martian1

Joined on Jul 11, 2007

Comments

Total: 19, showing: 1 – 19
In reply to:

Aleks7: I think I don't understand the speeds in detail, it really can do ~3GB/s sustained read speed thorough the hdd or is it more like burst using the buffer?
Looking at the numbers one could say the WD Blue 3D NAND SSD M.2 is 5-7times slower with it's read/write speeds about 500MB/s, but is it in real life?
I'm after 2T M2 SSD so I'd appreciate if someone could shed a bit more light here

This is not a burst using the buffer, it is actual sustained read and write speeds - sequential with a Queue depth (QD) of 32. For QD= 1 / QD= 4 CrystalDiskMark 6.0.0 measures my 960Pro 1TB still at 1806 / 2703 MB/s read and 1592 / 2028 MB/s write speed.
Please note that depending on the cooling in your computer after continuous use of roughly 2-10min, the SSD may apply a thermal throttle (Samsung calls it 'Dynamic Thermal Guard') to protect the SSD, this can be delayed/reduced by an additional cooler. However, already 1min at e.g. 2000 MB/s mean a total of 120 GB, which far exceeds normal use.

Link | Posted on Apr 24, 2018 at 22:26 UTC
In reply to:

Tom Holly: All this for one stop? Life’s too short.

I shoot velvia occasionally. That has 5 stops of dynamic range. Hacking files to increase from 14 to 15 stops (or whatever it is). Couldn’t be bothered. I’d just buy a Sony if DR was that important.

Such prep work is only needed when you want to ensure that you get the max. dynamic range, while in many other situations you will get a (partial) benefit even without prep.

Please note that bracketing will capture images at slightly different times, therefore producing artifacts as soon as any discernable subject or camera motion occurs between the images. In contrast, dual pixel images are taken at exactly the same time (as already mentioned in the article).

Link | Posted on Feb 12, 2018 at 22:47 UTC
In reply to:

cpugourou: If some find there dslr already big, heavy and hot, wait for encryption !
To encode live 40MP raws at 10 fp/s you need a multicore that will dry out your batferie in no time, a cooling system with a fan, a buffered ddr5 RAM deck, etc etc
Forget it

Encryption after shooting will still use a considerable amount of battery, time and processing power - and when the request for encryption is at some time after shooting, why not simply encrypt on a laptop/PC, all of which already today provide a range of encryption options.
Additionally, please note that a camera-based encryption is likely to be hacked rather sooner than later and cannot be readily updated as it is part of the camera's firmware. A laptop/PC based encryption may also be hacked, but the encryption method can well be changed when needed.

Link | Posted on Feb 7, 2018 at 04:38 UTC
On article Panasonic Lumix DC-GH5S Review (589 comments in total)
In reply to:

martian1: Reduced sensor resolution of 10.2MP (from 20.2MP for GH5):

As DPR already explain a reduced sensor resolution actually does not yield any benefit in terms of dynamic range or noise (when scaled to a common size), while it sacrifices resolution at least in good light with the 'only' advantage of faster read-out speed (half the data to read equates to double read-out speed) and thereby reduced rolling shutter in video.

Sony's RX10 Mark IV (and RX100 V) substantially reduces rolling shutter using fast read-out enabled by a sensor with DRAM chip, AND down-samples video from 20.1MP (with clear advantages in detail capture and reduced Moiré over a native 4K sensor) - apparently this is simply a better solution compared to the GH5s' sensor.

I do not know whether today a 20MP 4/3 sensor with DRAM chip is available, but it would clearly enable better video quality in terms of detail capture and Moiré without sacrificing other image quality including rolling shutter, dynamic range and noise.

@liberty:

Actually I think you are mostly right concerning noise.

