What sort of camera/camcorder would you recomend for Video making and editing

grahamlthompson

In memoriam
Joined
Aug 3, 2006
Messages
25,815
Reaction score
4,434
Points
3,988
Location
Redditch
Never really used the slo-mo capability of the camera so set up a test filming a stopwatch. Firstly I don't think it's something you couldnt emulate post prod using any editor capable of frame interpolation.

The camera produces a 11.881sec clip which from looking at the video equates to around 4.5seconds of real time footage (interpolates approx 2.4 times the original frames) to produce 1080i 1440 x 1080 25 fps content.

It didn't record any audio guess it can't handle pitch correction.

Judge for yourself, test clip available here

https://www.adrive.com/public/QPrxMn/00001.MTS

Clip details

Code:
General
ID                               : 0 (0x0)
Complete name                    : E:\SloMoTest\00001.MTS
Format                           : BDAV
Format/Info                      : Blu-ray Video
File size                        : 14.7 MiB
Duration                         : 11s 881ms
Overall bit rate mode            : Variable
Overall bit rate                 : 10.4 Mbps
Maximum Overall bit rate         : 18.0 Mbps

Video
ID                               : 4113 (0x1011)
Menu ID                          : 1 (0x1)
Format                           : AVC
Format/Info                      : Advanced Video Codec
Format profile                   : [email protected]
Format settings, CABAC           : Yes
Format settings, ReFrames        : 2 frames
Format settings, GOP             : M=2, N=13
Codec ID                         : 27
Duration                         : 11s 880ms
Bit rate mode                    : Variable
Bit rate                         : 9 701 Kbps
Maximum bit rate                 : 16.0 Mbps
Width                            : 1 440 pixels
Height                           : 1 080 pixels
Display aspect ratio             : 16:9
Frame rate                       : 25.000 fps
Color space                      : YUV
Chroma subsampling               : 4:2:0
Bit depth                        : 8 bits
Scan type                        : Interlaced
Scan order                       : Top Field First
Bits/(Pixel*Frame)               : 0.250
Stream size                      : 13.7 MiB (93%)

Audio
ID                               : 4352 (0x1100)
Menu ID                          : 1 (0x1)
Format                           : AC-3
Format/Info                      : Audio Coding 3
Mode extension                   : CM (complete main)
Codec ID                         : 129
Duration                         : 11s 968ms
Bit rate mode                    : Constant
Bit rate                         : 256 Kbps
Channel(s)                       : 2 channels
Channel positions                : Front: L R
Sampling rate                    : 48.0 KHz
Bit depth                        : 16 bits
Compression mode                 : Lossy
Delay relative to video          : -80ms
Stream size                      : 374 KiB (2%)

Text
ID                               : 4608 (0x1200)
Menu ID                          : 1 (0x1)
Format                           : PGS
Codec ID                         : 144
Duration                         : 11s 375ms
Delay relative to video          : -80ms
 
Last edited:
Thanks for that, I think this is in answer to a n other thread, (but still makes sense.). However, using my Media Player I'm only seeing about 3 Sec of that stopwatch running (one must accept the thing runs true). Maybe that makes my SloMo between 75-100fps and not 200-250 as I thought.
In Sony/Vegas Movie Studio I can stretch the SloMo clip from 11s to 48s - then SVMS won't allow any more (except by rendering and bringing back). I didn't do that as the SloMo was quite impressive on its own.

You're right, there is no Audio in this Recording Mode.
That's a fine stopwatch. Can you say "when" you pressed the "Rec" button? Was that at the same moment you released the Stopwatch, perhaps? +It would be interesting to start "Rec" say at the 5-sec marker and see when the Clip begins/ends . . . .

Same tech I guess applies to GolfShot (+ I don't like the thin frames), I tried a falling brick, but the image wasn't too good and I was finding it difficult to operate the camera and brick. Knowing the length of the brick accurately and "g" wouldn't be too taxing, but I've not got that far yet. In particular the brick seemed to drift sideways slightly as I released it.

I understand that a SloMo effect might be possible using Software, but you'd only have 25 frames x 11 sec (say) to play with. Whereas here we have more frames to start with . . . the Q is: are these actual frames, or interpolations?

