High Definition with Elphel model 333 camera

Forrest Schultz · April 18th, 2006, 09:14 PM

I also think at 1600 by 896, i am really maxing it out. it seems on certain footage, it will do 24fps fine, but when i have footage in motion and pans, i lose frames. i think i need to scale down the resolution. 1600 by 896 is not really needed anyways. could this also cause the excess electronic rolling shutter artifact? the fact that i am trying to grab more frames than possible at that resolution. thanks

Andrey Filippov · April 18th, 2006, 10:58 PM

Quote:

Originally Posted by Forrest Schultz

I also think at 1600 by 896, i am really maxing it out. it seems on certain footage, it will do 24fps fine, but when i have footage in motion and pans, i lose frames.

Forrest,

I think it will be easier to locate problems with frame timestamping. Only what you can try right now - set low quality and see if it helps (eliminates frame drop). You see, the FPGA compressor can now process all the data from the sensor at any quality settings without drops - this frame rate is calculated and reported in web interface (of cause there might be bugs and with some settings the calculation of the frame rate can be wrong). But the sending of the compressed data out requires CPU activity and higher bandwidth needs more CPU power, so it can limit the frame rate. If reduced quality helps - that is probably the case.

Did you use autoexposure? It also uses CPU and can steal some of it from the streamer - especially if the illumination of the scene really changes (when you move the camera).

Forrest Schultz · April 18th, 2006, 11:27 PM

I think the main problem is my CPU speed. im only at 1.5 Ghz right now. I will buy a 2.8 Ghz processer soon.,and that should help alot. because its when i go for better quality compression that it usually drops. thanks Andrey

Andrey Filippov · April 19th, 2006, 02:50 AM

Quote:

Originally Posted by Forrest Schultz

I think the main problem is my CPU speed. im only at 1.5 Ghz right now. I will buy a 2.8 Ghz processer soon.,and that should help alot. because its when i go for better quality compression that it usually drops. thanks Andrey

No, Forrest, I meant CPU in the camera (with "Axis" written on it- http://wiki.elphel.com/index.php?tit...d_overside.png ), not in your PC. Just recording is not so demanding application on the PC side, so frame drop is probably in the camera in that case.

Forrest Schultz · April 19th, 2006, 09:27 AM

Oh ok. I see now. thanks andrey

Forrest Schultz · April 21st, 2006, 01:05 AM

Hi guys. i was looking at my footage compared to the elphel cameras Andrey has set up.

camera4.elphel.com

and such

and i am noticing a big differece in resolution and im pretty sure its the lens im using. it dont think it was the best quality made c-mount lens. it doesnt even have a brand name written on it.

so i ordered 2 new lenses. One is a 12mm prime c-mount lens with f1.4 and the other is a computar f1.3 5-50mm cs mount zoom lens. The computar brand new costs 135. so i think that might do the trick. as for the other c-mount (the f1.4 one i bought) eh, i dont know if it will be good either, but the Computar i am hoping will get er' done.

both lenses are one their way, so when i recieve them. i will test with them.
PS, i have already tested with a 35mm adapter, and the results are good. great DOF. but i need to have a better lens because the quality of the glass really matters in this case. also, the ground glass i used for the test was just a clear plastic diffused on one side. it was a stationary peice so you can see all the stracthes and such. When i get all the new lenses and a new ground glass. I will post tests with that rigged up. And after that, its off to build an optical viefinder for the 35mm adapter. more on that later.

Wayne Morellini · April 21st, 2006, 09:42 AM

On Obin's 4:4:4 thread he had the same problem. They eventually tracked down the best c mount lens, but found lens from good film cameras very good (what format would that be, 8MM?). Have a read there, it should be within the first 1-10% of the thread.

Régine Weinberg · April 21st, 2006, 01:41 PM

a silly question
on the axis page
there is nothing to be found about their CPU's
maybe I'm stupid, seems so

Forrest Schultz · April 21st, 2006, 03:30 PM

I did a test today , so that when i get the new lenses, i will be able to compare the focus ablilities >The chart was hung in the shade out of direct sunlight.

http://i11.photobucket.com/albums/a1...colortest1.jpg

please note the colors on that chart are not your standard R G B. the blue is a deeper blue and the Green is a deeper green also. That is how they are on the paper.
when i get a chance, ill print out colorcorrect r g b and shoot that.

here is the same thing with the exposure set longer.

http://i11.photobucket.com/albums/a1.../colotest2.jpg

John Wyatt · April 22nd, 2006, 05:18 AM

Forrest -- a while back I started a thread called "Call for C-mount lens info" here in Alt Imaging, there might be some info there for you...
John.

