Tag Archives: video editor

Sony RX100 Mark IV Picture Profile Part 2

I have done some further research on the picture profiles and found out quite a bit of information.

In this post I will focus on the luminance and black gamma.

Luminance

Luminance determines the range of black and white that are in the footage. People familiar with the Panasonic GH4 will know that you can set luminance fundamentally in two ranges: 0-255 and 16-235.

What this means that you have in the first case 256 grades of grey and the in the second 220. Some people confuse luminance with dynamic range but they are not the same thing.

You will somewhere find the definition of limited for 16-235 and full for 0-255 this is what it means.

In particular video broadcasting legal luminance is 16-235 so if you use a clip produced at 0-255 the extremes will be clipped this is anyway resolved as the levels are mediated when clips are produced for compatibility.

On the other hand if you use a clip with range 16-235 on a computer that has 0-255 luminance range it will look as lacking deep blacks or whites.

Ultimately you need to decide what is that you are shooting for and if your clips are going to be played on a computer that works in RGB or on a Tv that has YUV.

Profile Name Description Luminance Range Dynamic Range
PP1 Movie 0-255 < 7 stops
PP2 Still 0-255 < 7 stops
PP3 Video Natural 0-255 7 stops
PP4 Video Vivid 0-255 7 stops
PP5 Cine1 109% 0-255 10 stops
PP6 Cine2 100% 16-235 10 stops
PP7 S-log 0-255 13 stops

So the only profile that is broadcast compatible before editing is PP6 or the Cine2 gamma curve.

You can see that the dynamic range is the same for PP5 and PP6 so the fact that the camera records more grey levels does not really change things as the cine2 curve is smoother so it can accept higher input signals.

Black Gamma

In order to increase depth of blacks it is possible to use the black gamma setting. This has 2 controls, one is the Range Narrow, Medium, Wide and the other is the level that goes from -7 to +7

If you are going to grade your footage of and if you are going to use the video gammas PP1-PP4 you should not touch the black gammas otherwise you risk crushing the blacks.

However if you intend to use the cine profiles straight out of the camera you can tweak the black gammas accordingly.

The range determines where the setting will be effective.

Broadly speaking Narrow works on the first 10% of the signal, medium around 20% and wide around 30-35%.

What it means is that Narrow really works on the deeper blacks as you move to Wide you are altering also the grey and effectively changing the balance of the whole image.

More details here http://helpguide.sony.net/di/pp/v1/en/contents/TP0000909110.html

For what concerns the level a positive value will move black towards grey and a negative value shift grey towards black.

There are two main uses of the black gamma: get deeper blacks without altering the overall contrast this is obtained using the Narrow setting and a value between -3 and -7 or use the Wide setting with negative values to give the whole image a darker tone.

A setting of Wide with level around -3 gives an overall darker tone to a Cinegamma if you don’t want to change the blacks in post and remains overall balanced.

Which leads to my current favorite profile that is a customised PP6 setting the parameters that I have changed are:

I have tested the various Gamma with backlight situation and I found that Cinema2 performs best on my Tv where I watch my clips.

I prefer Cine2 even on the computer to be frank but it is true that the blacks are a bit light.

Color mode: I have tried Cinema and Pro am now on Pro with Saturation +8

Black Gamma: Range Wide Level -7 as the cine2 mode is quite dull I like to push the blacks a little overall.

As the Sony RX100 Mark IV records at 8 bit my opinion is that using S-gamut is not worth without an external recorder. And to be honest the amount of grading possible is quite limited so my approach is to get the video as good as possible out of the camera.

This is a little test with my Kitten

Advertisements

Picture Profiles for the new RX100 and A7 Cameras

The new cameras in the RX100 and A7 series can record 4K video internally but what is more important come with significant enhancements for video shooters.

Probably the most important feature is the availability of picture profiles that have a number of presets that can be further customized by the user.

There are many forums on the internet with custom setting to try to extract cine like look from Panasonic GH4 clips but all in all we can say that the controls in a still camera have been quite limited so far.