However, the noise advantage of a lower resolution sensor compared to a higher resolution sensor of the same size is small, as the sensor area capturing light is the same (except the tiny distances between adjacent pixels which are more for more pixels), and the noise advantage is primarily valid when a sensor is continuously in use, e.g. during video capture, thereby generating more heat in a period of time compared to single frame capture which generates less heat in the same time period. Therefore an additional source of noise is the resulting increased sensor temperature. Obviously, the sensor temperature also depends on sensor cooling and therefore overall camera design, so any actual benefit varies by camera.

To me it is not clear, whether capture stops to prevent overheating of the sensor before any significant increase in noise, or sensor overheating only occurs when noise has already substantially increased.

Link | Posted on Jan 9, 2018 at 23:28 UTC
On article Panasonic Lumix DC-GH5S Review (589 comments in total)
In reply to:

martian1: Reduced sensor resolution of 10.2MP (from 20.2MP for GH5):

As DPR already explain a reduced sensor resolution actually does not yield any benefit in terms of dynamic range or noise (when scaled to a common size), while it sacrifices resolution at least in good light with the 'only' advantage of faster read-out speed (half the data to read equates to double read-out speed) and thereby reduced rolling shutter in video.

Sony's RX10 Mark IV (and RX100 V) substantially reduces rolling shutter using fast read-out enabled by a sensor with DRAM chip, AND down-samples video from 20.1MP (with clear advantages in detail capture and reduced Moiré over a native 4K sensor) - apparently this is simply a better solution compared to the GH5s' sensor.

I do not know whether today a 20MP 4/3 sensor with DRAM chip is available, but it would clearly enable better video quality in terms of detail capture and Moiré without sacrificing other image quality including rolling shutter, dynamic range and noise.

@Life recorder:

Thanks for the info on the sensor.

However, oversampling a higher resolution image and downsampling to 4K will produce better detail and reduces Moiré, at least for sensors using a Bayer filter pattern to estimate colours at each pixel using information of the surrounding pixels, i.e. each individual pixel 'only' provides green, red or blue information - and the GH5s as almost all cameras (except Foveon) uses a Bayer filter pattern.

Link | Posted on Jan 9, 2018 at 22:27 UTC
On article Panasonic Lumix DC-GH5S Review (589 comments in total)

Reduced sensor resolution of 10.2MP (from 20.2MP for GH5):

As DPR already explain a reduced sensor resolution actually does not yield any benefit in terms of dynamic range or noise (when scaled to a common size), while it sacrifices resolution at least in good light with the 'only' advantage of faster read-out speed (half the data to read equates to double read-out speed) and thereby reduced rolling shutter in video.

Sony's RX10 Mark IV (and RX100 V) substantially reduces rolling shutter using fast read-out enabled by a sensor with DRAM chip, AND down-samples video from 20.1MP (with clear advantages in detail capture and reduced Moiré over a native 4K sensor) - apparently this is simply a better solution compared to the GH5s' sensor.

I do not know whether today a 20MP 4/3 sensor with DRAM chip is available, but it would clearly enable better video quality in terms of detail capture and Moiré without sacrificing other image quality including rolling shutter, dynamic range and noise.

Link | Posted on Jan 8, 2018 at 22:33 UTC as 125th comment | 5 replies
In reply to:

Pelvet: His punishment should be either to not be allowed to enter an airplane again or to have to be at the airport 8 hours before his flight for the rest of his life (or else he is not accepted in the plane). Then he will appreciate the lost hours that he caused for the passengers... And surely never fly a drone again in his life.

Actually, the drone pilot has to pay all damages caused by his drone (in addition to any fine for violating the law and endangering air traffic), in this case this may well completely ruin the drone pilot unless he is really rich - my very rough estimate is at least GBP 250'000,-
So please, drone pilots, pay attention and really avoid such dangerous and extremely costly behaviour!

Link | Posted on Nov 3, 2017 at 21:24 UTC
In reply to:

thinkinginimages: "Creative Cloud" was always a nightmare to explain to anyone. But, I want to keep to Lightroom.