Régine Weinberg · April 22nd, 2006, 06:51 AM

Dear Wayne
16mm lenses are there with a real very good quality
adapter C-mount Pl are there also.

I used on an Aaton S16 Nikon lenses,
the wide angel was the best for money
we could get.

C-mount is industrie, for watching chicken farms or so quite ok LOL

Wayne Morellini · April 22nd, 2006, 07:19 AM

I think that was another of Obin's thread.

Ronald, 16MM lens, but aren't these sensors are far smaller than 16mm?

Be prepared, I have found something interesting, and if I get it home, I may post on it in future.

Wayne Morellini · April 22nd, 2006, 07:21 AM

I found nothing but trouble trying to run camera4.elphel.com . I get one photo with box and one web page that says loading.

Noah Yuan-Vogel · April 22nd, 2006, 09:01 AM

I'm impressed with the color and ease of use of this camera, and I know it can create sharp images using a nicer lens since i am using the same sensor. what kind of IR filter is being used? is it all built in? because ive had some problems getting good color out of my camera. Otherwise what im really interested in is the quality of the mjpeg compression, what is the least compression it can apply? all the way up to the ~7MBps that 100BT can handle? that was a little unclear. Another really important thing is how is the framerate control? accurate? you really need an accurate and consistent framerate for sound sync. Also does it give you control over vertical blanking to keep rolling shutter artifacts under control?

BTW this is the same sensor obin was using and decided to scrap due to rolling shutter or something. Also, these sensors arent that much smaller than 16mm, its 1/2" and 16mm i think is somewhere between 2/3 and 1". Although when running at 1600x896 only about a 1/3" area is being used.

Also Forrest, I noticed you mentioned rolling shutter artifacts at your high resolutions. rolling shutter is definitely a big problem with this sensor, ive even run it at 720x540 with 1200+ vertical blank and been able to notice significant rolling shutter artifacts in fast horizontal pans. Makes me kind of worry this sensor will never be appropriate for filmmaking. Now a 96MHz sensor sounds more like it, but if its 5mp its likely to have to be run at at a <2mp windowed resolution only using like 1/4" in sensor area if we want the minimal rolling shutter artifact and framerate we want...

Forrest Schultz · April 22nd, 2006, 03:02 PM

I agree with everything you said Noah, thats also nice to know i can get sharper picture with the better lens attached.

at 1600x896, you see more of the sensor, therefore rolling shutter is more evident. but that resolution isnt needed, plus it isnt too stable. 1440 by 800 is much more stable, and even 1280 by 720. I will do tests with all these when i get the new lens. I havent seen frames drop except in 1600 by 896 mode. I dont know if there is control over vertical blanknig. i dont know how that works. Thank you for your help.

April 21st, 2006, 01:41 PM	#83
Régine Weinberg Major Player Join Date: Jan 2004 Location: Bordeaux, going to Bangkok, 2011 Posts: 232	have a look a silly question on the axis page there is nothing to be found about their CPU's maybe I'm stupid, seems so Last edited by Régine Weinberg; April 22nd, 2006 at 07:01 AM.

April 22nd, 2006, 06:51 AM	#86
Régine Weinberg Major Player Join Date: Jan 2004 Location: Bordeaux, going to Bangkok, 2011 Posts: 232	Lenses Dear Wayne 16mm lenses are there with a real very good quality adapter C-mount Pl are there also. I used on an Aaton S16 Nikon lenses, the wide angel was the best for money we could get. C-mount is industrie, for watching chicken farms or so quite ok LOL

April 18th, 2006, 09:14 PM	#76
Forrest Schultz Major Player Join Date: Mar 2005 Location: Atwater, CA Posts: 246	I also think at 1600 by 896, i am really maxing it out. it seems on certain footage, it will do 24fps fine, but when i have footage in motion and pans, i lose frames. i think i need to scale down the resolution. 1600 by 896 is not really needed anyways. could this also cause the excess electronic rolling shutter artifact? the fact that i am trying to grab more frames than possible at that resolution. thanks

April 18th, 2006, 11:27 PM	#78
Forrest Schultz Major Player Join Date: Mar 2005 Location: Atwater, CA Posts: 246	I think the main problem is my CPU speed. im only at 1.5 Ghz right now. I will buy a 2.8 Ghz processer soon.,and that should help alot. because its when i go for better quality compression that it usually drops. thanks Andrey

April 19th, 2006, 09:27 AM	#80
Forrest Schultz Major Player Join Date: Mar 2005 Location: Atwater, CA Posts: 246	Oh ok. I see now. thanks andrey