The new software that comes with the RX100 and A7 is a real step forward in this respect and contains all characteristics and functionality previously only present on Sony professional and semi professional video cameras.

 

The Seven Presets

The starting point of using a picture profile is to find one that we like as a starting point for further customization; the new Sony cameras have 7 profiles already loaded with example settings. The profiles have many characteristics that can be configured but probably the most important is the shape of the gamma curve because it also determines the minimum ISO and the availability of other features.

Profile Name Description Minimum ISO Knee Mode
PP1 Movie 125 Auto/Manual
PP2 Still 125 Manual Only
PP3 Video Natural 200 Auto/Manual
PP4 Video Vivid 200 Auto/Manual
PP5 Cine1 109% 200 Manual Only
PP6 Cine2 100% 200 Manual Only
PP7 S-log 1600 Manual Only

The descriptions above are my personal short description as the user guide is quite confusing.

Profiles PP1 to PP5 are usable straight from the camera, PP6 is optimized for editing and PP7 can’t be used without editing.

 

Standard Profiles

The profile PP1 and PP2 emulate the camera behavior in movie and still mode in terms of contrast and saturation. The main difference between the two is that the Still mode has stronger contrast and saturation than movie. Both use a minimum ISO of 125 and they allow customization of the standard camera modes.

PP1 Movie
PP1 Movie

Within a profile you have access to controls for color and sharpness at much granular detail that you have with the camera in normal mode. For example the saturation slider goes from -32 to +32 and the sharpness can be completely controlled manually. Colour depth is available in the range -7 to +7 for the whole RGCCYM palette. The customization of those controls lets you have the colours you want off the camera this can be particularly important to people looking for a specific look of certain colors typically red and blue.

PP2 Still
PP2 Still

For a complete description see this link:

http://helpguide.sony.net/di/pp/v1/en/contents/TP0000909111.html

 

Another feature that is incremental to the normal movie mode is the knee setting. The knee is particularly useful with highlights and with back lit shots. What it does it to compress the highlights within the usable signal range.

Knee Point and Slope
Knee Point and Slope

 

 

The knee is a way for those who like video look footage to capture more highlights and avoid hard clipping of back lit images.

 

Video Profiles

The profiles PP3 and PP4 use the ITU709 gamma curve that is the standard for high definition video television. The gamma curve has a low light gain of 4.5 dB and for this reason the minimum ISO is 200.

The video profiles bring substantial increase in  both shadows and highlights to the image compared to the standard profiles.

PP3 ITU709
PP3 Video Natural

The difference between those two profiles is the selection of the color mode. The PP3 profile has a Pro colour mode that the manual says ‘it is similar to professional Sony broadcast cameras’ the PP4 instead has an ITU709 Matrix colour mode.

PP4 Video Vivid
PP4 Video Vivid

What it means is that the Pro colour is slightly below the standard saturation whilst the ITU709 Matrix is extremely saturated looking like the Tv program that go these days. Both settings allow usage of the knee setting in auto and manual mode. The idea is to restore the highlights and make the footage look less harsh. The issue is of course that this has only effect on the highlights and not on the shadows.

 

Cine Profiles

The profiles PP5 and PP6 use the two different Cine gamma curves one is limited at 109% and the other at 100%.

Gamma curves
Gamma curves

Those are the official descriptions:

Cine1: Softens the contrast in darker image areas and emphasizes gradation changes in lighter image areas, producing a subdued tone overall (equivalent to HG4609G33). Gamma curve that obtains a dynamic range of 460% when the exposure is adjusted to 33% video output with 18% reflectance gray. The maximum value of video output is 109%.

Images shot with this gamma can be used without grading, but since the images have smooth gradation characteristics, this can be used to grade and finalize the viewed image in the post-production process.

PP5 Cine1
PP5 Cine1 109%

Cine 2: Similar results to [Cine1] but optimized for editing with up to 100% video signal (equivalent to HG4600G30). Gamma curve that obtains a dynamic range of 460% when the exposure is adjusted to 30% video output with 18% reflectance gray. The maximum value of video output is 100%.

