meaning to look them up but hadn't gotten around to it yet. I know they are highly regarded for their builds. As I've been told, the trick is to rip their specs and get the actual build done cheaper (if possible).

    Also, is it really beneficial to run two lower tier cards, such as the RTX 2060 or 2070 units over a high spec card like the 2080 super or an RTX quadro card? 

    I'm reading as much as I can on these topics, but it's good to ask questions and get responses from folks who've had experience.

    Again, it depends of the App , but generally video editing software struggle to efficiently use more than 4-8 CPU. Past this number of cores, the correlation decrease sharply between the performance and the price:
    Eg: an 8 cores CPU at $100 would give you 100 points of performance but a $800 16 cores would only reach 120 points. So you would have to spend 8 times the money to gain only 20% of performance. Sure, Puget would say "buy the monster 16 cores for absolute performance" but you may want get the 20% slower CPU, save $700 and put this money on NVMe SSD, additional DDR and better GPU if you use resolve. That what I said previously about balancing your rig. Don't waste your money on "status" items.


    As for the card, I think that you are talking about SLI and Crossfire? Be careful with that, make sure your app really use that. Last time I checked, most software are not optimized for parallel GPU use so it might be completely useless. Again, look at the benchmarks online.
    Also, you are entering in a different world with dual GPU in terms of cooling, space and power supply.. First make sure that your tower can fit two large cards and you must scale your PSU accordingly.

    PS: be carefull about video game benchmark (most of the bench online actually) as video games are very poorly optimized for parallelism. Said otherwise benchmarks based on video games tend show better scores on fast frequency / low core counts CPU. And GPU is what matters the most for video games.

  19. The PC industry is full of BS these days, Intel is struggling to innovate since 2015 and the graphic card market is dominated by Nvidia but Radeon is making a come back. I wrote about that in detail:

    https://fstoppers.com/originals/little-lies-and-big-problems-computer-industry-lack-innovation-artificial-351357

    For the rest, as someone mentioned, there is not such thing as "best" CPU or GPU, etc. It depends of your use and your budget. The key is to pick the right gears for your user profile and maximize your money.  Build a balanced setup and avoid bottlenecks like spending a lot of money on memory when your CPU is the limiting factor. As for Intel vs AMD, it depends of your editing app and your expectations (want to improve editing experience or focus on rendering?). The way to go is to check these benchmarks:

    https://www.pugetsystems.com/recommended/Recommended-Systems-for-DaVinci-Resolve-187/Hardware-Recommendations
    https://www.pugetsystems.com/labs/articles/What-is-the-Best-CPU-for-Video-Editing-2019-1633/#DaVinciResolveStudioCPUPerformance

    Long story short, for Resole, AMD tends to be better than Intel but this app relies mainly on GPU. For Premiere and adobe in general, it relies mostly on CPU frequency and does a poor job with parallelism. Better have a 4-8 cores CPU for Premiere, you would mainly waste your money beyond that.

    FINALLY, please make sure to purchase a good power supply and have enough ventilation in your computer. I cannot stress enough the importance of installing a good power supply unit (PSU) in your PC. As a rule of thumb, do not even consider a power supply below $40. Cheaply made units have poor efficiency, they will waste energy by producing a lot of heat requiring a noisy fan to evacuate this thermal load. Finally, low-end PSUs generate bad quality voltage and amperage which will stress the precious electronic parts of the PC, in turn reducing their lifetime. In the worst case scenario, the machine may become instable and crash. Don’t be cheap and save yourself a lot of trouble by investing a little bit more on a decent PSU. Rely on established brands and spend between $50–$120 depending of the power needed.

     

  20. 3 hours ago, Video Hummus said:

    Highly doubt Canon would sell you an R camera with better specs and less money than a C500.

    I have my doubts too, but perhaps, I mean PERHAPS, it's slowly start to penetrate Canon executives brains that all their stupid crippling is not translating into more C500 sales but increase in Pana/Sony/Fuji sales.

×
×
  • Create New...