[ Normal Mode 3.5s x 25fps = 87 frames.
SloMo mode 11s @ 25fps = 275 frames . . . . about 3x improvement ].
If it's interpolation then the duration could be any length, but it's limited - I suspect to avoid overheating the sensor - by that logic this really is SloMo and all the frames are dumped into scratch-memory, then filed after "processing".
+If it was in-camera "post", it could be applied later to any clip, but it can't.

It really is surprising Sony doesn't explain the working (at least so Users can obtain best images) . . . . I find it a bit hit'n'miss. Furthermore, the "Processing Delay" is a bit irritating as the camera is then Locked-Out.


I've used it "on Location" where it generates some interest, since the final clip can be viewed "slow" and it's possible to see if the "hit" (or whatever), was actually caught. If a real-time clip was slowed in Post, I wonder that the "hit" might be lost anyway, due to motion-blurring? The SlowMo feature does appear to reduce blurring.
Needs a definitive look, eh?


Thanks.
 
Last edited:
You are right, wrong thread. Perhaps a moderator will move the two posts (and this one) to the correct thread. I will time the delay from pressing the button to the recording start. It's my wristwatch by the way. Pretty sure the camera films a short burst of standard frames (1440 x 1080) into a record buffer. The delay is creating the intermediate frames and copy to the SD card. Afaik this is different to the golf shot which will capture a short burst of frames at a high frame rate and simply stores these at 25fps.
 
Last edited:
Recording in Slo-Mo mode starts near instantly, no obvious delay. Golf mode is different (and different to my other Sony camera which is more like slo-mo mode). The camera is constantly buffering what is sees and uses frames from about 1.5 seconds before you push the button, subsequent frames are superimposed as vertical slices (you have to see the result to understand this)

I pushed the button with the sweep second hand at 12 o clock to get the following result

https://www.adrive.com/public/5aqd7t/Golf.MTS
 
I looked at that(link), thanks - but I can't see any movement in the hands . . . am I looking in the wrong place?

I fathomed the Golf-Shot OK, with the falling brick, it's the SloMo that needs further investigation (for me). . . . although that GS link maybe suggests I need further investigation, myself...Ooops!

Thanks.
 
I looked at that(link), thanks - but I can't see any movement in the hands . . . am I looking in the wrong place?

I fathomed the Golf-Shot OK, with the falling brick, it's the SloMo that needs further investigation (for me). . . . although that GS link maybe suggests I need further investigation, myself...Ooops!

Thanks.

Golf Shot

The second hand starts at 1.5 seconds before the 12 o clock position (where I pushed the record button) and in the last frame finishes 0.5 seconds after 12 o clock.

So when you push the record button you get a recording from about 1.5 seconds before you start the recording with a duration of about 2 seconds.

Playback time is 11 seconds

SloMo

Records about 3.5 seconds of content. Recording starts when you push the record button. Playback time is 9 seconds
 
Thanks glt, that's good to know and kinda matches what I hoped - but my SloMo is definitely 11sec playback according to MediaPlayerHC and MovieStudio time-line. eg align start at 02min00sec + 00frames ENDS 02.11 +24frames -
the +Frames figure is a running frame-count as you probably know..... That's Definitely not 9 sec., but maybe different camera-software
? I use that 11-Sec as a TAG for slomo files, as there is no other indication on my COPIED files, otherwise.
Golf-Shot pretty much tell me to take the swing and then press the Rec. button, so you don't record "misses!" However, I'm not too keen on such narrow frame slices, it would be a nice touch if sony numbered each slice, so enlargements (on paper) would be easy to assemble (in the Clubhouse).

Isn't it odd that SONY doesn't shout about these features? . . . expecting the Market to understand by a murmur on the Wind, perhaps? Other filmmakers are unaware, until shown.
_But then the BOSS stabaliser is somewhat hidden too. Any ideas how BOSS compares with Pana X920, at nearly twice the price ? I'm able to hold 800mm without a tripod, using some other support.


If I take yr Record figure 3.5sec and 11sec playback, I make the SloMo 78fps - does yr research suggest this is interpolated - or are these real/recorded frames as I suspect/hope, etc.?

