I'm not sure whether I should be posting this to the Audio Forum or the FCP forum and so please let me know if I should post it to the FCP one instead:
I am having a problem with audio sync lagging/drifting when I capture from my Canon HV40 tapes into FCP ProRes 422. Don't know if this is a user setting problem or something intrinsic to my hardware-perhaps my computer simply is too underpowered.
I've been seeking answers for this problem for nearly a month in other forums and the best I've come up with is here:
http://discussions.apple.com/thread.jspa?threadID=2105274&stqc=true
So I'm wondering whether anyone in this community can suggest anything better.
To reiterate (as per the above Apple forum link):
I am shooting on a HV40 shooting in its so-called native 24p mode, and so I presume there are no pulldown issues) and editing in FCP 6.06 (FCS v2) on a PowerMac G5 dual Core 2GHz machine with 8.5 Gb of ram, OS 10.4.11.
The specific capture preset I'm using is Apple ProRes 422 1440 x1080 24p 48kHz as I believe these setting match the specs of my camera: MPEG-1 Audio Layer 2 , 16bit. 48kHz, transfer rate 384 kbps(2 channels).
I capture from my camera (not a deck) which writes HDV files to tape during videography. If I capture in HDV into FCP with a HDV 1080 24P easy setup, i.e. without transcoding to ProRes, there is no sync problem. However because I intend to do a lot of colour grading, I have switched from using an HDV easy setup to Apple ProRes 422, but that is when I find that my audio is drifting/lagging widely over the course of the capture.
I observe that when I capture in HDV I can access the full log and capture interface and retain some FW device control of my camera through the FCP interface and that as the clips come are captured, they appear one by one in my browser. Whereas when I capture into ProRes, even though I begin the process by using the log and capture command, there is no interface presented to me and capturing begins immediately from wherever the tape happens to be sitting on the tape head of the camera and continues until the end of the tape. That is, I forfeit device control. Moreover (and I suspect my problem has a lot to do with the following observation) the clips don't appear in my browser until the tape stops rolling and the process is otherwise completed. Apropos this, it may be the case that the software doesn't break out the clips and reset the sync until the whole capture is done, and that it is in this phase of the process the lag is introduced, proportionally greater the longer the footage and the closer one gets to the final clip in the capture. If that is the case, it would perhaps be an FCP bug.
Over the course of a one hour tape, audio falls roughly 5 seconds behind.
BTW, (a) I've tried experimenting with various device protocols in the Device Control Preset Editor tab under Audio/Video Settings and cannot find a combination that regains me device control, and (b) I've also raked carefully through the various menus and sub menus to see if there is any customization I can apply to the presets (capture, sequence. a/v devices or otherwise) that might make more sense in terms of syncing audio to video, but nothing makes any better sense than the ones I'm already using.
If I were just capturing a short clip and could conveniently select clips in my camera LCD screen and navigate with my on-camera control to the in-point this sync delay probably wouldn't be a problem, but as this is a documentary with mountains of footage, such a workflow is unworkable. I need to be able capture (and preferably, log and capture) whole tapes and be able to use the more ergonomic environment of a computer to make my rough selections of clips. Captured in that way audio drift/lag becomes evident after a few minutes and as I say, extreme by the end of the tape... so extreme that if it is a short clip of a few seconds, most of the audio that should be synced with one video clip may show up in the next video clip.
As an established photographer new to video, if I understand correctly the following analogy is accurate: shooting HDV is like shooting JPGs both of which create compression artifacts on their respective initial saves to the on-board media but however whereas from a compact flash card one can copy a jpg still to a harddrive during the ingest process with no further re-compression artifacts, in comparison with HDV one has to effectively open up the file (i.e. play the video in camera out into the computer) and then re-save the HDV to the hard drive during the capture process incurring a second generation of re-compression artifacts. Once on the harddrive in Photoshop a still JPG can be worked on and saved as a tif or PSD whereas in FCP working in HDV it would save again as HDV generating yet subsequent re-compression generations. For the video then, transcoding to ProRes engenders no further re-compression ( anologous to saving a JPG as a TIF or PSD), and therefore the sooner one can transcode to ProRes, the less data loss can occur.
