How sharp can you get this?

Brendan Marnell · October 22nd, 2007, 03:42 PM

Thanks for asking ...

SD / 4:3 / 720 x 576 / Orig. clip 1min 12secs / Bitrate 1536kbps / Frame Rate 25/sec File 262Mbs / OIS : On

Edited/cut down in PPro 1.5 to 3sec + 2sec clips. Both clips duration increased by 1 second (to prolong the good bits!!). Final .avi = 7secs. and 26Mbs @ Bitrate 1024 kbps. 720 x 576
No other editing.

Compressed to .mov + .wmv files with Sorenson Squeeze.

Shot with Canon XM2 in April 2007.

[First clips of flight using tripod taken Oct.2007]

Eric Shepherd · October 22nd, 2007, 03:45 PM

It sounds like the OIS is different in the XM2 than the one in the XL2. I think that may be the problem.

Russ Holland · October 22nd, 2007, 06:41 PM

Quote:

Originally Posted by Brendan Marnell

Both clips duration increased by 1 second (to prolong the good bits!!).

Surely slowing the clip down a bit will make the footage slightly jerky, and perhaps ghost the image in places? Have you deinterlaced the footage?

Eric Shepherd · October 22nd, 2007, 06:43 PM

Quote:

Originally Posted by Russ Holland

Surely slowing the clip down a bit will make the footage slightly jerky, and perhaps ghost the image in places? Have you deinterlaced the footage?

Yeah, I was thinking that as well. Stretching stuff can cause all kinds of weird random stuff.

Chris Soucy · October 22nd, 2007, 10:14 PM

Sorry, had gales here and lost my broadband most of the day.

Eric,
I take your point about the OIS, nevertheless, OIS when mounted on a tripod does weird things.

Brendan,

As a matter of interest, have you done a frame by frame check on this footage before you laid a glove/ hand/ NLE on it? (I mean of the bits you finally ended up with).

Do any of the artifacts displayed in the final clip exist in the originals?(ghosting etc).

It would help to figure out just where in the food chain this stuff came from.

Don't know whether your XM2 has the facility, I think on the A1 you can practically scroll through stuff a frame at a time on camera. If you can check the originals of the stuff you posted and it's clean as a whistle, then it's the NLE somewhere. If it ain't clean on the camera, "Houston, we have a problem!".

CS

Brendan Marnell · October 23rd, 2007, 03:43 PM

Quote:

Originally Posted by Eric Shepherd

It sounds like the OIS is different in the XM2 than the one in the XL2. I think that may be the problem.

I can't establish what other settings I had at the time Eric but my understanding of OIS was even less at the time.

Russ said ...
Surely slowing the clip down a bit will make the footage slightly jerky, and perhaps ghost the image in places? Have you deinterlaced the footage?

You're spot on Russ, on both counts ... I also forgot to deinterlace

Chris
I did do a frame by frame; but only after making some of the mistakes listed above and before making the others.

Am I the only guy around who thinks he'll live long enough to make all the mistakes, twice??

What I really mean is thank you all sincerely for helping me identify my latest batch of blunders.

Eric Shepherd · October 24th, 2007, 04:48 AM

Don't worry, there's plenty of upcoming blunders. We'll all get our fair share. ;)

Russ Holland · October 25th, 2007, 08:52 AM

Just one more thing...

Did you stabilise the shot using SteadyMove or something similar? I looked over some footage of mine last night and noticed very similar ghosting in some of the shot and I definately remember that this particular shot i used SteadyMove on it.

Just another thought...

Brendan Marnell · October 25th, 2007, 01:00 PM

I don't know anything about Steadycam, Russ. I agree completely with your idea that "stretching" handheld OIS footage is a recipe for the wobbles, especially when I consider that stretching 2 seconds by 1 second amounts to a 50% stretch ... I remind myself of the folly of the chap who told his girlfriend not to worry because she was "only a little bit pregnant " Do I hear a chorus of "Wassat got to do with anything ?"

Eric Shepherd · October 25th, 2007, 03:40 PM

Quote:

Originally Posted by Brendan Marnell

I don't know anything about Steadycam, Russ. I agree completely with your idea that "stretching" handheld OIS footage is a recipe for the wobbles, especially when I consider that stretching 2 seconds by 1 second amounts to a 50% stretch ... I remind myself of the folly of the chap who told his girlfriend not to worry because she was "only a little bit pregnant " Do I hear a chorus of "Wassat got to do with anything ?"