+ The resulting File is real-enough, as I can stretch it in SVMS like any other clip [CTRL/Drag]....up to 2m47 + 21f in above example; showing overall timestretch 3.5s(yr figure) to: 47s +21f (=47.84sec), giving fps=340.....with the last "stretch" interpolated by the Editor.
. . . . . [ SVMS stretch-limit above = 47.84/11. ~ 4.3 times.
. . . . . So overall, it's 4.3 x 78 = 340 fps
. . . . . ]

That is amazingly slow for a consumer low-end camcorder....!



Thanks.
 
Last edited:
For Slo Mo the frames are most likely interpolated, hence the on screen processing message, and also why I said you could do similar with an editor with frame interpolation capability.

The file will be real enough, other than quality there's no way to tell which frames were invented in the resulting recording file.

I took 3 slo mo samples. The durations were one at 12 seconds and two at 11 seconds, presumably the duration isn't an exact figure.

When I get a bit of time I will try using a standard recording mode and process to get a similar effect to see what the footage looks like. I imagine a 1080p50 clip will give somewhat better results.
 
glt/Others...


Been thinking about a suitable "Test" -( as you know I believe the CX410 SloMo really is fast-framed hence the brevity of clips - rather than an interpolation technique ), . . . BUT whoKnows?

This is a "random walk"-thought along some possibilities . . . such a shame Sony doesn't Tell.

. . . . . . . . . . . . . . . . . .....Fast-track = quick-read is below the - - - - dotted line.....

The trouble is that any mechanical Test I think of has perfectly acceptable mid-points which could be interpreted by software, so proving very little. This rules out moving discs, wheels, animals/ legs etc.

Mt thoughts were electronic, using a randomly generated pattern of LEDs, such that the position of the fifth flash (say) was completely different from 4 and 6 - so without interpretation, we'd see the (real) new position whereas interpolation would show incorrectly, something midway.

I suspect an array of ten LEDs might suffice, arranged in a "random order" to prevent a happy chance interpolation. In addition it would need a 2-digit counter or similar, so the 10-Lots can be seen to be in the right order. (just in case LED 2 came next to LED 9, as it rolls-over)
Of course more LEDs + different colours, so making the random effect not just "position" - but "colour" too - If Red follows Blue, then interpolation would show yellow/green "midway" whereas we should see either the Red, or the Blue.

I used "" quotes for Midway as that will probably be determined by the position of the next conventional cine-frame ie 1/25 sec later and several LEDs

Ideally there might be ten LEDS covering all the intermediate "frames" so the next frame would be No1. *=Start) any interpolation would stay at No1, there being no other evidence available. Only a true SloMo would show the sparkling LEDs dancing about.

To achieve this the frame-rate (and LED-rate) needs to be near-equal [to avoid huge numbers of LEDs].
at 100fps 10 LEDs would only last 1/10 sec so the clocking waveform needs to be 1kHz - not difficult with crystals, although getting the exact freq might mean choosing an odd starting crystal. (if using a Binary count-down).

This shows that to maintain the Test for several seconds needs a capacious Test Method.... and I wonder that 10 LEDs really is sufficient. It's easy enough to increase the LEDs eg using CD1070 1/10 counter chips so that 40 or more LEDs can be zipping about . . . . although keeping track of them in Edit might be severely Testing.

I suspect that results should show if the SloMo is interpolation, but it would be nice to think of a simple analogue equiv . . . but I can't - since analogue is not random, it is SHM


- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

My head hurts (and anyone reading this far!) - a further thought which might be easier to build is a strobe light (using LEDs as source) illuminating a spinning top ( ideally with letters around the edge?) - if the frames are interpolated then the whole would be a blur, until the strobe freq comes very close to 25/50fps - at which point you'd get many missing frames as the sync won't be exact. Then the interpolation will be obvious - the "average" of any two illuminated frames.... whereas if the flash freq is increased the same effect would only occur at the filming-speed, which we believe to be about 78fps.
+++This scheme has the bonus that a strobe light might be useful elsewhere, whereas a random pattern generator proposed earlier is no use at all.
 