PP2 Cine2 100%
PP2 Cine2 100%

Both curves are cine like and therefore do not use the knee function by default this is because a cine gamma curve is already clipping at much higher signal levels. The Cine2 curve has less contrast than Cine1 and is limited to 100% as opposed to the standard 109% so it is optimized for editing and the footage will lack whites and look fairly grey.

Practically the cine profile achieve a higher dynamic range using a smoother curve, this compares to the video profiles where the curve is pretty much the same until the highlights get compressed using the knee function.

S-log2

The PP7 profile is for S-log2, this has the highest dynamic range but the minimum ISO is 1600, the footage looks really bland out of the camera but this can be really enhanced through editing to bring out all the details.

PP7 s-log2
PP7 s-log2

 

My Take on the Picture Profiles
I think the whole cinema like video is a very overrated subject for underwater use. An acid test to understand where you stand is your own TV. If you use the picture settings Vivid or Standard then you are a video like person, waste no time pursuing cinema like look as in the editing you will effectively bring it back and eliminate all the detail you capture by increasing contrast and saturation.

If instead your TV is set to Cinema or is even calibrated then you are a Cine like person.

Personally I think S-log2 for underwater use at 100 mbps is a total waste of time and it is only worth with external recorders. The bitrate is too low to capture that amount of detail and resist extensive grading and actually shooting at 1600 is pretty difficult and requires ND filters (that the RX100 has) to perform decently.

You can take still shots with the picture profile so the first thing to do is to take a few pictures and see what image you like the most, then you can start tweaking.

I found myself to like the cine settings in terms of contrast but I find then too bland in color so I have settled for +20 saturation right now.

Between Cine1 and Cine2 is quite difficult to choose because although Cine2 is in theory optimized for editing actually it looks pretty good straight from the camera for my liking. I particularly like the fact that the highlights almost never clip with this setting even in harsh conditions. It is possible to mix the Cine profiles with the Pro color mode but I found that using saturation was giving the same result.

PP5 Saturation +20
PP5 Saturation +20
PP6 Saturation +20
PP6 Saturation +20

 

Detail setting (aka Sharpness)

The last feature of the RX100 and A7 new controls is the fine tuning of the sharpness setting. I find that the default works pretty well except at high ISO where you want to start adjusting slightly the behavior.

The full guide is here http://helpguide.sony.net/di/pp/v1/en/contents/TP0000909112.html

But it takes a bit of time to understand the settings.

The starting point for me is the crispening setting that in perfect Sony tradition means exactly the opposite of what you would think. So a setting of +7 means actually limit the sharpening on speckles and noise where a clear shape can’t be detected. Once you set crispening to the max and you shoot at high ISO you see how it reduces the noise. However the issue is that now the sharpness has also reduced so in order to restore the effect you go back to the total level of detail and set it to 7. Once you do that you find out that now the sharpening applied correctly is actually excessive so the limit setting effectively clips the amount of sharpening applied to the edges I use this at 3, as this results in a reduction of highlights sharpening (glass reflecting objects look dull) I then set the highlight detail to +2. After I did all of that the end result was very similar to the standard setting and a bit softer on the edges, practically it was only better at very high ISO so I went back to the original setting.

 

 

100,000 visits – In Depth into Sharing Videos on the Internet

Two years and few months later I am pleased my blog hit 100,000 visits. Considering that there is sponsorship and this is pretty much content produced during free time I am well pleased.

So as commemoration topic I want to put a few considerations that spin off a post on the editing and share section of wet pixel.

Many people spend a lot of money on underwater video rigs and use sharing websites such as youtube and vimeo to host or promote their content. The reason is clear those sites have a huge audience and if you have original content you can get a bit of advertising revenue as well that is never a bad thing.

However most of us have noticed that once you upload a file on those websites it looks worse than the original some time much worst. Why does that happen?

The answer lies in two words video compression.

