Sanyo HD1 footage!

Joseph Aurili · February 14th, 2006, 04:57 PM

Resolution chart test clip:

www.gamersden.com/hd1test/chart.MP4

Tell me how many lines?

I see that distracting adjustment problem again in the first few seconds of video, and this is fixed lighting...

One cool thing about the camera is you can control the entire menu system from the remote, so I am able to change the images settings without even touching the camera. And the remote is even easier to control the the joystick on the camera.

--- Joseph

Joseph Aurili · February 14th, 2006, 05:12 PM

OK, I got it done early. Here are the results from the different image settings. I would go with normal as the best. The vivid has too much color. The soft just looks more like SD. The lighting gives a huge improvement in the noise. The camera selected F3.5 at 1/30 of a second. It does not get much better then that ;) With controlled lighting I am actually impressed with the quality. Of course nothing moving in the shot which should make it real easy for the compression.

--Normal kitchen lights normal setting--

www.gamersden.com/hd1test/kitchen.JPG
www.gamersden.com/hd1test/kitchen.MP4

--Better lighting--

Normal setting:

www.gamersden.com/hd1test/normal.JPG
www.gamersden.com/hd1test/normal.MP4

Vivid setting:

www.gamersden.com/hd1test/vivid.JPG
www.gamersden.com/hd1test/vivid.MP4

Soft setting:

www.gamersden.com/hd1test/soft.JPG
www.gamersden.com/hd1test/soft.MP4

Soft Vivid setting:

www.gamersden.com/hd1test/softvivid.JPG
www.gamersden.com/hd1test/softvivid.MP4

--- Joseph

Robert Jackson · February 14th, 2006, 05:37 PM

Quote:

Originally Posted by Joseph Aurili

The thing that is bother me the most now is that in many of the video clips, the camera is adjusting the picture is a very visible way for the first 5 seconds or so.

About an hour ago someone let me mess with an HD1 for about 15 minutes or so. It was a gray-market camera, so it may not be a good example, but I found this same thing to be problematic. It hunts around for a few seconds before it settles down. Same thing with focus, though. It seemed very slow to focus and did some hunting. Not so much of an issue for interviews, since you have the opportunity to set up the shot before shooting, but it seems like it would make spontaneous shooting kind of tough.

Joseph Aurili · February 14th, 2006, 05:51 PM

Robert, i'm hoping I can use one of the manual modes to eliminate that problem. I hate to make the clips larger then they need to be just to avoid that issue.

Robert Jackson · February 14th, 2006, 05:57 PM

Quote:

Originally Posted by Joseph Aurili

Robert, i'm hoping I can use one of the manual modes to eliminate that problem. I hate to make the clips larger then they need to be just to avoid that issue.

I didn't have time to really get into the camera at all, but it seems to me that the manual modes will probably be very helpful. It's so small. I have big hands and couldn't really control it at all. I'm sure that something you get used to, but I felt completely out of my element. I took a couple of stills of my girlfriend with it, though, and they seemed pretty nice. It will be great to me able to take production stills with the same camera I'm using to do interviews.

Jef Bryant · February 15th, 2006, 02:32 AM

Interesting but disheartening notes from the manual:

"If the exposure setting is "M" (manual) and the ISO setting is set to OTHER THAN ISO-A (auto), the exposure correction setting will be disabled."

Well, ain't that a kick in the head.

And a special note to PAL users:

When the TV output mode is set to PAL and the S-AV cable is connected, "during recording the image appears only on the (on-camera) monitor, it does not appear on the (external) monitor."

I guess an external monitor will work in ntsc, though.

I suppose this cam is not really intended for use by those who would want good manual control and external monitors anyway...

David Kennett · February 15th, 2006, 07:30 AM

Jef,

I read that "exposure CORRECTION", not manual settings. That is, if you set everything manually, there's no need to TRIM exposure.

David Kennett · February 15th, 2006, 09:00 AM

Joseph,

I read the chart as about 550 lines H & V. H being just a tad better than V. You could go just a bit tighter, but I doubt that it will improve the numbers much. The total H res is about 1000 lines, so the numbers come in a little below the theoretical maximum. Not bad for a camera of this ilk.

Most troubling is the red cast on the chart. Did it print out that way? Now that I look closer, I think the problem is not too much color, but an imbalance in the colorimetry. Are you using auto color? Try an appropriate fixed, or the manual with a white sheet of paper under the light you'll be using. If the chart printed out with an appropriate grey scale, you should see each of the grey squares as grey, and you should be able to see ALL of the squares. Of course you have to get it printed properly first. I will PM you a frame from JVC HD10. It's JPEG, but degraded only a little.