In the Apple forum (as per the link inserted above) one very helpful person has suggested capturing as HDV, editing in an HDV sequence, and simply changing the Sequence Settings to Render Effects in ProRes 422. This certainly reduces storage demands as the ratio of storage for the original capture is roughly 4:1 ProRes:HDV. This perhaps is a work around, though if I understand correctly still far from ideal since doing it this way means (a) I'm still incurring a second generation of compression artifacts (and based on the above storage ratio, throwing away 3/4 of the data at that point) at the time of capture when I save the already compressed-once-on-tape HDV from that tape onto the harddrive (rather than getting much less drastically compressed data captured in ProRes right away) and (b) after rendering in the sequence in ProRes it makes it difficult to do further editing with new HDV because that would mean mixing ProRes and HDV in the same sequence (which I believe can't be done, correct?) and furthermore, I believe that I may also become disconnected from the original source media file once the rendered sequence transcodes to ProRes. On the plus side I'd be throwing out less data going forward each time I saved a render that had been left in HDV.
My questions:
Can anyone verify my observations/understanding/conclusions?
Is it going to be possible to capture into ProRes 422 and retain A/V synchronization, and if so how?
Or is the workflow suggested in the other forum (i.e. Capture to HDV, edit natively and render effects in ProRes.) the best I can hope for for now?
Does it have something to do with 24P?... that the native "24P" of the HV40 is really 23.98 whereas the 24p timeline is truly 24p, that lag is there in both HDV capture and ProRes capture and I'm just not seeing it in the HDV capture because syncing is reseting at the beginning of each clip?
Or are my issues simply a question of needing a more powerful computer than my PowerMac G5, and that for example people working in the ProRes capture workflow on Intel Macs are not having this problem?
Is it going to be possible to restore my ability to control the camera and play/read device while I am logging and capturing to FCP in ProRes 422, and if so how?
Further Background: Here is a summary of my settings, software and hardware specs
FCP 6.0.6
Mac OS X 10.4.11
PowerMac G5 (Dual Core 2 GHz)
QTPro 7.6.2 (518)
HDV (as far as I know there is only 1 type) shot on Canon HV40 and my setup is at Native 24P recorded on Panasonic DVM63PQ tape
Scratch is a secondary internal SATA drive
Playback Control is default i.e RT = Safe; Video Quality=Dynamic, Frame Rate = Full; Pulldown Pattern = 2:3:2:3; Gamma Correction = Accurate Fame Offset = 28 Record = Full Quality
Additionally:
For my General Prefs: All the following boxes are checked/activated:
Sync audio capture to video source if present
Report dropped frames during playback
Abort ETTT/PTV on dropped frames
Abort capture on dropped frames'
Warn when importing non-optimized media
On timecode break: Make New Clip
PowerMacG5 Mac OS X (10.4.11)
I am having a problem with audio sync lagging/drifting when I capture from my Canon HV40 tapes into FCP ProRes 422. Don't know if this is a user setting problem or something intrinsic to my hardware-perhaps my computer simply is too underpowered.
I've been seeking answers for this problem for nearly a month in other forums and the best I've come up with is here:
http://discussions.apple.com/thread.jspa?threadID=2105274&stqc=true
So I'm wondering whether anyone in this community can suggest anything better.
To reiterate (as per the above Apple forum link):
I am shooting on a HV40 shooting in its so-called native 24p mode, and so I presume there are no pulldown issues) and editing in FCP 6.06 (FCS v2) on a PowerMac G5 dual Core 2GHz machine with 8.5 Gb of ram, OS 10.4.11.
The specific capture preset I'm using is Apple ProRes 422 1440 x1080 24p 48kHz as I believe these setting match the specs of my camera: MPEG-1 Audio Layer 2 , 16bit. 48kHz, transfer rate 384 kbps(2 channels).
I capture from my camera (not a deck) which writes HDV files to tape during videography. If I capture in HDV into FCP with a HDV 1080 24P easy setup, i.e. without transcoding to ProRes, there is no sync problem. However because I intend to do a lot of colour grading, I have switched from using an HDV easy setup to Apple ProRes 422, but that is when I find that my audio is drifting/lagging widely over the course of the capture.
I observe that when I capture in HDV I can access the full log and capture interface and retain some FW device control of my camera through the FCP interface and that as the clips come are captured, they appear one by one in my browser. Whereas when I capture into ProRes, even though I begin the process by using the log and capture command, there is no interface presented to me and capturing begins immediately from wherever the tape happens to be sitting on the tape head of the camera and continues until the end of the tape. That is, I forfeit device control. Moreover (and I suspect my problem has a lot to do with the following observation) the clips don't appear in my browser until the tape stops rolling and the process is otherwise completed. Apropos this, it may be the case that the software doesn't break out the clips and reset the sync until the whole capture is done, and that it is in this phase of the process the lag is introduced, proportionally greater the longer the footage and the closer one gets to the final clip in the capture. If that is the case, it would perhaps be an FCP bug.
Over the course of a one hour tape, audio falls roughly 5 seconds behind.