Like Photoshop it's became a mess of added features, but never really addresses the underlying issues: it's getting slower as computers are getting faster, and camera files are getting bigger. The streamlined workflow is getting bogged down. I can't imagine how long importing would take with medium format files.

Naming the desktop version "Classic", and everyone hearing that the web version "is the future" just reinforces that notion that the desktop version is on the way out. Add the very slow version updates, even patches, adds a bit of fuel to that fire.

Even I'm thinking "It's time to find something else - how do I get out of LR." That's been going on for a while. The updates aren't solving the core problems, primarily speed, and feature bloat.

Now there's Lightroom Classic CC, Lightroom CC, Lightroom (web) and Lightroom mobile. Four versions? What? No LR Elements?

@ Reilly Diefenbach:
Taking into account raw image sizes of 40MB and even 10-12MB (obviously larger will be even worse) just shows how unreasonable a 'cloud'-based workflow is, as such files will take quite considerable time to upload and download, taking huge amounts of bandwidth when you shoot many of them... - to me a cloud-based workflow is plain nonsense.
Additionally, there might be a network or server interruption at any time, so you just do not have access to your images any more until a connection is re-established - who knows when? This can cause a major mess, when you have a customer waiting!

Link | Posted on Oct 23, 2017 at 20:24 UTC
In reply to:

CaliforniaDave: Thanks for the excellent article with good explanations of technical issues. I learned where the term aliasing comes from. Keep these articles coming.

Remember when the highest quality video cameras had 3 CCDs and a color separating prism? That is certainly a possibility for still cameras, but of course high end and high resolution video cameras today generally just use one sensor with a Bayer pattern. (I fully expect someone to point out some cases in which that isn't true, though.)

@FLruckas
Actually loosing light in the prism is far less than with a Bayer type sensor.
- In a dichroic prism assembly used for 3CCD/3CMOS the light loss is mainly reflection at the interfaces air-glass and prism-prism (note that some reflections are total internal reflection, so no loss occurs), all of the reflection losses are largely reduced by using coatings (same as for a lens) and matching the refraction indices with an optical epoxy.
Reflection loss of glass is about 4% per surface, however it is greatly reduced by the measures mentioned, and also the colour filters of a Bayer sensor have reflection losses, so really the light loss in a dichroic prism assembly is far less than for Bayer.

For further reading (and if somebody doesn't believe me) please see the Wikipedia article 'Three-CCD camera':
en.wikipedia.org/wiki/Three-CCD_camera

Finally, not only Sony uses 3CCD, also Panasonic, Canon and Toshiba use 3CMOS (Panasonic call it 3MOS).

Link | Posted on Apr 10, 2017 at 20:17 UTC
In reply to:

CaliforniaDave: Thanks for the excellent article with good explanations of technical issues. I learned where the term aliasing comes from. Keep these articles coming.

Remember when the highest quality video cameras had 3 CCDs and a color separating prism? That is certainly a possibility for still cameras, but of course high end and high resolution video cameras today generally just use one sensor with a Bayer pattern. (I fully expect someone to point out some cases in which that isn't true, though.)

Thanks CaliforniaDave for the precise description!
Some video camera manufacturers improved further: a 3CCD system can provide roughly 2x higher spatial resolution when the green sensor is shifted in horizontal and vertical direction relative to the red and blue sensors by 1/2 pixel, although then colour resolution is reduced accordingly, however the human eye is significantly less sensitive to colour resolution, so overall the system achieves substantially improved spatial resolution, maintains good colour resolution (better than Bayer) and captures all light from the lens (except for the small light loss caused by the prism).
Note that instead of a 3CCD system as well a 3CMOS system could be used, cashing in the advances of CMOS over CCD.

Link | Posted on Mar 29, 2017 at 22:37 UTC
In reply to:

belle100: With these cards and their corresponding development, can we forget XQD or CFast 2.0 and the like please.

