Just wondering what people think of the overall lighting from these clips? I exposed the children for the harsh spotlights, but due to very large gaps between the lights they got darker when not under direct light. I tried to compromise as best as I could. I used the XF300 for the closer shots, and an unmanned, stationary XF100 for the wide view.

RDA Summer Show 2017 Highlights on Vimeo

Any feedback appreciated.

Thanks

IMHO it looks a bit underexposed. I'd maybe open it up a stop or two, or bump up the gain. Understandably there seems to be a big spotlight location on stage, but alot of the shots the dancers don't seem to encounter it.

I agree with David !
When the camera "sees" that huge black background it should actually over-expose but it's done the opposite so one has to ask did you expose manually?? Yeah I would have gone one stop up as well if it was available ... I will often see what an auto exposure looks like before they start and then I have a ball park figure to work with ! Then again the parents who will watch this are not really going to be too fussy as long as their "precious" is in focus.

It's an interesting concept using closeups too ...I find that most dance teachers prefer the wide shot so they can assess the whole performance. Most we have done don't even want to look at a version with closeup cutins ... Has anyone found this??

Yeah, I agree about the lighting... It was my first time filming a performance and it was set to manual exposure in order to keep the look of the video the same instead of constantly changing exposures.

It wasn't for assessment purposes, so I did tend to go quite close at times, and I cut to the wide view during zooming, although I think next time I'll go wider, and open up the iris a little more.

Thanks for feedback.

I saw no over exposure of the people in the highlights so your camera could have had the lens opened up a fair bit - two or maybe three stops, but you need to test.

To be fair the biggest problem was the stage was lit with too few lights to get an even wash. You have something very typical for multi use stages of this size. Random focussing. probably aimed at something important but long gone. Most entertainment doesn't want even wash light at all because it's boring. Dance where people physically use all the space need even illumination as the minimum, and most theatre with conventional lighting can't offer this. New installs with LED wash lighting can do wide, even and bright - albeit usually in loads of colours, the designers hating white light as it's not theatrical!

Success in a video needs advance discussion and negotiation with the venue.

Actually I'm going to give a contrary opinion here. Yes there were some underexposed shots when they got out of the light but on some shots, unless you were shooting in log, you were pretty much on the edge for highlights and couldn't have gone much hotter. There are a couple of under exposed close ups - so if you are going manual you have to ride the iris and focus all the time. That's easy (well easier) on an ENG camera but much harder on handycams or still frame glass.

As others have said it is more a problem with lighting. If you can negotiate with the venue before hand start of with a 25% white wash across the whole stage and get them to add colours after that it makes the job easier and IMO is a better experience for the audience who can actually see their children instead of only the ones at the front of stage.

The whole argument about "theatrical" is bogus - costume colours actually pop in white light, not in coloured light.

Feedback on wide shots. I am a big proponent of the "rule of thirds". Many dance videos that I have seen usually have the wide shot with the downstage edge on the bottom of the screen. This then usually puts all the heads in the center of the screen with huge amounts of wasted space above. Dance competition do it because many times their logo is on backdrop behind and up high so audiences always see their logos. then it's in the videos as well. Personally, I prefer to put the heads in the normal upper third position this puts all the body movements more centered and filling the screen with dance. Now if you have to do a lockdown shot for the wide shot, then you have to come up with a happy medium to compensate for a full stage or jumps and lifts. So it depends on if you have an operator or not. I would tend to have an operator and push in when groups cluster to the center and pull out as they then fan out wide. Following the action much like a wide shot of a soccer game. Depending on the venue, you will sometimes see the backs of heads of the audience in the first couple of rows. Again, I think the improved framing looks better, but of course it's a matter of choice. Just a thought. I like your close-ups. Watch out going to them too much in a finished edit, and be prepared for some parents to complain that they didn't see their child enough.
I use to work with one studio that didn't want to deal with it so I just stayed wide most of the time unless there were solos and the dancers grouped together. That's where I started using more zooming in and out to follow the action technique. Trying to have a decent medium. Good luck.

The whole argument about "theatrical" is bogus - costume colours actually pop in white light, not in coloured light.
Sorry - you misunderstood me. The question is to do with the look in general. It's really a common issue.

Do you light for the audience there, or the audience at home? The only person who can set the rule is the commissioner - the client or producer. The theatre lighting designers light for the people in the audience. If the critical factor is the video, then you will need to provide them during the programming stage (when the video people are rarely there) video and a monitor, so they can light for your camera. LED lighting is now so common that for the first time, you can wash a large stage with solid colour - something previously impossible, so the lighting can be visually, wonderful. However, most cameras hate blue with a passion, and go sparkly with red. costumes can behave very oddly colour wise, to the benefit or detriment of the camera (and audience member) so if this is important it needs planning.

I run a 1400 seat venue, and we can light for any purpose. However, when the video person turns up half an hour before the show and demands 30% white, then unless the client has got me to programme this in the rehearsals we've spent all day at, they can go and jump because at this stage, it cannot happen without wrecking our input AND it's our name in the programme that says lighting design FRED SMITH or whatever. If the audience see terrible flat lighting, then I don't want my name on it. If the programming can cater for video at the clients request, then I'll do it properly - NOT - just add in a pile of white light.