Wassat got to do with anything? ;)

SteadyMove is a post production stabilizer plugin. That's what I was originally asking, if anything had been added to stabilize this, giving the weird lines at the top of the footage and the slight blurring, etc.

I'm sure the stretching hasn't helped, but it sounds like your XM2's stabilizer (is this like the GL2 in the US?) has an electronic image stabilizer (EIS) rather than optical, as I described way back. It works by sliding around a blown up version of the image, to keep it stable, but doesn't always work perfectly.

How does your footage look without time stretching it?

Eric

Brendan Marnell · October 31st, 2007, 04:19 AM

Eric and anyone, one more look ... no stretching or doctoring, only deleting in PPro & compression with Sorenson Squeeze.. .mov file was refused upload.

On page 44 of XM2 (GL2) Manual OIS claims specifically to be superior to EIS.

Alan Craven · October 31st, 2007, 04:47 AM

Eric, the XM2 is indeed the camera you know as the GL2. Like the GL2 its built in stabiliser is optical.

Brendan, did you have the camera stabiliser turned off when you took this shot? You really need to when you are following a bird in flight, as otherwise the optical stabiliser sees any movement of the bird within the frame as needing stabilising. With the stabiliser on the result is that the bird often appears to have a rather step-like motion.

This occurs because the stabiliser "holds up" the detected motion until it reaches its limit, and then has to let go.

Canon recomment in the manual that you do not use the camera stabilisation when using a tripod.

Eric Shepherd · October 31st, 2007, 07:06 AM

As far as I know, the lens doesn't 'watch' the subject, it only senses the movement of the lens, I believe. That's why it tries to stabilize the shot when you go from non-movement to a pan or tilt and the image 'sticks' and then releases and moves. That's why it's recommended on the DV Creators.net XL2 video for zoomed in shots on a tripod. (granted that's a fixed or zooming shot, without moving the camera's position via tilt or pan) Then it can correct for shaking ground, wind, etc.

I've had good results using the stabilizer on my Miller tripod, but I've been following prop and jet aircraft at airshows, not with birds yet. From my experience it doesn't seem to 'stick' with the shot as long as the camera keeps moving around, so it's good for following movement as long as I don't stop moving. :)

Is it possible that the jarring of the lens causes the prisms to flicker a little bit, and it's more of a limitation of the stabilizer than an artifact of the stabilizer being enabled? It doesn't look like the lens would be jarred much in this shot though, but maybe the smaller lens is less tolerant?

I guess the best way to find the answer here would be just to find some random wildlife and follow it with and without the stabilizer and then check both versions on playback and see which looks better?

Eric

Brendan Marnell · October 31st, 2007, 08:12 AM

Quote:

Originally Posted by Alan Craven

Brendan, did you have the camera stabiliser turned off when you took this shot? You really need to when you are following a bird in flight, as otherwise the optical stabiliser sees any movement of the bird within the frame as needing stabilising. With the stabiliser on the result is that the bird often appears to have a rather step-like motion. This occurs because the stabiliser "holds up" the detected motion until it reaches its limit, and then has to let go.

Canon recomment in the manual that you do not use the camera stabilisation when using a tripod.

Alan, I hope these statements mean that OIS is suitable for handheld clips. Please correct me if I'm wrong. I could post several handheld clips of vulture flight using OIS in which the bird is fairly sharp and the background is jagged, especially cliffs in near background. I wonder is that the step-like motion you refer to? Virtually all my bird flight footage is handheld + OIS. I have only used a tripod for landings and take-offs at one favourite roost, never for gliding or soaring. I think it was Don DesJardin told me that I could help myself anticipate footage that might be worth stretching by upping my frame rate in advance to 50fps.