Last edited:
Different approach

Used an Android device with a timer app (Nexus 7), timer is digital with 100th second counter.

At 50fps this will give a frame resolution of 1/50th second (2 digit increments on the 100th second counter)

Filmed 1 minute from near enough 10 seconds to 1 minute 10 seconds using 1080p50.

Using Magix edit pro (with frame interpolation enabled).

Speeded up the clip a factor of 4 times to get 15 seconds of time lapse.

Slowed down the clip by a factor of 0.25 to create slomo clip of 4 minutes duration.

Both results output at 1080p25 to mp4 files.

Based on the resulting quality, I firmly believe the two functions on the camcorder are a bit gimmicky, especially as you aren't limited to a few seconds of footage, and 1440 x 1080 resolution.

Judge for yourselves

1080p50.mts (original)
https://www.adrive.com/public/fyBgmc/1080p50.MTS

FourTimesFaster (TimeLapse)
https://www.adrive.com/public/BSJu6B/FourTimesFaster.MP4


FourTimesSlowMo
https://www.adrive.com/public/xzs88H/FourTimesSlowMo.MP4
 
Last edited:
That is interesting, but I have a question regarding the display and the software that runs it - there will be no point in displaying characters faster that the display - and in all probability any faster than the (average)human eye, which is possibly 25fps at the centre - as you know the peripheral vision is quite a lot faster.

Unfortunately my Media Player doesn't allow single-frame advance but even at "normal" speed it shows some blurred frames where numbers changes (and the Media-Player chances upon the transition)

Can you provide any insight on this Android Display?
A digital stopwatch is fine for its intended purpose, but I'm yet to be convinced it would match random LEDs (or the Strobe I suggested), since LEDs are quite capable of delivering fast flashes beyond 10kHz which is about 30x the rate we are discussing, even after time-stretch, whereas a screen as seen, is really intended to display slowly-moving images. Having a slight lethargy is ideal since it will make the movies appear much smoother.

The problem is that when in SloMo it is not possible to see clearly the change between numbers . . . so I'm not able to see this either as an interpretation (by software), or as true SloMo (since the transitions are already compromised).

Thoughts...?
 
Last edited:
That is interesting, but I have a question regarding the display and the software that runs it - there will be no point in displaying characters faster that the display - and in all probability any faster than the (average)human eye, which is possibly 25fps at the centre - as you know the peripheral vision is quite a lot faster.

Unfortunately my Media Player doesn't allow single-frame advance but even at "normal" speed it shows some blurred frames where numbers changes (and the Media-Player chances upon the transition)

Can you provide any insight on this Android Display?
A digital stopwatch is fine for its intended purpose, but I'm yet to be convinced it would match random LEDs (or the Strobe I suggested), since LEDs are quite capable of delivering fast flashes beyond 10kHz which is about 30x the rate we are discussing, even after time-stretch, whereas a screen as seen, is really intended to display slowly-moving images. Having a slight lethargy is ideal since it will make the movies appear much smoother.

The problem is that when in SloMo it is not possible to see clearly the change between numbers . . . so I'm not able to see this either as an interpretation (by software), or as true SloMo (since the transitions are already compromised).

Thoughts...?

The screen is the highest resolution device currently available, it has more pixels in a 7.2" display than your full-HD TV

7.02" 1920x1200 HD display (323 ppi)
1080p HD IPS
Scratch resistant Corning® glass

It must have a refresh rate of at least 100 frames/second (you can see every number (0-99) in the 100th second countup in absolute crisp detail). As this is twice the capability of the camcorder at least it's hard to imagine anything better. If you used led's you are not any better off. Any multiple flashes within the same frame would appear as just one flash. A frame will record everything in 1/25 or 1/50 second depending on frame rate whether there's 1 flash or 50 of them. Within 1 frame the led would be on or off, it cannot be both. Then add to the mix that the process is lossy. The camcorder itself does not record every pixel for every frame. Only the I-frame in a GOP has the full pixel count.