Today a comparable Sandisk SDXC Extreme Pro UHS-II 128GB with 300MB/s read / 260MB/s write / U3 is available in Switzerland for CHF 225,- or roughly USD 225,- and I expect Sony's pricing to be similar, otherwise they are not competitive.

Link | Posted on Feb 23, 2017 at 22:48 UTC
On article SanDisk 1TB SDXC card prototype unveiled at Photokina (109 comments in total)
In reply to:

mpgxsvcd: The cost is the only concern here.

Actually SD cards are primarily used in compact and enthusiast cameras, meanwhile also for DSLRs, i.e. MF is in comparison a very small market segment. Cost is a primary concern, as such cards need to be competitive, in particular to smaller, however still high-capacity SD cards like 256GB or 512GB. Additionally, a 1 TB card is so new that concerns about the yet unknown risk of data loss will not help sales, if the price is not competitive.

Link | Posted on Sep 20, 2016 at 20:45 UTC
In reply to:

Gpruitt54: Totally stupid to assign ownership of the image to the ape. Of course, the photographer owns the image. Who’s gear was used to capture the image? Who set up the camera? Who made it possible to capture this image? Answer; the photographer.

Gpruitt54 - Nobody assigns the ownership of the image to the ape - except that you are making up this stupid idea in order to jump to a statement that the photographer whose gear was used owns the image.
Your simplistic statements do not reflect reality, while the legal situation appears clear - the images are in the public domain, because the ape who took the photos is not human.

Link | Posted on Aug 7, 2014 at 19:02 UTC
On article SD card labeling for 4K video announced (50 comments in total)
In reply to:

erotavlas: 4k? why don't they skip 4k and just go directly to 6k

6k is unlikely to be introduced at all, as the increase in resolution is only 22% (measured linearly - sqrt(6) / sqrt(4) = 1.22), which is just above the threshold for visibility of 20%, while the increase in required bandwidth, storage space and computing power is roughly 50% (6/4 = 1.5). The change from FullHD (2k) to 4k gives a resolution increase of 41% (measured linearly - sqrt(4) / sqrt(2) = 1.41) and therefore is highly visible.

Link | Posted on Nov 9, 2013 at 13:01 UTC

Colour resolution at full 1920 x 1080 can be provided by a 3 CMOS system using a prism assembly to split the 3 colours red, green and blue onto 3 CMOS sensors. This is today employed by many professional video cameras, albeit using relatively small 1/3", 1/2" or 2/3" sensors.

Note that splitting the incoming light using a prism assembly utilises ALL incoming light (except for minor reflection losses), while any colour filter in front of a pixel loses a major portion the incoming light by filtering out other colours, e.g. for red one has to filter out green and blue, for green one has to filter out red and blue, etc - in principle this loses 2/3 of the incoming light, however note that in a standard Bayer pattern the filters are setup a bit different.

Additionally, at each pixel full colour information is available through the 3 sensors, while for a single sensor using Bayer pattern filters, colours are always interpolated, as each pixel has either a filter for red, green or blue.

Link | Posted on Mar 5, 2013 at 00:12 UTC as 41st comment | 1 reply
On article Canon announces EOS C100 professional video camera (93 comments in total)
In reply to:

Donnie G: Just left the Canon Learning Center website, and I must say that their new C 100 is an exciting bit of cinematography kit that is destined to wind up in many a run and gun event/wedding/spot news multi-tasking image maker's camera bag. This new compact and light weight one chip design is sure to replace all of Canon's lineup of 3 chip camcorders. I can hardly wait to hear what real end users have to say about this video camera vs. the Red Scarlett, Blackmagic cinema camera, and the rest of its competition. I'm excited, and I don't even do video, YET. :)

Dear Francis,
Apparently you do not understand the point about 3-chip cameras - as stated before by Shogi and Donnie, in Canon Professional's current portfolio 4 out of total 7 cameras do have 3 sensors (either CMOS or CCD) and indeed with a prism. This is similar for Sony Professional, e.g. current XDCAM series - of total 11 cameras all have 3 sensors (PDW700, PDWF800, PMW200, PMW320, PMW350, PMW500, PMWEX1R, PMWEX3, PMWTD300) except 2 cameras with 1 sensor (PMW100, PMWF3). And similar for Panasonic Professional (here actually several consumer cameras have 3 sensors), etc ...
Hope this clarifies things.