Wayne Morellini · February 15th, 2006, 09:06 AM

Just looked at some of the kitchen images, I won't bother suggesting a 15 watt light bulb version, the results would be far to predictable.

Typical reduced light range, high noise and reduced low light performance of a small mega-pixel CCD. They would have been better sourcing a cmos sensor from fill-factory or smalsensor (now both owned by Cypress semiconductor) or Prixim (which seems too much like smalsensor). They offer multi-slope and autobrite latitude extending techniques that produces latitude ranges upto 20 stops even on small cheap chips, and fill-factory has a technique that samples the light missing the sensor pad, eliminating the screen door technique, with high Signal to Noise ratio and high native latitude.

The multi- sampled colours are as I predicted, great, and some people around these forums might actually prefer the noise as a film look. Actually I quiet like it, I think it might turn out to be a funky legend of a consumer camera ;)Only around 15-45 days before we see the h264 versions.

About the camera. If you want to increase compression performance let the highlights over expose. I'm examining the videos as I watch House, so I'm in one of those analytical moods again. The amount of noise in the scene will be dragging compression performance down which could be better used on existing detail. Upping the exposure should reduce this. The great anti-aliasing screen door effect isn't helping either. I don't even think it is doing pixel binning (combines pixel signal output from a group of pixels to act as a larger pixel and increase the single strength being sampled, reducing noise and increasing relative latitude. As it is, it looks like they need to use a better interpolation routine to reduce the screen door effect (to cross filter out the anti-aliasing using the separate RGBG pixel channels). It is curious the the sample shots of the baby and ice cream worked so well, but I think it is because it was actually because the light brightness and latitude on the day was good enough, but it almost seems like another camera. Was that guy using a lens attachment?

Joseph, there is possibly a way to get a better image from this camera. There has been commercial slr lens adaptors that use a triplet condenser instead of a screen, so you get clean extra bright image with a number of stops more light. Used in conjunction with a suitable ND filter you could possible extend the range and cleanness of the camera. Tiffen had some interesting filter kits, and contrast filters.

http://www.optexint.com/sales/filters/filterkits.html
http://www.tiffen.com/tiffen_filters.html

Worth looking through the pdfs on each type in the kit (I think there are newer kits available to).

http://www.tiffen.com/tiffen_product_literature.html
http://www.tiffen.com/userimages/filter-lens.pdf

This contrast filter thing it interesting, appears counter intuitive, I am interested in how the technology behind it works. Because, despite their glowing description, I don't really how it is much of an improvement over a ND in most cases, anyone know?

http://www.tiffen.com/userimages/Con...ctSheet_Lo.pdf

I just thought of something funny, there was a guy on one of the web forums that was buying a JVC HD1/10 to film operations for his class, this camera would be much better. Imagine if he drops it in the patient, "Oops, I'll get that" it might get lost ;)

Joseph Aurili · February 15th, 2006, 09:28 AM

The chart printed gray. I use a good quality color laser printer. I did not even think about the white balance ;) it is on auto. I bet the mix of fluorescent light from the kitchen light and the halogen front light threw off the auto color. I will try again soon.

Joseph Aurili · February 15th, 2006, 09:46 AM

Thanks for the ideas Wayne. I think I will be getting a h264 camera if the quality is better. I would pay $5000 for a camera of this size if I could get top quality results.

Jorge Gil · February 15th, 2006, 10:02 AM

Joseph, in the camcorderinfo's first impressions, the reviewer complained about the possible noise caused by the lens motor and catched by the integrated microphone (as happened with an earlier model from fisher), and argued that was the reason to put an audio in in such a small camera.

Was he right? Can you hear it? Is an external micro needed or the quality of the camera's one is enought?

Thanks for all the time you are spending for us!

Joseph Aurili · February 15th, 2006, 10:07 AM

I can not say I have tested the sound at all, but in the instruction manual is says that motor noice from the optical zoom or focus is normal is not a malfunction. That kind of says it all... ;)

Joseph Aurili · February 15th, 2006, 10:09 AM

I will try to get a sound clip soon also.

Wayne Morellini · February 15th, 2006, 11:48 AM

Yeah, I've been examining the macro blocking and seeing if the soft mode does pixel summing (related to binning) but I can't tell if it's there. I think 25Mb/s H264 would be a better bet for low end work, though Samsung might be doing 19Mb/s camera, and they do vertical models.

I hope somebody starts a switch to h264 resolution TV soon, because the artifacting from mpeg2 compression is literally/distractingly appalling. I would be thinking of getting up and walking out if that was on a theatre screen. At least with all these new DVD players with alternative HD codec support, eventually we won't have to rely on mpeg2 disk standards and will be able to distribute on h264 instead. In China they are developing their own DVD standards, and last I heard they were licensing the latest version of the Ogg Theodore codec, which is better than h264.