At 323ppi it's way sharper than any TV. 1080p50 video looks fantastic. In any case the results when viewed on a HD TV speak for themselves. The footage looks way better than from the camera despite the fact I didn't take any particular care (see the reflections in the glass - it would look better if filmed in the dark with the only light source from the tablet. The bearing in mind the limitations of the source is only 2 numbers. I wouldn't fancy trying to sync a strobe device to the camera
 
Last edited:
Thanks, but those 100th that are seen-sharp are (I believe) the brain luring the viewer. When you stop the display the processor chooses which numbers to display - but so it does when the numbers tumble - to attempt to make the display look "right". That's not a weakness in the display, just the software designer choosing what's omitted to keeps things going - but if the display is not able to display very fast movement, why try to force it, with the additional battery-drain? Don't know how one might test the display, but my money's on using a different technique, probably using LEDs which are demonstratably fast light sources. LCD's are by comparison notoriously slow by orders of magnitude.

Maybe I didn't explain precisely but Test-1 was suggesting LEDs in different positions, such that the mid-point (to detect interpolation) would not be between the previous and next positions.

The other Test-2, uses the fast-flash possible from LEDs to make a strobe . . . several have been described for DIY'ers

LEDs can be used at very high frequencies, with each pulse a separate flash of light - LEDs are used as flash-guns up to very short durations, - so whilst there could be some errors in the sensor/photography element of the Test, it should not be the LED light source.

Syncing a strobe to the camera isn't necessary, as my intention was to move from under- to- over frequency, such that there is a time-region when the sync is accurate . . . . the frequency can be recorded, provided it lasts long enough for the camera to record it, say a second or so, then move the frequency.

I fear we're no nearer discovering how the Sony SloMo works and it's an interesting point you raise about the digital camera recording:
"...Only the I-frame in a GOP has the full pixel count..." as I wonder if this is done "only" to save memory - could it be convenient to reduce the data-flow through the sensor circuitry?
I'm still trying to understand why the SloMo is for such a short time and then is processed . . . why do it (if it's interpolation) only for such a short clip? this is the reason I believe there is a much faster scan in place, but proving it is far more difficult - Two owners of the same camera can't agree....
 
Maybe I didn't explain precisely but Test-1 was suggesting LEDs in different positions, such that the mid-point (to detect interpolation) would not be between the previous and next positions.

The other Test-2, uses the fast-flash possible from LEDs to make a strobe . . . several have been described for DIY'ers

LEDs can be used at very high frequencies, with each pulse a separate flash of light - LEDs are used as flash-guns up to very short durations, - so whilst there could be some errors in the sensor/photography element of the Test, it should not be the LED light source.

Syncing a strobe to the camera isn't necessary, as my intention was to move from under- to- over frequency, such that there is a time-region when the sync is accurate . . . . the frequency can be recorded, provided it lasts long enough for the camera to record it, say a second or so, then move the frequency.

I fear we're no nearer discovering how the Sony SloMo works and it's an interesting point you raise about the digital camera recording:
"...Only the I-frame in a GOP has the full pixel count..." as I wonder if this is done "only" to save memory - could it be convenient to reduce the data-flow through the sensor circuitry?
I'm still trying to understand why the SloMo is for such a short time and then is processed . . . why do it (if it's interpolation) only for such a short clip? this is the reason I believe there is a much faster scan in place, but proving it is far more difficult - Two owners of the same camera can't agree....

You are failing to understand how a moving image is recorded. At 50fps a single frame contains video information that happens during the period the shutter is open, outside this period no extra information will be captured until the next shutter opening 1/50 second later captures more information. Anything that happens between the two events isn't even captured. If the shutter is open for the full frame then everything that happens will appear as happening at the same time as the resolution is only 1/50 second. So variable number one is the shutter speed.

Next you need to know what makes High Definition video possible in the first place. It's the lossy mpeg compression system required to allow storage of the large amount of data. HD camcorders use the most efficient current form of mpeg compression (H264/AVC - aka AVCHD).

Mpeg compression works basically by capturing one full frame of data, then records subframes that contain only difference information. The compression system is lossy (like jpeg used for still photos). The i-frame and it's subframes form a group of pictures (GOP), the lower the bitrate the more information is discarded and the more information has to be guessed in the mpeg decoding process.

Look at the basic maths.