The notion of a white state and then adding in colour is not lighting design, it's illumination. Quite honestly, I'd not want my name on something so unprofessional. Broadcast TV is exactly the same - you put in face light where it's needed, and then add to create mood. With piles of equipment, you can ensure everyone's faces are lit properly, but to do this in a theatre is tricky with moving people. People glued to mic stands are easy.

For dance, white light is a form wrecker. Professional dance shows use lots of side light from quite low angles, it reveals form and shape, especially in ballet and contemporary where costumes can often be figure hugging.

Light styles, as in classical ballet will be lit in pastel colours traditionally - so pale pinks, lavenders and golds are popular, often split left and right. Open white light is rare there's simply no need for it, and the American system was often to split colours left and right into warm and cool - pinks and blues. For the past 59 years, Stanley McCandless's ideas have been used, in the US, and some of his ideas leaked into UK practice too - so solid a foundation they are.

The notion of spoiling the performer to audience link with a bucket of white light just isn't on. It looks terrible and while a good compromise for the video people, it is such a blunt weapon. I do do it - but only when the client insists the video is paramount in the heirachy, and on these occasions I demand a monitor from the video person - who will be expected to be there at rehearsals and plotting sessions. If they cannot do this, then we tell the client that it's their call, and we will simply do what the video people want, and the terrible look for the audience is fine with us - we also insist we do NOT have a credit for lighting in the programme. In this industry, you're only as good as your last job, and with phones so prevalent, Facebook reveals horrible things.

Video might be the important thing, we're happy with that, but the attitude that you just shove a few faders and that's it, is the same type of nonsense as when people insist autofocus and exposure will be fine, and two different types of camera will match in the edit, and of course the sound from the onboard mic will be excellent. Everyone is an expert in everyone else's work area. We do sound, lighting and video, and consequently are always having to find out the most important area, and working the rest around it - BUT - it needs proper planning, or the end result is dire!

Actually I'm going to give a contrary opinion here. Yes there were some underexposed shots when they got out of the light but on some shots, unless you were shooting in log, you were pretty much on the edge for highlights and couldn't have gone much hotter.

I've heard of log before, but I don't actually understand what you mean by shooting in log. Can you or someone else explain please?

Simplistically, shooting log ( different companies have different log varients Look up Slog for Sony or Vlog for Pansonic etc ) expands the dynamic range of the video beyond the rec709 of normal TV video. In editing it allows for more detail in the shadows and more detail in the highlights beyond what is possible in simple recording in rec709. This video can then be graded back into rec709 dynamic range for viewing then of course the detail in the highlights will not be lost or for the shadows. I do not believe the XF series can shoot in log but the newer Canon's can shoot with 800% gamma to extend the range.

I tend to agree with John and individual clips can be graded from your video to expand the dark areas. I edit in Edius and would use the YUV filter to bring up the levels as needed. It is easier to bring up the levels than recover highlight detail that has been lost.

Thanks Ron.

So from what I understand, put simply of course, log shooting creates a lower contrast and blander image which utilises the full colour gamut the camera can produce to be edited in post, giving an overall better quality range of darks, highlights and colours?

I'm pretty certain the XFs don't have this option, however they do shoot in 50mbps which is good for retaining a lot of data for high quality Blu-Rays.

Just wondering what people think of the overall lighting from these clips? I exposed the children for the harsh spotlights, but due to very large gaps between the lights they got darker when not under direct light. I tried to compromise as best as I could. I used the XF300 for the closer shots, and an unmanned, stationary XF100 for the wide view.

RDA Summer Show 2017 Highlights on Vimeo (https://vimeo.com/225118583)

Any feedback appreciated.

Thanks
It is under exposed (I understand why you had to do it), in post I would definitely add more gain and rolloff the highlights.

Hey folks. Got another clip to get some feedback on. This time they had some better lighting. I used a manual f-stop at 2.8 I believe for the manned camera.

In post, I took the highlights down a little, and slightly increased the darker side of the mid range as I felt it need a tiny lift, especially when the dancers enter from the sides, and for the feet/leg shots being a little dark for my liking.

I also added a tiny bit of unsharp mask for a little more definition but this was very little.

Fiinally, I used the 3-way colour corrector to make a more blue-ish/greenish tint as the faces were too red before-hand, and took the saturation down to 75 points (from 100 default).

Overall, I'm fairly happy with the end result.

Will appreciate all feedback regarding the video quality and any improvements for the future.

My shooting details are attached. I felt that 3700k was right at the time but in post the faces were too red.

Rennie Dance Academy Irish Tap Dance on Vimeo

The codec almost completely breaks down in the shadows.

Could you dropbox one minute right out of camera original of this footage? I'd like to take a look.

You could cut the video without re-encoding with ffmpeg, for 20 seconds you would use:

ffmpeg -i <video> -ss 1 -t 20 -c:v copy -an <result>

Hi Cary,

Do you mean give you a minute from the raw file or do you mean before it's colour-corrected etc? Because the raw file for one minute is around 400MB.