Thanks for the clips Joseph, they really give me an idea of what to expect.

February 15th, 2006, 02:32 AM	#141
Jef Bryant Regular Crew Join Date: Jan 2004 Posts: 31	from the manual Interesting but disheartening notes from the manual: "If the exposure setting is "M" (manual) and the ISO setting is set to OTHER THAN ISO-A (auto), the exposure correction setting will be disabled." Well, ain't that a kick in the head. And a special note to PAL users: When the TV output mode is set to PAL and the S-AV cable is connected, "during recording the image appears only on the (on-camera) monitor, it does not appear on the (external) monitor." I guess an external monitor will work in ntsc, though. I suppose this cam is not really intended for use by those who would want good manual control and external monitors anyway...

February 15th, 2006, 07:30 AM	#142
David Kennett Major Player Join Date: Jul 2003 Location: North Ridgeville, Ohio Posts: 407	Jef, I read that "exposure CORRECTION", not manual settings. That is, if you set everything manually, there's no need to TRIM exposure. __________________ Dave

February 15th, 2006, 09:00 AM	#143
David Kennett Major Player Join Date: Jul 2003 Location: North Ridgeville, Ohio Posts: 407	Joseph, I read the chart as about 550 lines H & V. H being just a tad better than V. You could go just a bit tighter, but I doubt that it will improve the numbers much. The total H res is about 1000 lines, so the numbers come in a little below the theoretical maximum. Not bad for a camera of this ilk. Most troubling is the red cast on the chart. Did it print out that way? Now that I look closer, I think the problem is not too much color, but an imbalance in the colorimetry. Are you using auto color? Try an appropriate fixed, or the manual with a white sheet of paper under the light you'll be using. If the chart printed out with an appropriate grey scale, you should see each of the grey squares as grey, and you should be able to see ALL of the squares. Of course you have to get it printed properly first. I will PM you a frame from JVC HD10. It's JPEG, but degraded only a little. __________________ Dave

February 14th, 2006, 04:57 PM	#136
Joseph Aurili Major Player Join Date: Feb 2006 Posts: 326	Resolution chart test clip: www.gamersden.com/hd1test/chart.MP4 Tell me how many lines? I see that distracting adjustment problem again in the first few seconds of video, and this is fixed lighting... One cool thing about the camera is you can control the entire menu system from the remote, so I am able to change the images settings without even touching the camera. And the remote is even easier to control the the joystick on the camera. --- Joseph

February 14th, 2006, 05:12 PM	#137
Joseph Aurili Major Player Join Date: Feb 2006 Posts: 326	OK, I got it done early. Here are the results from the different image settings. I would go with normal as the best. The vivid has too much color. The soft just looks more like SD. The lighting gives a huge improvement in the noise. The camera selected F3.5 at 1/30 of a second. It does not get much better then that ;) With controlled lighting I am actually impressed with the quality. Of course nothing moving in the shot which should make it real easy for the compression. --Normal kitchen lights normal setting-- www.gamersden.com/hd1test/kitchen.JPG www.gamersden.com/hd1test/kitchen.MP4 --Better lighting-- Normal setting: www.gamersden.com/hd1test/normal.JPG www.gamersden.com/hd1test/normal.MP4 Vivid setting: www.gamersden.com/hd1test/vivid.JPG www.gamersden.com/hd1test/vivid.MP4 Soft setting: www.gamersden.com/hd1test/soft.JPG www.gamersden.com/hd1test/soft.MP4 Soft Vivid setting: www.gamersden.com/hd1test/softvivid.JPG www.gamersden.com/hd1test/softvivid.MP4 --- Joseph

February 14th, 2006, 05:51 PM	#139
Joseph Aurili Major Player Join Date: Feb 2006 Posts: 326	Robert, i'm hoping I can use one of the manual modes to eliminate that problem. I hate to make the clips larger then they need to be just to avoid that issue.