One pixel has 3 subpixels, one each of red greeen and blue. Each subpixel normally uses 8 bits of data making a total of 24 bits for each pixel (around 64000 variations in colour). A 1920 x 1080 picture contains 2073600 pixels or 49766400 bits of information. The data is stored in parcels of 8 bits known as a a byte (each pixel needs 3 bytes). Therefore 1 screen requires a minimum of 6220800 bytes for each frame. At 50 fps thats 311040000 bytes for each second of video. Divide by 1000 x 1000 * 1000 giving 0.31 GB/second (A data rate of 310MB/sec or 2480 Mbps ). A 16GB storage card would hold just 51 seconds of uncompressed video.
 
A useful Digital explanation, though I doubt our modern cameras contain shutters operating at anything like 1/50 sec - the noise would be obvious. Or was this an "Examplar" not actuality?

Yet, it really matters little how the frames are created, since my real interest is whether they come from interpolation or high-speed actual recorded pixels, - as I see it.

Looking at the Timeline on an Editor of a digital movie-clip - you can examine each "frame".
- It may be these are not real (as in a cine film), but they do represent the movement of recorded light between each element of Time - and it is these changes that we see when the Rendered file is watched on a TV/Projector.


Whilst it is useful/informative, I'm not sure how accurate knowledge of mpeg-compression helps determine how the Sony SloMo works - since the mpeg process is applied whatever frame-rate is set to.

Indeed we cannot determine (can we?), what method is used at the sensor to produce the stream the mpeg compression starts with. I suspect camera-makers are very coy about telling the Public how their sensors turn an image into a stream of pixel data. They tell us how many pixels there are, since this equates (in the Buyers' mind) to "image quality" any further detail might be commercially-sensitive . . . but for most Folk it would be meaningless.
( Many of us hear bold statements: "I like Brand xxx because I had issues with Brand yyy" )

What I was trying to agree was a Test method that would identify whether the SloMo frames are interpolation, or gathered at a higher speed and processed to extend time, compared with the normal process.

Whilst yr images of the Android stopwatch is v. interesting, I don't think LCD screen-speed is sufficient to allow any definite conclusion . . . and any analogue dial is also flawed, since the high-speed image will be very similar to one obtained by interpolation . . . . or can you propose anything from the footage you got?
That's why I think an optical method of using pulsed LEDs should produce some useful results . . . if only I had the time to do it.
Presently I'm rewiring the workshop to control heating/power and hope soonish to have the ability to make some of the many "projects I have started" to be concluded - it won't be v.soon, I suspect.


...Good Luck....
 
Last edited:
A useful Digital explanation, though I doubt our modern cameras contain shutters operating at anything like 1/50 sec - the noise would be obvious. Or was this an "Examplar" not actuality?

The shutter is electronic not mechanical and is completely silent.

The Sony spec is low on details apart from mentioning low light performance at 1/25 and 1/30 second. Why would a shutter time half as long as 1/25 second make any noise :confused:

Look at the spec here for a Panasonic HC-X920, shutter speed range is 1/50-1/8000 of a second.

Camcorders - HD Camcorders - HC-X920 - Specification - Panasonic UK & Ireland

EDIT - found this for the CX410VE

Standard:6 lux(1/50 Shutter Speed), Low Lux:3 lux(1/25 Shutter Speed)

http://www.sony.co.uk/product/cam-h...ry-stick/hdr-cx410ve/technical-specifications
 
Last edited:
Found an Android Timer with both analogue and digital displays. Filmed it using slomo gave a recording of 229 frames (11.96 seconds)

Frame analysis reveals the effect is created by frame hold and intermediate lap dissolves.

eg

Frames 1-5 are identical with the elapsed time at 56.725 sec
Frame 6 is a dissolve to Frame 7
Frames 7 - 18 elapsed time is 56.792 sec
Frame 19 dissolve to Frame 20
Frames 20-31 elapsed time 56.858 sec
Frame 32 dissolve to Frame 33
Frames 33-34 elapsed time 56.924 sec
Frame 45 dissolve to Frame 46
Frames 46- 58 elapsed time 56.991 sec
Frame 59 dissolve to Frame 59
59 elapsed time 57.057

Gave up here :D, if anyone wan't to look at the clip post and I will upload it to FS site.