Video compression is a technical subject and my previous post tried to share some of my finding in regards of the reasons why a camera produces video better than another even if the second produces better still images. It is all in the compression effectiveness and the same issue applies when we share our videos on line.

Unfortunately many people do not really know much about this subject and assume that the video editing program they purchased has all the answers and everything is optimised. Well that is not the case. Video produced off the shelf by such programs with default settings may be watchable but are not great and usually worse than the source clip of a good deal.

Another common misconception is that you need to convert a file produced by your device to another format so you can edit.

Finally many people convert files many times and wonder why the result is far off the original clips, not realising that video compression is lossy so each time you manipulate a clip are you are making things worse.

Obviously am talking consumer and prosumer here not RAW video recording at stellar bitrates.

So what is the best way too produce an underwater clip that looks good without spending too much time on it and that when uploaded on the web looks still decent?

To give an idea why a clip like this one shot with a compact camera

Does not look to far off this other clip shot with a semipro camcorder Sony AX100

or a Panasonic GH4

What all 3 clips at 1080p on youtube and honestly evaluate if there price difference is justified you will probably think no and think the second clip is actually a pro.

So why is that?

50% of the problem comes from the editing, I don’t have the details of how the other two clips are done but I know my clip is edited with iMovie, surely not the most advanced tool on the market you would think.

However there are a few tricks of the trade that I will explain to you one at time:

1. Never let your editor convert the files at the import.

Unless your workstation can’t physically process them leave the clips as in. Even think about getting a better computer in the long run if you can’t process files as is.

Many editors convert the files at import, in intermediate formats like prores or Avid that have no temporal compression. Those files unlike the originals have each frame stored like a complete image so that it is easier to edit. If your editor allows you use the original file without any conversion. You can do this in Final Cut using proxy and cheating also in iMovie creating manually event folders and copying mov or mp4 compliant files manually into them.

2. Once you finish your editing use the highest quality option available for export.

This is sometimes a tricky issue as the default options of those programs mention sometimes just a quality option with a slider from low to best. Many programs though, like final cut offer other options and modules for advanced compression.

If you have spent money on the editor spend the extra funds on the advanced codecs as they are worth every penny.

Once you have the advanced codecs (x264 is the one I use and is free plug in for iMovie) use constant quality with factor of 18 and the slowest preset your workstation can bear.

X264 preset go from very fast to placebo, my workstation can tolerate a very slow for 1080p that applies all the most advanced compression settings. This together with quality at 18 gives me an output very similar to the input but much more efficient with a smaller file.

At this point you are nearly there and ready to upload on vimeo and youtube.

Between the two services which one has the best quality?

Vimeo plain and simple, the same file will look better than youtube with less artefacts at the same resolution, however vimeo requires you to have a plus account to upload and share in 1080p whilst youtube is free.

So this is the reason why your files do not look as good as the clips you shot with the camera when you share them.

Now onto the second part why do clips produced with my very expensive equipment look worse than someone with a much cheaper set up and inferior equipment?

This second problem has to do with the way videos are shot.

Many people look on the internet for guidance on how to produce a video clip that looks decent and are tempted by some esoteric terms such as: flat profiles, colour grading, gamma curves etc etc.

They then go into water with their camera set like they have read on the internet and then spend a long time editing their clips, after all that effort the result image is a bit soft and the colors are washed out.  This seems to be quite a common issue especially with pros.

http://www.peterwalker.com/komodo.html

Note that the two videos above are probably two of my favourites of the last few years. However check the difference between the close up shots with lights or the land shots and the wide angle with natural light? Very different

This instead is an example of someone who knows how to work with the limitation of the set up:

Flat profiles and color grading may work very well when the environment is controlled in a studio situation or where there is plenty of light but in water this is seldom the case. So the best help is to get it right first time and if needed use a filter for your ambient light shots.

Many people including me used to be a white balance evangelist but I have to say with years I have lost interest and I think is greatly overrated.

This video from ikelite is my absolute favourite

The best part is at 0:45 comparing filter with auto white balance and filter with manual white balance. The clips says looks at the purple that comes with the manual white balance but actually that is a horrible hue there!