The vulture clips at this site are all handheld + OIS @ 25fps without stretching ...
http://www.hbw.com/ibc/phtml/votacio...?idVideo=15072

Eric, I will try your suggestion " guess the best way to find the answer here would be just to find some random wildlife and follow it with and without the stabilizer and then check both versions on playback and see which looks better"

Alan Craven · October 31st, 2007, 11:28 AM

Eric,

"Sees" was an unfortunate choice of word - there is obvious ambiguity! The movement of the bird within the frame that I was talking about would be due to a jerky movement of the camera, due to the pan/tilt action needed to follow the bird not being entirely smooth as the bird changes flight velocity and the operator tries to follow it. You are right about the image not "sticking" provided the camera motion is smooth(ish), but birds in flight are rather more skittish than the average aircraft! You should try to follow a randy male hen harrier when it is sky-dancing to impress potential mates! The old RAF Harrier jump-jet is a couch potato by comparison.

Brendan,

There are some very odd effects in the background of your vulture shot, especially when the camera is moving quickly. I have never seen any effects like that with my attempts at similar shots. This looks as though you could have been to Jaca in the Spenish Pyrnees?

I have just been working with some hand held shots of Red Kite in flight and found that there is visible "stepping" of the motion - seen in the background rather than the bird - in shots where I had the OIS on.

I also have a small Canon with electronic stabilisation, shots of Kites taken with that are far worse where I forgot and left the stabilisation on.

I am gradually coming to the conclusion that in camera stabilisation is a waste of time at longer focal lengths. It is certainly effective for hand held wide angle, but, I fear not much else. With the camera on a good tripod/head combination, it does not seem to work well, and ruins many shots, especially if there is deliberate camera movement.

Incidentally, even with a huge heavy tripod, and the best of heads, you will get image movement with a long lens simply because of air-disturbance in the light path. Image stabilisation has almost no chance of correcting that because the motion isn't uniform over the image. You just have to live with that.

October 22nd, 2007, 03:42 PM	#16
Brendan Marnell Major Player Join Date: Dec 2004 Location: Dublin, Ireland Posts: 938	Details Thanks for asking ... SD / 4:3 / 720 x 576 / Orig. clip 1min 12secs / Bitrate 1536kbps / Frame Rate 25/sec File 262Mbs / OIS : On Edited/cut down in PPro 1.5 to 3sec + 2sec clips. Both clips duration increased by 1 second (to prolong the good bits!!). Final .avi = 7secs. and 26Mbs @ Bitrate 1024 kbps. 720 x 576 No other editing. Compressed to .mov + .wmv files with Sorenson Squeeze. Shot with Canon XM2 in April 2007. [First clips of flight using tripod taken Oct.2007]

October 22nd, 2007, 10:14 PM	#20
Chris Soucy Inner Circle Join Date: Feb 2007 Location: Fairfield, Dunedin, New Zealand Posts: 3,691 Images: 18	Hi Guys........ Sorry, had gales here and lost my broadband most of the day. Eric, I take your point about the OIS, nevertheless, OIS when mounted on a tripod does weird things. Brendan, As a matter of interest, have you done a frame by frame check on this footage before you laid a glove/ hand/ NLE on it? (I mean of the bits you finally ended up with). Do any of the artifacts displayed in the final clip exist in the originals?(ghosting etc). It would help to figure out just where in the food chain this stuff came from. Don't know whether your XM2 has the facility, I think on the A1 you can practically scroll through stuff a frame at a time on camera. If you can check the originals of the stuff you posted and it's clean as a whistle, then it's the NLE somewhere. If it ain't clean on the camera, "Houston, we have a problem!". CS

October 25th, 2007, 08:52 AM	#23
Russ Holland Regular Crew Join Date: Mar 2007 Location: Burton on Trent, UK Posts: 193	Just one more thing... Did you stabilise the shot using SteadyMove or something similar? I looked over some footage of mine last night and noticed very similar ghosting in some of the shot and I definately remember that this particular shot i used SteadyMove on it. Just another thought... __________________ Sony Alpha a57 \| RODE VideoMic \| Adobe Premiere CS5 Manfrotto 785b \| Manfrotto 718b

October 22nd, 2007, 03:45 PM	#17
Eric Shepherd Major Player Join Date: Feb 2007 Location: Orlando, FL Posts: 404	It sounds like the OIS is different in the XM2 than the one in the XL2. I think that may be the problem.

