Video losing quality

[DB83]

Confused.

If you can save BMP direct from a monitor why are you bothering with any consumer capture device ?. Really, you can try 100's of these but quality is hardly going to improve.

[jagabo]

I've confirmed with the colposcope manufacture today again, and can confirm
1. Camera >> S-Video >> TV monitor (or other monitor) >> Quality is good, and can save to BMP file

When you say you "can confirm" this -- did you actually hook up a TV or monitor to the s-video port and see it for yourself? Or are you just taking the manufacturer's word for it? And if you can save a BMP image this way why are you bothering with the EasyCap and a computer?

Did you try any of the earlier suggestions? Adjust the video proc amp settings? Change the deinterlace method? Display the digital capture pixel-for-pixel rather than scaled to some arbitrary dimensions?

[PuzZLeR]

Hi,

As indicated above, I've contacted Matrox and Hauppauge, and got a suggestion of 950q (see link below). Not sure if anyone used this before?

http://www.hauppauge.com/site/products/data_hvr950q.html

No, but I wonder why Hauppauge didn't suggest this: http://www.hauppauge.com/site/products/data_usblive2.html The USB Live 2 is a decent standard definition USB analog capture device with no TV tuner. While the WinTV-HVR-950Q does have the ability to capture standard definition analog video and stereo audio, it also includes an analog/digital combo TV tuner, which adds to the cost but isn't going to be useful to you in your work.

Hi, figured I'd chime in. I have a 950Q (along with my dozens of other capture devices I just can't seem to stop buying...).


Please note that it does not support YUY2 (YUV 4:2:2) video (or maybe I haven't been able to make it work with this), hence I would not recommend it for many types of capture. It's a tuner primarily, and its complete capabilities do not superset those of the USB-Live2 when it comes to analog capture.

[nzcncf]

Good to know that you have used 950q and the feedback is valuable. In this case, I may try USB-Live2 to replace EasyCap, and see how it goes.

Thanks you all for the help and advice!

[DB83]

Good to know that you have used 950q and the feedback is valuable. In this case, I may try USB-Live2 to replace EasyCap, and see how it goes.

Thanks you all for the help and advice!

So how about responding to both mine and jagabo's comments about the use of a monitor and the method of 'comparing' ?

And I'll say this again, and until I am blue in the face, ALL analogue capture devices record to a maximum pixel size of 720*480 (720*576 for PAL). If your picture is larger then you might simply be creating these issues with the magnification. Another reason to see an actual sample.

[nzcncf]

1. In actual fact, we only try to captuer image from the popup window, not the full screen with the resolution of 720x480.
2. Image on the ULead windows is the same as our software, and clinicians aren't happy with the quality at the stage, and same with the saved image.
3. When S-Video directly connects to TV monitor, the quality is ok
4. Unfortunately real sample can not be uploaded due to privacy

Therefore, I guess the issue is due to EasyCap converter, and we would like to try another converter, and see if any difference. I will report the outcome when I receive the new converter (950q). Thanks

[DB83]

I can appreciate the matter of privacy but, without naming names, one female anatomical part will look like another.

Now I do not understand what you mean by 'popup window'. Do you mean the preview window on the Ulead Software or your own software ?

If it is your own then noone can really help unless that has been badly written and it just captures a very small image which can never look good at a reasonable size for clinical analysis.

A suggestion. Allow the camera to film the palm of your hand at the sort of distance where it would do its normal work. Then upload the capture of that from the popup window/your software NOT the Ulead Screen image.

But if the image you see on the tv monitor is ok can not you capture just that ?. There is software available to grab a screen image. There would be an additional process to include that image in a database but it seems to me that your software is trying to do both. Maybe it was written with high quality imaging in mind.

[Brad]

I can appreciate the matter of privacy but, without naming names, one female anatomical part will look like another.

Legal considerations...

[DB83]

I can appreciate the matter of privacy but, without naming names, one female anatomical part will look like another.

Legal considerations...

The OP said 'privacy' which I suppose can encompass legal issues. But I have suggested a compromise

[jagabo]

Just wrap your fist around the end of the device and send a video of your hand [b]moving a little[b], along with a screen cap the way you are capturing now. The problem probably isn't the EasyCap, but the software's deinterlacing, or the capture device driver's proc amp not being set sharp enough. But jesus, trying another capture device will only cost you $50 to $100 dollars so you might as well try it.