I have spent the entire 2012-2014 trips trying to perform custom white balance with various cameras, with various degree of success. When I was in Raja Ampat I once left the camera in auto and realised the color where the best I ever got. Though this was a mistake but after few months when I reviewed the clips and how they were taken I realised the truth, even since I have never hit the custom white balance button once on my RX100 and I am preparing to do exactly the same on the GX7.

So my five cents into video editing and doing something decent for sharing on the internet is based around the following key principles:

  1. Get the clip right in camera. Use the settings that make the clip look great at the outset, experiment until you are happy of the results. Forget about theory focus on what you like.
  2. Don’t let your editor alter the clips at all and use no or minimum grading or even try to do no correction at all including contrast and exposure any time the editor touches the clip something is damaged.
  3. Export with advanced settings using all the CPU power you have at hand to produce a high quality but as small as possible file

Good luck for your next trip, I am very much looking forward to mine!

 

Underwater Video Tips: Working with AVCHD 2.0 and 1080p60 or 1080p50 files in iMovie

As hardware becomes more and more powerful video format evolve to allow higher quality capture.

AVCHD is a format that still relied on interlaced video and the classic 24p until version 2.0 where higher frame rate 1080p50 and 1080p60 have become standard with a maximum bit-rate of 28 Mbps.

To date many non linear editing programs are not capable to process such files actually most of the low cost programs are not even able to import those files at all, this is quite frustrating after spending a good amount of money on a camera.

I use iMovie for all my edits as after testing programs like Adobe Premiere I did not really find them to add many benefits to justify the price and I also find them quite slow and counter intuitive so when I got my Sony RX100 I had the issue of processing AVCHD 2.0 files 1080p50.

An AVCHD container is made of streams that have a video and an audio track plus another track of text. The video is encoded in H.264 as other formats like mp4 and the audio is AC3 usually two channels. Usually video editor like files with an H.264 video track and a stereo audio track in AAC or MP3.

So if you re-wrap the information in an mp4 or mov format there is a good chance that a program like iMovie or final cut will digest it.

After various attempts I managed to find on the internet the tools I needed, I will list them here:

  1. LAME for Mp3 encoding (mandatory)
  2. FAAC for AAC encoding (optional but I have it in my build)
  3. FFMPEG
  4. Growl
  5. Clearpipe automator Action
  6. Automator FFmpeg action
  7. MTS2MP4 automator agent

For instruction on how to build your own ffmpeg (as the static builds did not work for me) look here:

http://sesam.hu/2012/09/05/installing-ffmpeg-on-os-x-mountain-lion/

Then install growl version 1.2.2 http://growl.googlecode.com/files/Growl-1.2.2.dmg

Get clearpipe, automator ffmpeg action and the mts2mp4 finder service here http://blog.laaz.org/apps/automator/ and install in sequence.

This creates the option to right click on an MTS file and re-wrap it into an Mp4, note that there are also commercial programs that do this like clipwrap and iVi however our finder service is free and quick…

I have created this little video to show how it works in practice, as you can see it swallows entire folders which is great. So here I create an output folder in the iMovie events folder so that iMovie can edit the 1080p50 file later skipping the import, this means no time is wasted and after generating thumbnails you are ready to edit your original video at high frame rate, a feature ‘officially’ not supported…this is how I edit my video natively in iMovie. If you have a GoPro that saves 1080p50 or 1080p60 mp4 files you can start from the manual creation of an event folder.

From there onwards you can import your double frame rate video into iMovie projects, that will anyway be 24,25,30 frames per second by default but can also exported in 50/60p using x264 decoder that you can find here http://www003.upp.so-net.ne.jp/mycometg3/

This means that you can process with iMovie and also final cut pro 50/60p projects with no problems!

Update for those struggling this is the link where all the files including the ffmpeg build are: https://www.dropbox.com/sh/6m4527odhpw3hcc/nHODxg3_DL I have modified the ffmpeg automator action as I was getting a problem with growl