April 21st, 2006, 01:05 AM	#81
Forrest Schultz Major Player Join Date: Mar 2005 Location: Atwater, CA Posts: 246	Hi guys. i was looking at my footage compared to the elphel cameras Andrey has set up. camera4.elphel.com and such and i am noticing a big differece in resolution and im pretty sure its the lens im using. it dont think it was the best quality made c-mount lens. it doesnt even have a brand name written on it. so i ordered 2 new lenses. One is a 12mm prime c-mount lens with f1.4 and the other is a computar f1.3 5-50mm cs mount zoom lens. The computar brand new costs 135. so i think that might do the trick. as for the other c-mount (the f1.4 one i bought) eh, i dont know if it will be good either, but the Computar i am hoping will get er' done. both lenses are one their way, so when i recieve them. i will test with them. PS, i have already tested with a 35mm adapter, and the results are good. great DOF. but i need to have a better lens because the quality of the glass really matters in this case. also, the ground glass i used for the test was just a clear plastic diffused on one side. it was a stationary peice so you can see all the stracthes and such. When i get all the new lenses and a new ground glass. I will post tests with that rigged up. And after that, its off to build an optical viefinder for the 35mm adapter. more on that later.

April 21st, 2006, 09:42 AM	#82
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	On Obin's 4:4:4 thread he had the same problem. They eventually tracked down the best c mount lens, but found lens from good film cameras very good (what format would that be, 8MM?). Have a read there, it should be within the first 1-10% of the thread.

April 21st, 2006, 03:30 PM	#84
Forrest Schultz Major Player Join Date: Mar 2005 Location: Atwater, CA Posts: 246	I did a test today , so that when i get the new lenses, i will be able to compare the focus ablilities >The chart was hung in the shade out of direct sunlight. http://i11.photobucket.com/albums/a1...colortest1.jpg please note the colors on that chart are not your standard R G B. the blue is a deeper blue and the Green is a deeper green also. That is how they are on the paper. when i get a chance, ill print out colorcorrect r g b and shoot that. here is the same thing with the exposure set longer. http://i11.photobucket.com/albums/a1.../colotest2.jpg

April 22nd, 2006, 05:18 AM	#85
John Wyatt Regular Crew Join Date: Jan 2006 Location: West Country, UK Posts: 141	Forrest -- a while back I started a thread called "Call for C-mount lens info" here in Alt Imaging, there might be some info there for you... John.

April 22nd, 2006, 07:19 AM	#87
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	I think that was another of Obin's thread. Ronald, 16MM lens, but aren't these sensors are far smaller than 16mm? Be prepared, I have found something interesting, and if I get it home, I may post on it in future.

April 22nd, 2006, 07:21 AM	#88
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	I found nothing but trouble trying to run camera4.elphel.com . I get one photo with box and one web page that says loading.

April 22nd, 2006, 09:01 AM	#89
Noah Yuan-Vogel Major Player Join Date: Dec 2004 Location: New York City Posts: 613	I'm impressed with the color and ease of use of this camera, and I know it can create sharp images using a nicer lens since i am using the same sensor. what kind of IR filter is being used? is it all built in? because ive had some problems getting good color out of my camera. Otherwise what im really interested in is the quality of the mjpeg compression, what is the least compression it can apply? all the way up to the ~7MBps that 100BT can handle? that was a little unclear. Another really important thing is how is the framerate control? accurate? you really need an accurate and consistent framerate for sound sync. Also does it give you control over vertical blanking to keep rolling shutter artifacts under control? BTW this is the same sensor obin was using and decided to scrap due to rolling shutter or something. Also, these sensors arent that much smaller than 16mm, its 1/2" and 16mm i think is somewhere between 2/3 and 1". Although when running at 1600x896 only about a 1/3" area is being used. Also Forrest, I noticed you mentioned rolling shutter artifacts at your high resolutions. rolling shutter is definitely a big problem with this sensor, ive even run it at 720x540 with 1200+ vertical blank and been able to notice significant rolling shutter artifacts in fast horizontal pans. Makes me kind of worry this sensor will never be appropriate for filmmaking. Now a 96MHz sensor sounds more like it, but if its 5mp its likely to have to be run at at a <2mp windowed resolution only using like 1/4" in sensor area if we want the minimal rolling shutter artifact and framerate we want...

April 22nd, 2006, 03:02 PM	#90
Forrest Schultz Major Player Join Date: Mar 2005 Location: Atwater, CA Posts: 246	I agree with everything you said Noah, thats also nice to know i can get sharper picture with the better lens attached. at 1600x896, you see more of the sensor, therefore rolling shutter is more evident. but that resolution isnt needed, plus it isnt too stable. 1440 by 800 is much more stable, and even 1280 by 720. I will do tests with all these when i get the new lens. I havent seen frames drop except in 1600 by 896 mode. I dont know if there is control over vertical blanknig. i dont know how that works. Thank you for your help.