Link | Posted on Sep 11, 2012 at 19:42 UTC
In reply to:

mick232: You people seem to ignore a couple of things:

1. there is more than one workaround available (don't open TIF files from untrusted sources, scan TIF file with virus scanner)

2. fixing a bug in old software is more expensive by orders of magnitude than in upcoming or current software; even if the fix is a one-liner, the software has to be re-built, re-tested, re-released. Don't underestimate the effort needed for all these steps. These processes have to be re-run for the fix whereas with software currently being developed they run anyway.

That is why any software company will assess the severity of a bug. No company will fix any bug, even if it is a security issue. That's just how it is and all your rants are not going to change it.

Sorry mick232,

I have read your last paragraph initially correct and just to confirm, read it again - you now have changed your original wording:
'No company will fix any bug, even if it is a security issue. That's just how it is and all your rants are not going to change it.'

Additionally, please note that CS6 is not yet available in some countries through most retailers, e.g., in my country Switzerland, therefore CS5.5 still is effectively the current version and not 'old software'.

Link | Posted on May 14, 2012 at 22:42 UTC
On article Variation Facts and Fallacies (231 comments in total)
In reply to:

PicOne: "If you ever think 'for $1,500 I demand perfection,' this is not the article for you, it will just get you upset. The laws of physics are not suspended, nor are techniques of manufacturing altered, just because you demand it be so."

Perhaps not... but for the more money spent, the more I would at least presume the QC procedures prevent crappola from actually hitting the retailer shelves. Your term "allowable tolerance" I would also presume is a floating definition. Again I would presume that more expensive lens productions lines have more stringent definitions of 'allowable' applied.

Indeed you seem to argue the opposite, in that the potential tolerances are much greater with lenses with more elements/groupings (ie. more expensive lenses) than lenses of simpler construction, and that therefore buyers should expect more expensive lenses to perform worse than cheaper lenses.

Lenses with more elements/groupings do not necessarily perform worse than simpler lenses, even for the same allowable tolerance, as not every element/grouping exhibits the same sensitivity, i.e. a certain variation may cause different levels of image quality degradation depending on where in the construction it occurs.

In a good lens design the allowable tolerances are specifically defined for each construction element depending on the impact of variation on image quality.

Consequently, a well-designed lens on average performs better than a lesser lens, independent of the amount of lens elements/groupings.

Link | Posted on Nov 26, 2011 at 01:22 UTC
On article Adobe faces criticism for change of upgrade policy (375 comments in total)
In reply to:

IEBA1: Two things of note here is that Adobe both accelerated the release schedule (more versions more often) and decreased the number of previous versions eligible for upgrade pricing. So, in essence, instead of upgrading 2, 3 maybe even more years apart, Adobe now wants to see new cash every year or so.

To their credit- they are making dramatic improvements in their software. So the ROI is still there IMHO.

Whether any improvements in their software are 'dramatic' highly depends on each customer's need, while it is a fact that many users do not update to every new version.

I currently use CS4 and planned to upgrade to CS6, however Adobe puts me under pressure to pay for CS5 by end Dec 2011 on very short and even strangely indirect notice via a blog post - just to be able to use CS5 for the short period until CS6 is out, when I could have enjoyed CS5 for a much longer time, if I knew this earlier.

Therefore to me Adobe's way of communicating is unprofessional and to do it on such short notice is unfair business.

Link | Posted on Nov 22, 2011 at 23:45 UTC
Total: 19, showing: 1 – 19