[Omer Aydin]

1) A scene change detector would be great. It can drop markers or split the clip when scene detection threshold is met.

2) Stream-copy rendering and keyframe navigaton for rendering/extracting clips without re-encoding.
(just like in VirtualDub or Avidemux)

3) An audio sample searching feature for finding specific audio patterns in audio clip events.

[Omer Aydin]

I guess that’s because Sony MP3 encoder inserts a ~0.25 sec blank(depends on the audio frame size) in both intro and outro of MP3 streams just like LAME Encoder does.
Why?
You can read the explanation about the LAME encoder here; https://lame.sourceforge.net/tech-FAQ.txt

[Omer Aydin]

Thanks for your response.
What do you mean by `all of that`?
It’s just `hitting ESC key to close a dialog box` and `switching tasks during a render’…
These are very common things in a workflow.

I’m sure you’ll get it if you follow the steps properly.

BTW, I have already confirmed this bug on a users forum and submitted it to Sony.
https://www.sonycreativesoftware.com/forums/showmessage.asp?forumid=4&messageid=940592

[Omer Aydin]

Thanks for the quick response John, I checked them with MediaInfo and the only difference was one of the files had a meta data saying it was rendered in ProTools.

I opened it in SoundForge, made a null change and saved over the original. Now Vegas is able to write marker data into that file.
I guess Protools writes some info in the area that Sony uses for marker data so Sony avoids overwriting it.

[Omer Aydin]

Thanks for your reply Aaron,

[Aaron Star] “Drop some MP3 material on the timeline and then check the properties of the file. What dll is vegas using for mp3? “

MP3 file properties says;
Plug-In
Name: mp3plug2.dll
Folder: C:\Program Files\Sony\Vegas Pro 12.0\FileIO Plug-Ins\mp3plug2
Format: MP3 Audio
Version: Version 12.0 (Build 394)
Company: Sony Creative Software Inc.

I have noticed that Sound Forge v10 also uses another version of mp3plug2.dll in it’s own plugin folder so I copied that file into Vegas folder after renaming the original file to test and I get the same result. Vegas loads the DLL without any problem but renders very slow (more than 2mins). Same DLL works very fast in SoundForge, takes less than a second to render the same MP3 file.

[Aaron Star] “Have you installed any other software that may have insall mp3 encoding changes to your system?”

Yes, I have other software on this system for more than a year but this mp3 problem appeared recently.
The latest app I’ve installed was REAPER and it uses LAME mp3 DLL for encoding. It takes 12 secs to render the same file.

[Aaron Star] “Have tried AAC,that is mp4 based and bit more updated vs mp3. Do you get the same slowness converting to AAC?
“

AAC renders just fine. I tested it using a MainConcept AVC/AAC template (video track disabled) and it renders very fast to a MP4 file as usual.

[Omer Aydin]

Nice to hear that. I’m glad it worked.

About the velocity issue; I don’t know any easy solutions for that.
I suggest using Time-Remapping tool instead of Time Stretch.
(I am assuming you are able to locate the exact position of the slider camera by eye observation.)

Lets say your footage has 100 frames.
Enable Time-Remapping and add key frames at frame 1 and 100.
Jog to the point where you think the slider camera is at the middle of the slider path and add a key frame there.
This key frame should be at the 50th frame if the footage was captured with a constant slider speed. Move the key frame to the 50th frame to fix the time shift.
Repeat this procedure for sections between frame (1-50), (50-100).
If the problem still exists, repeat for sections (1-25), (25-50), (50-75), (75-100)…
You get the idea… Repeat until you go crazy.

Hope it helps.

[Omer Aydin]

For instance, if your footage is 1920×1080, a rectangle mask of 2×1080 positioned at the center of the frame will reveal only a slice of a video frame so the motion tracker will capture the vertical movement only.
Lets say the background is white and the shelf is red. You said the shelf moves up and down while the slider moves across. When you playback the footage, the masked image will look like a thermometer indicator moving up and down. If you track the upper red edge, you’ll capture the vertical motion of the shelf.

[Omer Aydin]

Here is how to do;

1) Create a new comp (Comp1) using your footage
2) Create a few pixels wide, vertical mask slice.
3) Create a new comp (Comp2) using `Comp1`
4) Stabilize Motion in Comp2
5) Copy Anchor Position of `Comp1` layer in Comp2
6) Return to Comp1 and paste on to the footage layer.

That should stabilize your footage* vertically.

*(I’m assuming the `furnishing` you have captured has an upper edge with enough contrast)

[Omer Aydin]

I suggest creating a 2-pixel-wide mask at the middle of the footage that reveals the vertical position of the furnishing. Use Stabilize Motion with a 2-pixel-wide-tracker on the masked layer and capture the vertical motion, then apply it.

[Omer Aydin]

Anything that causes a peak in data rate will cause a false detection, like fast panning, fast zooming, flashing etc. so it wouldn’t be a good idea to use a scene detection automation on such footage.

But it will work properly with footage where peaks are due to cut transitions only.

BTW I have found an app called HandySaw DS that does it. It creates an EDL file for the video and splits the scenes into video events in timeline.