[Michael Mccune]

Glenn and Tero;

Does not seem to be a DPX (or any other format) export or import issue. The Sony Slog3 footage imports and displays in the edit and color pages normally and as expected. One can color the footage in the color page, sometimes even using the Color Space Transform tool to dip into another color space for a specific purpose and then back out. (BTW, found in someone’s tutorial that the ArriC-to-709 LUT shows a better 709 than the Transform 709 conversion. This LUT and many others are in the node context menu.)

Tero is probably right that a Fusion display LUT is what is needed. But I haven’t found one–after trying literally a hundred combinations of Project color management settings and Fusion display LUTs.

I do understand that the manual says that Fusion takes its color space and gamma from the timeline. But I have spent many hours trying Project color management settings and Fusion LUTs with no appreciable improvement.

I believe Tero is right but the result of all of my experimentation is simply no change at all in the linear display of the image. Which is baffling in itself. Or the changes that can be made may change the displayed image but only slightly.

Tero, do you have a suggested set of settings for my situation???? That is, a log source with your Project settings for input, timeline and output color space, and gamma, along with your specific Fusion Display Lut choice???

The goal is to have Fusion show an image that is not linear. Now Fusion always displays a linear (highly expanded luma and chroma) image.

Or perhaps I am missing something basic here.

Thanks again, guys. Mike

[Michael Mccune]

Duke; glad to hear that the tooth fairy brought you a nice present! Has happened to me, too, when I am at my wits end!

Great tip regarding O&O software; they certainly have an impressive range of diagnostic products.

Seriously, sometimes it seems that the tooth fairy is the only explanation. Or Whoever runs this universe finally has mercy on the poor soul beating their head against a wall.

I have also come to have this view: some software, such as Resolve or any major NLE or graphics program such as Photoshop, Maya or AutoCAD, is now so complex that no one human being can fully embrace even one of them.

Or, I have found that taking a nice relaxing walk helps: seriously, it seems that the pieces of the puzzle have a chance to reorganize themselves and, suddenly, a bright idea pops up. Or, at least, one may feel there is some new promising way to tackle a problem.

Onward and upward! Mike

[Michael Mccune]

Duke; Aside from the possible Resolve issues, as Tero says, a failing mechanical drive can cause dramatic slowdowns as the very-smart drive system tries to work around failing read and writes. It is really rather amazing how much error correction happens under the hood.

A failing mechanical drive–as distinct from an SSD drive–may also take forever to boot up and to shut down. Again, it is trying to sort things out.

Housecleaning a PC is a routine chore for paranoid folks like myself. Routinely run CCleaner (free download, https://www.ccleaner.com/“>https://www.ccleaner.com/https://www.ccleaner.com/) and Windows command line chkdsk for each drive.

Even though file detritus seems innocuous it is in fact a demanding burden as the PC tries to enumerate and organize tens or hundreds of gigabytes of old files. Takes time. That’s one reason a new PC seems so fast but quickly becomes noticeably slower.

Also, I have Hard Disk Sentinel (https://www.hdsentinel.com/) running in the background. It monitors temperature, fill rates, error messages, SMART and has various read/write tests. Another bit of confidence and insurance when things start acting in unexpected ways.

Presumably you are not using the main C drive for Resolve files; nevertheless another clue that a drive is failing is a failure to clone. You probably have a disk image or clone made every month or so to prevent that worst of disasters, an unrecoverable boot drive. If there are drive sectors that cannot be read then the clone image fails, even though the drive –mostly–seems to behave normally.

Also, SSDs get old. And when they do, even when they are supposedly far from their stated number of rewrites, then they can so south in an instant, which is not usually the case with spinning drives, which tend to try to chug along.

Best wishes! Mike

[Michael Mccune]

Thank you, Michael. Using Resolve 14.1.1 Studio and also Adobe Media Encoder for this work.

You said, “… the audio can be extracted as broadcast wav files… “

That is news to me but welcome indeed. I see there are programs to export BW metadata from existing BW files.

But, with the exception of Steinberg CUBASE software (https://steinberg.help/cubase_pro_artist/v9/en/cubase_nuendo/topics/export_audio_mixdown/export_audio_mixdown_file_format_wave_files_r.html), I do not see a method to re-encode audio in a BW format.