October 24th, 2007, 04:48 AM	#22
Eric Shepherd Major Player Join Date: Feb 2007 Location: Orlando, FL Posts: 404	Don't worry, there's plenty of upcoming blunders. We'll all get our fair share. ;)

October 25th, 2007, 01:00 PM	#24
Brendan Marnell Major Player Join Date: Dec 2004 Location: Dublin, Ireland Posts: 938	I don't know anything about Steadycam, Russ. I agree completely with your idea that "stretching" handheld OIS footage is a recipe for the wobbles, especially when I consider that stretching 2 seconds by 1 second amounts to a 50% stretch ... I remind myself of the folly of the chap who told his girlfriend not to worry because she was "only a little bit pregnant " Do I hear a chorus of "Wassat got to do with anything ?"

October 31st, 2007, 04:47 AM	#27
Alan Craven Major Player Join Date: May 2003 Location: York, England Posts: 518	Eric, the XM2 is indeed the camera you know as the GL2. Like the GL2 its built in stabiliser is optical. Brendan, did you have the camera stabiliser turned off when you took this shot? You really need to when you are following a bird in flight, as otherwise the optical stabiliser sees any movement of the bird within the frame as needing stabilising. With the stabiliser on the result is that the bird often appears to have a rather step-like motion. This occurs because the stabiliser "holds up" the detected motion until it reaches its limit, and then has to let go. Canon recomment in the manual that you do not use the camera stabilisation when using a tripod.

October 31st, 2007, 07:06 AM	#28
Eric Shepherd Major Player Join Date: Feb 2007 Location: Orlando, FL Posts: 404	As far as I know, the lens doesn't 'watch' the subject, it only senses the movement of the lens, I believe. That's why it tries to stabilize the shot when you go from non-movement to a pan or tilt and the image 'sticks' and then releases and moves. That's why it's recommended on the DV Creators.net XL2 video for zoomed in shots on a tripod. (granted that's a fixed or zooming shot, without moving the camera's position via tilt or pan) Then it can correct for shaking ground, wind, etc. I've had good results using the stabilizer on my Miller tripod, but I've been following prop and jet aircraft at airshows, not with birds yet. From my experience it doesn't seem to 'stick' with the shot as long as the camera keeps moving around, so it's good for following movement as long as I don't stop moving. :) Is it possible that the jarring of the lens causes the prisms to flicker a little bit, and it's more of a limitation of the stabilizer than an artifact of the stabilizer being enabled? It doesn't look like the lens would be jarred much in this shot though, but maybe the smaller lens is less tolerant? I guess the best way to find the answer here would be just to find some random wildlife and follow it with and without the stabilizer and then check both versions on playback and see which looks better? Eric

October 31st, 2007, 11:28 AM	#30
Alan Craven Major Player Join Date: May 2003 Location: York, England Posts: 518	Eric, "Sees" was an unfortunate choice of word - there is obvious ambiguity! The movement of the bird within the frame that I was talking about would be due to a jerky movement of the camera, due to the pan/tilt action needed to follow the bird not being entirely smooth as the bird changes flight velocity and the operator tries to follow it. You are right about the image not "sticking" provided the camera motion is smooth(ish), but birds in flight are rather more skittish than the average aircraft! You should try to follow a randy male hen harrier when it is sky-dancing to impress potential mates! The old RAF Harrier jump-jet is a couch potato by comparison. Brendan, There are some very odd effects in the background of your vulture shot, especially when the camera is moving quickly. I have never seen any effects like that with my attempts at similar shots. This looks as though you could have been to Jaca in the Spenish Pyrnees? I have just been working with some hand held shots of Red Kite in flight and found that there is visible "stepping" of the motion - seen in the background rather than the bird - in shots where I had the OIS on. I also have a small Canon with electronic stabilisation, shots of Kites taken with that are far worse where I forgot and left the stabilisation on. I am gradually coming to the conclusion that in camera stabilisation is a waste of time at longer focal lengths. It is certainly effective for hand held wide angle, but, I fear not much else. With the camera on a good tripod/head combination, it does not seem to work well, and ruins many shots, especially if there is deliberate camera movement. Incidentally, even with a huge heavy tripod, and the best of heads, you will get image movement with a long lens simply because of air-disturbance in the light path. Image stabilisation has almost no chance of correcting that because the motion isn't uniform over the image. You just have to live with that.