I have been using VS10 to capture/edit/output hi-def video from Sony HDR-HC3 and back to tape in same camcorder. Somewhere during the process, I am losing video quality. The final output to tape, when played back on hi-def Sony CRT TV, is not nearly as good as the original taped video. I am capturing with the following settings:
NTSC drop frame (29.97 fps)
MPEG files
24 bits, 1440 x 1080, 29.97 fps
Upper Field First
(MPEG-2), 16:9
Video data rate: 25000 kbps
Audio data rate: 384 kbps
MPEG audio layer 2, 48 KHz, Stereo
Have tried both deinterlace during capture and not
I am capturing via Firewire, and playing back to TV via HDMI output
I use the same settings to render and output back to tape in HDR-HC3
The final output is better quality than SD, but just not as crisp as the original uncaptured and unedited raw HD tape.
Any suggestions, or is VS10 just not capable of retaining original quality?
Thanks,
Chuck
Quality loss after edit and output high-def video to tape
Moderator: Ken Berry
-
jchunter
Chuck,
Please read the High Definition Tutorial at http://phpbb.ulead.com.tw/EN/viewtopic.php?t=13872 - follow this procedure.
Your capture properties are perfect. Don't try to deinterlace or change any other properties while capturing.
Judging quality: What is your procedure for comparing quality? When I first got my HDR-HC1, I printed out a large image of the ISO 12233 resolution chart and shot some video, which let me compare resolution on my HDTV (thru the camcorder's component video interface) with the final edited video on HDTV thru my AVLP2 DVD player. There was no visible reduction in picture quality. You can also use Virtual Dub Mpeg to accurately measure the resolution of a single frame on a PC monitor.
When playing back in the PC, a lot of software is not able to handle HD - especially in a slower PC. Media Player Classic seems to work the best right now.
Please read the High Definition Tutorial at http://phpbb.ulead.com.tw/EN/viewtopic.php?t=13872 - follow this procedure.
Your capture properties are perfect. Don't try to deinterlace or change any other properties while capturing.
Judging quality: What is your procedure for comparing quality? When I first got my HDR-HC1, I printed out a large image of the ISO 12233 resolution chart and shot some video, which let me compare resolution on my HDTV (thru the camcorder's component video interface) with the final edited video on HDTV thru my AVLP2 DVD player. There was no visible reduction in picture quality. You can also use Virtual Dub Mpeg to accurately measure the resolution of a single frame on a PC monitor.
When playing back in the PC, a lot of software is not able to handle HD - especially in a slower PC. Media Player Classic seems to work the best right now.
-
maddrummer3301
- Posts: 2507
- Joined: Fri Dec 10, 2004 10:24 pm
- Location: US
-
jchunter
MD,
You are correct that there is no way to change capture properties for HDV in VS10+ - as I stated in the HD Tutorial. I did not go into this detail with the OP because I assumed that either he was setting Project Properties or was skipping ahead to the Create Video stage. BTW, I could always set capture properties for DV camcorders - I wonder if you can with the HC3.
AhHa Moment!
Chuck, is it possible that you are inadvertantly capturing from the HC3 in standard definition? because I assume that you would have a choice of properties in this mode. You might be then receiving standard quality video and upconverting to 1440x1080, which would not restore the quality. This would explain the quality loss that you are seeing.
You are correct that there is no way to change capture properties for HDV in VS10+ - as I stated in the HD Tutorial. I did not go into this detail with the OP because I assumed that either he was setting Project Properties or was skipping ahead to the Create Video stage. BTW, I could always set capture properties for DV camcorders - I wonder if you can with the HC3.
AhHa Moment!
Chuck, is it possible that you are inadvertantly capturing from the HC3 in standard definition? because I assume that you would have a choice of properties in this mode. You might be then receiving standard quality video and upconverting to 1440x1080, which would not restore the quality. This would explain the quality loss that you are seeing.
-
maddrummer3301
- Posts: 2507
- Joined: Fri Dec 10, 2004 10:24 pm
- Location: US
-
csavall
Again,
Thanks for all the help on this problem. Here are some more things I have learned and tried to improve quality.
I used the Hi-Def tutorial referenced above to do all my original capturing, it appears I did not make all the correct settings prior to creating the video files after editing. I have now changed to following two settings in: Create Video File / Options / Compression Tab - Video data rate from Constant to Variable / 25000 kbps max, and de-delected I-frames only. Can anyone tell me what I-frames only changes, I can find no reference to it in the VS10 help files?
The above two file creation setting changes resulted in a dramatic improvement in output quality back to the camcorder tape and viewed on a Sony Hi-Def CRT TV. As good or nearly as good as original tape.
jchunter asked me what I was using to compare quality of original vs. edited output, quite low tech actually, just using the old eyeball from about 4 feet away from TV, but the quality differences is readily apparent even with this method, which is ultimately the best final judgment, what you actually see on the TV.