And in the CUBASE manual there seems to be little info on exactly how this works, referring only generally to: “…the embedding of additional file information…” by which presumably they mean timecode. How this is done is not mentioned.

FWIW, interesting factoids in the Broadcast Wave standard here: https://tech.ebu.ch/docs/tech/tech3285.pdf.

Thanks, Mike

[Michael Mccune]

Yes, you both are quite right: I did not mention professional recorders with time code.

This is from B&H, showing eight portable recorders with time code:

True, an original recording with double system sound with time code solves this problem.

But my situation is different: I have to sync audio –actually re-syncing audio–which began life as an otherwise acceptable original video with audio track (such as an Atomos Shogun with outboard Sennheiser). But the audio lost its pairing with the timecoded video due to transcoding the video into a frame sequence format which of course has no audio channels.

I guess one could re-record the audio: that is, play original video+audio, taking the timecode out of the playback device and feed that timecode and the audio into another audio recorder which does lay down timecode.

Probably too much work. But you all have at least help cover the waterfront on this issue.

Thanks, Mike

[Michael Mccune]

Rhys: Regarding the Clip mode and the Frame mode, one might first understand that these modes refer the polygon or Power Window changes. That is, changes — such as enlarging the polygon–apply to the entire clip OR they are applied to the frame only. Because Resolve will interpolate between frames, it smoothes out the frame changes made in Frame mode.

As a practical matter, I usually find myself capturing the object to be tracked (probably with the cloud option, or perhaps with the point option chosen–with a polygon window. Then start the tracking. (Usually I use the fewest of the options for Pan, Tilt, Zoom, Rotate and 3D, often just Pan and Tilt, for simple tracking in a 2D plane.)

I usually set the mode to Clip mode at the outset and then start.

It won’t be long before the window drifts from the object. Of course if one’s clip is only a few seconds then this drift issue might not occur. If is is minutes in duration, then it probably will drift.

When it drifts, I usually pause the tracking, using the pause button in the tracking window, and change the mode from Clip to Frame.

Look at the trace in the tracking window: when the track fails it usually shows a larger excursion and is fairly obvious. You might well use the tool in the tracking window to highlight and then Delete Selected Frames to clean up the track data path.

Then recenter your window on the target and recommence tracking. Rinse and repeat. This can be a tedious routine for long clips, to be sure.

This means that almost all of the tracking is done in Frame mode with lots of correcting re-do work.

With a window that basically captures and tracks your subject, then perhaps a filter applied to this limited polygon area might allow color adjustments to, say, a face.

Good luck! Mike

[Michael Mccune]

Simon; RE the new Fusion planar tracker: how well would this tool work as a polygon tracker?

That is, track a set of points outlining an object with irregular dimensions, such as a face outline. A kind of Face Refinement tool, as Resolve now offers. A kind of auto-rotoscope tool.

We see in these forums repeated mention of the fact-of-life that sometimes one must roto the shape. This is fine for short clips. But not humanly possible for long clips, such as an interview: simple enough in regards to the shape but infeasible in respect of the many thousands of frames required. And no slaves.

Fusion ChromaKey masking seems the better of the options I have tried for this sort of work.

Thoughts???

Thanks, Mike

[Michael Mccune]

Dylan; right, changing the External Matte Saver option under Matte from “A” to “RGBA” does indeed output a usable file and one that Resolve recognizes.

But of course it is no longer just an alpha channel, at <100KB per frame, but a video frame at about 4MB per frame.

Onward and upward!

Mike

[Michael Mccune]

Thanks. Found the External Matte Saver tool, as you suggested, under, “Resolve Connect,” after opening the tool menu from the open Flow area (not visible with an already- highlighted or selected tool nor via the Tools menu).

And it does generate a frame sequence output. The file size, less than 100KB, does indeed indicate only an alpha channel is generated.

BUT the file format seems to have only one output, .mov.

The format selection option in the file save menu lists the usual dozens file formats. But a selection from that list seems to have no effect. Only a .mov file format is generated in the output.

AND the resulting .mov file is not recognized by Resolve, so no way to import this Fusion output file.

MediaInfo (“https://mediaarea.net/en/MediaInfo”>https://mediaarea.net/en/MediaInfo) can’t seem to pull any format info for this .mov variant.

Any thoughts?

Thanks, Mike

[Michael Mccune]

AutoProxy was it, Simon. Thank you!
Mike