Looking at the last frame elapsed time is 59.750 giving on screen elapsed time of 59.750-56.725 = 3.025 seconds

Normal time would be 11.96 seconds giving near enough 0.25 times playback.

Should be very easy to emulate in a video editor, might try it with full 1080p50 source video.
 
As I read yr post #14 with mention of a shutter being open.
Sorry to say I'm not aware of the technique used in sensors to switch-on-off and disregard changes in the incoming light/image. Presently, I believe the sensor is continually bathed in light, with some variation due to the mechanical Iris, but looking at the spc it seems the iris has quite a small range of f/values - quite how that works with zoom, effectively changing the f/number. (Although some techniques claim to make a constant-f-stop is regularly possible).

That quote from Sony is in the Manual, but I've noticed very little real improvement - but when it's really dark it's better than nowt.

Whilst it a great effort (and Thanks), I am not a great believer in Tests using an LCD - these frame-skips/merges could be the software that drives the display. To be acceptable I think a different approach is needed - but for practical purposes it may not matter...

The people I've shown the Sony footage to, believe it to be superior to their own efforts using GoPro . . . and You'll recall that there was much excitement as GoPro developed their range, with significant claims as to SloMo (Black Model).
The problem appears to centre on the much smaller frame GoPro provides; this degrades the image, so Sony has scored much Kudos - but somehow it doesn't appear to have reached the Sony Marketing persons.
 
The change in f numbers is accomodated as in pretty well all modern cameras by the metering being behind the lens. It's simply impossible for the lcd display to have any effect, there are groups of 11 frames each one is identical so the camera is repeating the same frame 11 times. This is an effect known as frame hold in video editing programmes.

LCD displays are progressive they show a complete frame unlike a CRT display. They are capable of very high frame rates as in high quality gaming monitors. I imagine your TV will show a 1920 x 1080 progressive signal at 60 fps without breaking into a sweat.

By the way, your TV, the display on your camcorder, the monitor on your PC/laptop all use back lit lcd displays, do they all tell lies ? In the end everything you record finishes up on a LCD display (unless of course you use a CRT display)

If you look at the original chronograph footage you will see the sweep second hand freezes in exactly the same way for groups of 11 frames. Instead of postulating vague theories why not take close look ?


See for yourself :eek:

https://www.adrive.com/public/uyTRZ9/NewTimerslomo.MTS

Stop waffling and have a look the evidence is all there.

This frame is repeated 11 times, how could the LCD display do this :confused:

The next repeat of a group of 11 frames shows 56.858 seconds.

It may be a handy feature, it's deffo something you could do with any camera and a suitable video editor though it may be a bit laborious unless you have a video editor with macro control capabilities.

You could do it with a short clip in photoshop for instance.
 

Attachments

  • slomotest.jpg
    slomotest.jpg
    78.2 KB · Views: 60
Last edited:
This may be an interesting project for those really interested in high speed videography.

Imagine 1280x720 @ 700 fps!

edgertronic - The first affordable high speed video camera. by edgertronic — Kickstarter

Bank Balance not up to it :(

Exported 2 frames at the end of the 56.792 frames, the transistion frame and the start frame of the 56.858 block to Photoshop. Resized and cropped to a sensible size and pasted the layers into a seperate document so as to see the frames side by side. To me the intermediate frame looks like a single frame created by a 1/25 lap dissolve from the last frame of the first block and the first of the next.

Could be a usefull way of generating slomo from a standard clip if someone is clever enough to automate the process.
 

Attachments

  • FrameTransistion.jpg
    FrameTransistion.jpg
    100.4 KB · Views: 42
glt, not sure I follow this - I'm "on hold" for SloMo, but am giving some thought to a strobe using LEDs to allow filming of moving things . . . should be fun. Some Internet builders suggest that above 100flashes/sec the white LED's phosphor can create blurring of the effect, which doesn't surprise me. So, I might try a mix of RGB, - the snag being that Blue LEDs are low-output and more expensive....
This Strobe, need not use the CX410 SloMo feature, but with some trials might allow further research if of interest.