As far as deinterelace capture. I have tried both ways, even with my new file save options. There does seem to be some difference, but depends on how I feed the HDR-HC3 output to the Sony TV. I have tried both the HDMI cable, and a Firewire feed. When playing back the original tape, the quality is same whichever feed I use. However, when palying back captured and edited files, rendered and then uploaded to the camera tape, the files captured with "Image capture deinterlace" de-selected lose quite a bit of quality when played back via the Firewire feed, but not the HDMI feed. Now, this could be the interaction of the camera and the TV, but for now, I will capture with "Image capture deinterlace" set to the default of ON. The only reason I even tried to change this setting in the first place is that my final output will be to a CRT tube TV, not an LCD or Plasma display, so I thought changing the deinterlace setting might improve the TV picture.
Finally, VS10 does seem to allow changing most of the image capture settings options for the HDR-HC3 via the Project Properties / Edit menu. I do not see where thay are "locked." However, there is no help file that I can find that addresses the options in the "Video Save Options / Compression Tab" menu box. The help file only addresses the AVI tab, which is not present when editing Hi-Def video, only SD video.
Any more input would be great.
Thanks again, Chuck
Thanks for all the help on this problem. Here are some more things I have learned and tried to improve quality.
I used the Hi-Def tutorial referenced above to do all my original capturing, it appears I did not make all the correct settings prior to creating the video files after editing. I have now changed to following two settings in: Create Video File / Options / Compression Tab - Video data rate from Constant to Variable / 25000 kbps max, and de-delected I-frames only. Can anyone tell me what I-frames only changes, I can find no reference to it in the VS10 help files?
The above two file creation setting changes resulted in a dramatic improvement in output quality back to the camcorder tape and viewed on a Sony Hi-Def CRT TV. As good or nearly as good as original tape.
jchunter asked me what I was using to compare quality of original vs. edited output, quite low tech actually, just using the old eyeball from about 4 feet away from TV, but the quality differences is readily apparent even with this method, which is ultimately the best final judgment, what you actually see on the TV.
As far as deinterelace capture. I have tried both ways, even with my new file save options. There does seem to be some difference, but depends on how I feed the HDR-HC3 output to the Sony TV. I have tried both the HDMI cable, and a Firewire feed. When playing back the original tape, the quality is same whichever feed I use. However, when palying back captured and edited files, rendered and then uploaded to the camera tape, the files captured with "Image capture deinterlace" de-selected lose quite a bit of quality when played back via the Firewire feed, but not the HDMI feed. Now, this could be the interaction of the camera and the TV, but for now, I will capture with "Image capture deinterlace" set to the default of ON. The only reason I even tried to change this setting in the first place is that my final output will be to a CRT tube TV, not an LCD or Plasma display, so I thought changing the deinterlace setting might improve the TV picture.
Finally, VS10 does seem to allow changing most of the image capture settings options for the HDR-HC3 via the Project Properties / Edit menu. I do not see where thay are "locked." However, there is no help file that I can find that addresses the options in the "Video Save Options / Compression Tab" menu box. The help file only addresses the AVI tab, which is not present when editing Hi-Def video, only SD video.
Any more input would be great.
Thanks again, Chuck
-
jchunter
Chuck,
Don't use "I-Frames-Only" because the video file will inflate to about 5 times its compressed size and Video Studio will probably run off the rails trying to do that with HDV.
In Mpeg encoding, an I-Frame is a complete reference frame, with all its pixels just spatially compressed. There is one of these per Group of Pictures. The other frames are temporally encoded, using only differences from the I-frame. Thus, a still image video is encoded with a single I-Frame and the 14 other frames in that GOP are nearly empty (because they are identical to the I-Frame).
You can't control the CAPTURE properties using the Project Property settings. They only govern playback during editing. After you capture, Video Studio should ask if you want Project Properties set to match those of the captured video file. Answer 'Yes" !
The HD Tutorial shows you exactly how to set properties in the Create Video File stage.
Don't use "I-Frames-Only" because the video file will inflate to about 5 times its compressed size and Video Studio will probably run off the rails trying to do that with HDV.
In Mpeg encoding, an I-Frame is a complete reference frame, with all its pixels just spatially compressed. There is one of these per Group of Pictures. The other frames are temporally encoded, using only differences from the I-frame. Thus, a still image video is encoded with a single I-Frame and the 14 other frames in that GOP are nearly empty (because they are identical to the I-Frame).
You can't control the CAPTURE properties using the Project Property settings. They only govern playback during editing. After you capture, Video Studio should ask if you want Project Properties set to match those of the captured video file. Answer 'Yes" !
The HD Tutorial shows you exactly how to set properties in the Create Video File stage.