February 15th, 2006, 09:06 AM	#144
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	Just looked at some of the kitchen images, I won't bother suggesting a 15 watt light bulb version, the results would be far to predictable. Typical reduced light range, high noise and reduced low light performance of a small mega-pixel CCD. They would have been better sourcing a cmos sensor from fill-factory or smalsensor (now both owned by Cypress semiconductor) or Prixim (which seems too much like smalsensor). They offer multi-slope and autobrite latitude extending techniques that produces latitude ranges upto 20 stops even on small cheap chips, and fill-factory has a technique that samples the light missing the sensor pad, eliminating the screen door technique, with high Signal to Noise ratio and high native latitude. The multi- sampled colours are as I predicted, great, and some people around these forums might actually prefer the noise as a film look. Actually I quiet like it, I think it might turn out to be a funky legend of a consumer camera ;)Only around 15-45 days before we see the h264 versions. About the camera. If you want to increase compression performance let the highlights over expose. I'm examining the videos as I watch House, so I'm in one of those analytical moods again. The amount of noise in the scene will be dragging compression performance down which could be better used on existing detail. Upping the exposure should reduce this. The great anti-aliasing screen door effect isn't helping either. I don't even think it is doing pixel binning (combines pixel signal output from a group of pixels to act as a larger pixel and increase the single strength being sampled, reducing noise and increasing relative latitude. As it is, it looks like they need to use a better interpolation routine to reduce the screen door effect (to cross filter out the anti-aliasing using the separate RGBG pixel channels). It is curious the the sample shots of the baby and ice cream worked so well, but I think it is because it was actually because the light brightness and latitude on the day was good enough, but it almost seems like another camera. Was that guy using a lens attachment? Joseph, there is possibly a way to get a better image from this camera. There has been commercial slr lens adaptors that use a triplet condenser instead of a screen, so you get clean extra bright image with a number of stops more light. Used in conjunction with a suitable ND filter you could possible extend the range and cleanness of the camera. Tiffen had some interesting filter kits, and contrast filters. http://www.optexint.com/sales/filters/filterkits.html http://www.tiffen.com/tiffen_filters.html Worth looking through the pdfs on each type in the kit (I think there are newer kits available to). http://www.tiffen.com/tiffen_product_literature.html http://www.tiffen.com/userimages/filter-lens.pdf This contrast filter thing it interesting, appears counter intuitive, I am interested in how the technology behind it works. Because, despite their glowing description, I don't really how it is much of an improvement over a ND in most cases, anyone know? http://www.tiffen.com/userimages/Con...ctSheet_Lo.pdf I just thought of something funny, there was a guy on one of the web forums that was buying a JVC HD1/10 to film operations for his class, this camera would be much better. Imagine if he drops it in the patient, "Oops, I'll get that" it might get lost ;)

February 15th, 2006, 09:28 AM	#145
Joseph Aurili Major Player Join Date: Feb 2006 Posts: 326	The chart printed gray. I use a good quality color laser printer. I did not even think about the white balance ;) it is on auto. I bet the mix of fluorescent light from the kitchen light and the halogen front light threw off the auto color. I will try again soon.

February 15th, 2006, 09:46 AM	#146
Joseph Aurili Major Player Join Date: Feb 2006 Posts: 326	Thanks for the ideas Wayne. I think I will be getting a h264 camera if the quality is better. I would pay $5000 for a camera of this size if I could get top quality results.

February 15th, 2006, 10:02 AM	#147
Jorge Gil New Boot Join Date: Feb 2006 Posts: 6	Joseph, in the camcorderinfo's first impressions, the reviewer complained about the possible noise caused by the lens motor and catched by the integrated microphone (as happened with an earlier model from fisher), and argued that was the reason to put an audio in in such a small camera. Was he right? Can you hear it? Is an external micro needed or the quality of the camera's one is enought? Thanks for all the time you are spending for us!

February 15th, 2006, 10:07 AM	#148
Joseph Aurili Major Player Join Date: Feb 2006 Posts: 326	I can not say I have tested the sound at all, but in the instruction manual is says that motor noice from the optical zoom or focus is normal is not a malfunction. That kind of says it all... ;)

February 15th, 2006, 10:09 AM	#149
Joseph Aurili Major Player Join Date: Feb 2006 Posts: 326	I will try to get a sound clip soon also.

February 15th, 2006, 11:48 AM	#150
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	Yeah, I've been examining the macro blocking and seeing if the soft mode does pixel summing (related to binning) but I can't tell if it's there. I think 25Mb/s H264 would be a better bet for low end work, though Samsung might be doing 19Mb/s camera, and they do vertical models. I hope somebody starts a switch to h264 resolution TV soon, because the artifacting from mpeg2 compression is literally/distractingly appalling. I would be thinking of getting up and walking out if that was on a theatre screen. At least with all these new DVD players with alternative HD codec support, eventually we won't have to rely on mpeg2 disk standards and will be able to distribute on h264 instead. In China they are developing their own DVD standards, and last I heard they were licensing the latest version of the Ogg Theodore codec, which is better than h264. Thanks for the clips Joseph, they really give me an idea of what to expect.