I've never used ffmpeg before. So this just takes my MXF file and produces the raw, cut copy?

I always use Premiere Pro H.264 (usually around 10 mbps CBR for a full HD video) then Encore to do the processing. If anyone knows a program which gives better encoding then I'm all ears.

Hi Cary,

Do you mean give you a minute from the raw file or do you mean before it's colour-corrected etc? Because the raw file for one minute is around 400MB.

I've never used ffmpeg before. So this just takes my MXF file and produces the raw, cut copy?

I always use Premiere Pro H.264 (usually around 10 mbps CBR for a full HD video) then Encore to do the processing. If anyone knows a program which gives better encoding then I'm all ears.
Can you dropbox about 20 seconds of the out of camera file?

The ffmpeg will cut 20 seconds from the MXF without reencoding (make sure the destination extension stays MXF).

I think something must have gone wrong in the workflow, if you provide some unaltered out of camera original we can determine that. 10Mb/s for an intermediate file is rather low.
Just trying to help.

There is only one solution. Teach the Dance School how to light a stage. Even parents would dislike the lighting.

Hi Cary,

Took a few hours to set it up, but finally think I managed to get 20 seconds of raw file downloaded.

It can be downloaded from here on the original format: https://www.dropbox.com/s/2fn6z3b4ornbn0i/output.MXF?dl=0

It's a 117MB XMF file.

Let me know what you think. You may have noticed how red it looks as raw. I've tried to remove this however I doubt I'm using the best tools to do so, as I only use the three way colour corrector. I also end up losing some of the green on their dresses.

I first deinterlaced the footage to 50p, adjusted the overall gain, then I adjusted the red color cast by lowering the red gain and red gamma (in linear light). Then I added some contrast to add some more black definition. I noticed your camera has some pixel errors popping up once in awhile. I would make sure your intermediate files use a high bitrate and especially for dancing or other action I would deinterlace to 50p.

Here is an unpublished link:
dance01 on Vimeo

Hi Cary,

The XF300 uses 50i, and I always shoot on the maximum of 50mbps. As far as I'm aware you can't deinterlace 50i footage to 50p, it has to be 25p.

Deinterlacing will always replace the missing field by interpolation so it doesn't matter whether its to 25P ( drop every other field and interpolate the missing one in the remaining 25) or interpolate the missing field for all 50 fields to get 50P. 50i isn't 25p with the frames split into 2 fields. The timing of the fields is exactly like 50P with identical temporal motion. Going the other way if you shoot at 50P you can get 50i for DVD or Bluray by the NLE just extracting a field from the full frame. That is what I do as I have a mix of interlace cameras ( NX5U ) and 60P cameras. I edit on an interlace timeline and the NLE ( EDIUS in my case ) just takes a field from the progressive file to make a DVD or Bluray. You can go either way but you get more detail by shooting in progressive and the progressive interpolation from interlace is very dependent of course on how good the interpolation method is for the particular subject matter.

Hi Cary,

The XF300 uses 50i, and I always shoot on the maximum of 50mbps. As far as I'm aware you can't deinterlace 50i footage to 50p, it has to be 25p.
Well, the video is there as proof, you are wrong it is possible! Download the video and verify for yourself that each frame is unique.

A camera at 50i records 50 unique half vertical resolution events per second, by interpolation and motion compensation you can deinterlace that into 50 full resolution events. If you deinterlace 50i to 25p you throw away temporal resoution which is detrimental for sports and action scenes.

50i isn't 25p with the frames split into 2 fields. The timing of the fields is exactly like 50P with identical temporal motion.
That is exactly the reason why you should frame double, by converting the footage to 25p you destroy temporal resolution.

That is exactly the reason why you should frame double, by converting the footage to 25p you destroy temporal resolution.

Isn't that what I said. To me it isn't frame doubling either. I shoot everything at 60P or 60i they both have the identical temporal motion . It is unfortunate that 60i in the standard is 29.97 fps and this is not the same as 29.97P fps. People get very confused. NTSC interlace has a sample rate of 59.94 images a second. In interlace the recording is of course just a field. 60P may cause a problem of course with equipment that expects the video to have a 29.97fps timecode to sync with for the same time period. Yes in this case it does look like a doubling. History causes all sorts of problems !!!! The timecode coincided with the sync pulse for interlace ( every two fields, that told downstream equipment which field was being received ) I assume it was easier to sync the timecode to this pulse. Although 2 fields make a frame the two consecutive fields in an interlace video do not come from the same full frame they are about 1/60 sec apart in time of course.

OF course most TV's do exactly what we are talking about when they play interlace TV. They fill in the missing fields. Depending on the TV this can be quite elaborate in using many fields to interpolate from and make the progressive image as good as possible.

Isn't that what I said.
To me it is very simple: if you have a 50i or 60i source and you want to put it on YouTube or Vimeo you turn it into 50p and 60p unless the 60i source is telecined, in that case you turn it into 24p (or 23.976p).

So if that's what you say as well we seem to agree. :)

Yes, but my response was to the "frame double" comment really. Agree with everything else you said.