-
maddrummer3301
- Posts: 2507
- Joined: Fri Dec 10, 2004 10:24 pm
- Location: US
-
csavall
Guys,
Lots of great help there. Thanks a ton.
I am currently re-rendering several hours of hi-def, but the test samples look quite nice. The big difference for me was not using I-frames, and possibly selecting VBR. The only reason I used I-frames in the 1st place is that it was the default setting, and I didn't know whether to change it or not, since I could find no VS10 help file on it.
Excuse my ignorance, but I thought I had an idea what interlaced video was. Since my final output is to a CRT TV tube which is 1080i hi-def, I assumed that I would want to output files that are interlaced video. I would only want a "de-interlaced" video if I was outputting to a progressive scan display, such as a "p" tv, LCD or plasma. Therefore, I thought I should try to capture in interlaced mode, but the tutorial shows to set preferences to "Image capture deinterlace" Now this is where my confusion may be. I don't know it that affects the capture settings or the output setting or both. I may also have it backwards. To me, selecting deinterlace would mean that the program would un-interlace an interlaced image and convert it to frame based or progressive format. Since that is not what I want, it seems that I should de-select the default setting of "Image capture deinterlace", but apparently I may have that backwards. So for now, I still have "Image capture deinterlace" selected, and things are working fine.
I am attempting not to change ANY formats or settings from what I am capturing in, as I want to output and playback in exactly the same format. Now, down the road, I may want to convert some files to display on internet, and hopefully I don't have to re-capture everything, just change the output settings. I have tried Quicktime Pro before, but have had no luck converting to H.264 with it, just won't run, and they won't give you your money back.
Thanks again,
Chuck
Lots of great help there. Thanks a ton.
I am currently re-rendering several hours of hi-def, but the test samples look quite nice. The big difference for me was not using I-frames, and possibly selecting VBR. The only reason I used I-frames in the 1st place is that it was the default setting, and I didn't know whether to change it or not, since I could find no VS10 help file on it.
Excuse my ignorance, but I thought I had an idea what interlaced video was. Since my final output is to a CRT TV tube which is 1080i hi-def, I assumed that I would want to output files that are interlaced video. I would only want a "de-interlaced" video if I was outputting to a progressive scan display, such as a "p" tv, LCD or plasma. Therefore, I thought I should try to capture in interlaced mode, but the tutorial shows to set preferences to "Image capture deinterlace" Now this is where my confusion may be. I don't know it that affects the capture settings or the output setting or both. I may also have it backwards. To me, selecting deinterlace would mean that the program would un-interlace an interlaced image and convert it to frame based or progressive format. Since that is not what I want, it seems that I should de-select the default setting of "Image capture deinterlace", but apparently I may have that backwards. So for now, I still have "Image capture deinterlace" selected, and things are working fine.
I am attempting not to change ANY formats or settings from what I am capturing in, as I want to output and playback in exactly the same format. Now, down the road, I may want to convert some files to display on internet, and hopefully I don't have to re-capture everything, just change the output settings. I have tried Quicktime Pro before, but have had no luck converting to H.264 with it, just won't run, and they won't give you your money back.
Thanks again,
Chuck
-
maddrummer3301
- Posts: 2507
- Joined: Fri Dec 10, 2004 10:24 pm
- Location: US
-
maddrummer3301
- Posts: 2507
- Joined: Fri Dec 10, 2004 10:24 pm
- Location: US
-
jchunter
MD,
Those are very interesting results - you have shown that there is a definite negative interaction between two Ulead products: VS10 and MF. This needs to be somehow communicated directly to the Ulead development team because the interaction could extend to other Ulead products.
However, what do mean about MF capturing HDV "better" ? Capturing HDV is just a copying process, which can't affect picture quality...
Those are very interesting results - you have shown that there is a definite negative interaction between two Ulead products: VS10 and MF. This needs to be somehow communicated directly to the Ulead development team because the interaction could extend to other Ulead products.
However, what do mean about MF capturing HDV "better" ? Capturing HDV is just a copying process, which can't affect picture quality...
-
maddrummer3301
- Posts: 2507
- Joined: Fri Dec 10, 2004 10:24 pm
- Location: US
-
jchunter
I have a different take on the camera movement issue. There is no problem when people are moving in front of a stationary hi def camcorder. However, if the camera is rocking (e.g., on a boat), the whole video frame rocks on playback and the greater sharpness makes people dizzy because is seems more real. This perspective leads me to believe that it is a hi def. issue and has nothing to do with mpeg encoding.maddrummer3301 wrote:I just shot a birthday party and used my SD Canon cam. Reason for using SD was going to do lots of movement and shooting indoors closeup footage.
I started using a tripod when I got the HC1 but, in retrospect, I think it was as much due to taking my video techniques more seriously as it was due to high def being more picky about motion. When I look at some of my older pans and zooms, some were way too fast.