BTW, (a) I've tried experimenting with various device protocols in the Device Control Preset Editor tab under Audio/Video Settings and cannot find a combination that regains me device control, and (b) I've also raked carefully through the various menus and sub menus to see if there is any customization I can apply to the presets (capture, sequence. a/v devices or otherwise) that might make more sense in terms of syncing audio to video, but nothing makes any better sense than the ones I'm already using.
If I were just capturing a short clip and could conveniently select clips in my camera LCD screen and navigate with my on-camera control to the in-point this sync delay probably wouldn't be a problem, but as this is a documentary with mountains of footage, such a workflow is unworkable. I need to be able capture (and preferably, log and capture) whole tapes and be able to use the more ergonomic environment of a computer to make my rough selections of clips. Captured in that way audio drift/lag becomes evident after a few minutes and as I say, extreme by the end of the tape... so extreme that if it is a short clip of a few seconds, most of the audio that should be synced with one video clip may show up in the next video clip.
As an established photographer new to video, if I understand correctly the following analogy is accurate: shooting HDV is like shooting JPGs both of which create compression artifacts on their respective initial saves to the on-board media but however whereas from a compact flash card one can copy a jpg still to a harddrive during the ingest process with no further re-compression artifacts, in comparison with HDV one has to effectively open up the file (i.e. play the video in camera out into the computer) and then re-save the HDV to the hard drive during the capture process incurring a second generation of re-compression artifacts. Once on the harddrive in Photoshop a still JPG can be worked on and saved as a tif or PSD whereas in FCP working in HDV it would save again as HDV generating yet subsequent re-compression generations. For the video then, transcoding to ProRes engenders no further re-compression ( anologous to saving a JPG as a TIF or PSD), and therefore the sooner one can transcode to ProRes, the less data loss can occur.
In the Apple forum (as per the link inserted above) one very helpful person has suggested capturing as HDV, editing in an HDV sequence, and simply changing the Sequence Settings to Render Effects in ProRes 422. This certainly reduces storage demands as the ratio of storage for the original capture is roughly 4:1 ProRes:HDV. This perhaps is a work around, though if I understand correctly still far from ideal since doing it this way means (a) I'm still incurring a second generation of compression artifacts (and based on the above storage ratio, throwing away 3/4 of the data at that point) at the time of capture when I save the already compressed-once-on-tape HDV from that tape onto the harddrive (rather than getting much less drastically compressed data captured in ProRes right away) and (b) after rendering in the sequence in ProRes it makes it difficult to do further editing with new HDV because that would mean mixing ProRes and HDV in the same sequence (which I believe can't be done, correct?) and furthermore, I believe that I may also become disconnected from the original source media file once the rendered sequence transcodes to ProRes. On the plus side I'd be throwing out less data going forward each time I saved a render that had been left in HDV.
My questions:
Can anyone verify my observations/understanding/conclusions?
Is it going to be possible to capture into ProRes 422 and retain A/V synchronization, and if so how?
Or is the workflow suggested in the other forum (i.e. Capture to HDV, edit natively and render effects in ProRes.) the best I can hope for for now?
Does it have something to do with 24P?... that the native "24P" of the HV40 is really 23.98 whereas the 24p timeline is truly 24p, that lag is there in both HDV capture and ProRes capture and I'm just not seeing it in the HDV capture because syncing is reseting at the beginning of each clip?
Or are my issues simply a question of needing a more powerful computer than my PowerMac G5, and that for example people working in the ProRes capture workflow on Intel Macs are not having this problem?
Is it going to be possible to restore my ability to control the camera and play/read device while I am logging and capturing to FCP in ProRes 422, and if so how?
Further Background: Here is a summary of my settings, software and hardware specs
FCP 6.0.6
Mac OS X 10.4.11
PowerMac G5 (Dual Core 2 GHz)
QTPro 7.6.2 (518)
HDV (as far as I know there is only 1 type) shot on Canon HV40 and my setup is at Native 24P recorded on Panasonic DVM63PQ tape
Scratch is a secondary internal SATA drive
Playback Control is default i.e RT = Safe; Video Quality=Dynamic, Frame Rate = Full; Pulldown Pattern = 2:3:2:3; Gamma Correction = Accurate Fame Offset = 28 Record = Full Quality
Additionally:
For my General Prefs: All the following boxes are checked/activated:
Sync audio capture to video source if present
Report dropped frames during playback
Abort ETTT/PTV on dropped frames
Abort capture on dropped frames'
Warn when importing non-optimized media
On timecode break: Make New Clip
PowerMacG5 Mac OS X (10.4.11)