Chelters - WOW!
That is some camera for those with deep pockets . . . although for film-shoots a modest price . . . OR maybe for Hire? - it looks like a small team has come up with the right mind-set. In particular I like the touch where they give the DC connector spec...nice-one!

Their movie of an alarm-clock bell-ringing is an excellent slow-mo subject, being easy to try and not weather dependent . . . . . . it's the kinda-subject my above (proposed, only), LED-Strobe should be capable of doing . . . . obviously NOT as good, but seeing the hammer moving will be quite striking*, if it was an insert in a sinister movie of someone waking-up. I "see" the hammer being a shadow on the wall . . . of course I'd cheat! By using a point-source light and moving the hammer with a string . . . something about the best shots being fakes . . . er, I read somewhere.

But the LED-Strobe should be capable of very similar, but the colour-sources ( being spread ), would spoil it for a movie audience, which would destroy the scene...er, IMHO..


*Ho-Ho!
 
Last edited:
glt, not sure I follow this - I'm "on hold" for SloMo, but am giving some thought to a strobe using LEDs to allow filming of moving things . . . should be fun. Some Internet builders suggest that above 100flashes/sec the white LED's phosphor can create blurring of the effect, which doesn't surprise me. So, I might try a mix of RGB, - the snag being that Blue LEDs are low-output and more expensive....
This Strobe, need not use the CX410 SloMo feature, but with some trials might allow further research if of interest.

Chelters - WOW!
That is some camera for those with deep pockets . . . although for film-shoots a modest price . . . OR maybe for Hire? - it looks like a small team has come up with the right mind-set. In particular I like the touch where they give the DC connector spec...nice-one!

Their movie of an alarm-clock bell-ringing is an excellent slow-mo subject, being easy to try and not weather dependent . . . . . . it's the kinda-subject my above (proposed, only), LED-Strobe should be capable of doing . . . . obviously NOT as good, but seeing the hammer moving will be quite striking*, if it was an insert in a sinister movie of someone waking-up. I "see" the hammer being a shadow on the wall . . . of course I'd cheat! By using a point-source light and moving the hammer with a string . . . something about the best shots being fakes . . . er, I read somewhere.

But the LED-Strobe should be capable of very similar, but the colour-sources ( being spread ), would spoil it for a movie audience, which would destroy the scene...er, IMHO..


*Ho-Ho!

Can't see what you hope to achieve. The camcorders fastest shutter speed appears to be 1/50 second used when shooting 1080p50. Everything it sees in 1/50 second will be in the single frame. This is the physical limit to what the camcorder itself can do. Done some testing by shooting at 1080p50 and then processing the footage in Magix to 0.25 times speed and frame interpolation. The results look very good using the analogue/digital screen as a target. I need to find some real action shots to test further. Thinking of using granddaughter on garden swing but weather today precludes it.
 
You've misread my posting, I fear I may not have explained well. Para 3
"...alarm-clock bell-ringing is an excellent slow-mo subject, being easy to try and not weather dependent . . . . . . it's the kinda-subject my above (proposed, only), LED-Strobe should be capable of doing ...
The hammer image is repeated at high speed (on a 2-bell alarm clock). By adjusting the Strobe to the hammer frequency, it will appear stationary.
- - - - - - - - - - Thus you can film a stationary hammer.
Then by shifting the Strobe Freq. the image moves slowly (forward/back) although being a simple hammer it won't make much difference.
The shutter-speed, frame-rate, or whatever, makes no difference, -provided the exposure is right and image is behaving.
That's OK now.

To test the CX410 "SloMo" feature I think a different approach may be needed . . . I'm hoping to modify the (earlier Post) Strobe - possibly non-white LEDs in a circle which come on in almost diametric positions. If the camcorder uses interpolation, then it will either show both LEDs On (even though they never are), or a middle position which passes the centre (where there is no LED). ( Thoughts???)
I expect to add a counter display, so I can let it run for some cycles.
 
Last edited:

The latest video from AVForums

TV Buying Guide - Which TV Is Best For You?
Subscribe to our YouTube channel
Back
Top Bottom