Audio: 5.1 vs Stereo

amendegw wrote on 9/18/2011, 3:32 PM
Can someone 'splain this to me?

If I shoot a clip on my Panny TM700 with 5.1 audio and drop in to the Vegas timeline with Project properties set to "Stereo", I get the following (note the volume bias to the right channel).


If I change the project properties to Audio 5.1 and drop the clip to the time line I get the following (note the right & left waveforms for Channels 1/2 are approximately equal):


So, the question is... "If project properties are set to 'Stereo', shouldn't I expect to see Channels 1/2 (only) on the timeline?"

...Jerry

System Model:     Alienware M18 R1
System:           Windows 11 Pro
Processor:        13th Gen Intel(R) Core(TM) i9-13980HX, 2200 Mhz, 24 Core(s), 32 Logical Processor(s)

Installed Memory: 64.0 GB
Display Adapter:  NVIDIA GeForce RTX 4090 Laptop GPU (16GB), Nvidia Studio Driver 566.14 Nov 2024
Overclock Off

Display:          1920x1200 240 hertz
Storage (8TB Total):
    OS Drive:       NVMe KIOXIA 4096GB
        Data Drive:     NVMe Samsung SSD 990 PRO 4TB
        Data Drive:     Glyph Blackbox Pro 14TB

Vegas Pro 22 Build 239

Cameras:
Canon R5 Mark II
Canon R3
Sony A9

Comments

Geoff_Wood wrote on 9/18/2011, 3:53 PM
V10e ?

geoff
farss wrote on 9/18/2011, 4:00 PM
"If project properties are set to 'Stereo', shouldn't I expect to see Channels 1/2 (only) on the timeline?"

No. To get stereo from a 5.1 recording all channels need to fed into a matrix to create the stereo equivalent. The spatial arrangement of the front left and right channels in 5.1 is different (wider) to stereo.
I'm no guru on 5.1 but I do know in 5.1 the centre channel is there because of the greater spatial distance between the other front channels. Without it there's a hole in the soundfield. The centre channel generally carries the dialog. If Vegas did what you think it should channels 1/2 would be missing the dialog.

Bob.

amendegw wrote on 9/18/2011, 4:10 PM
@Geoff_Wood, Yes, 10.0e 64bit.

@farss, I guess I still don't understand. If you look at the 5.1 waveforms, I don't see anything that would cause a bias to the right channel.

...Jerry

System Model:     Alienware M18 R1
System:           Windows 11 Pro
Processor:        13th Gen Intel(R) Core(TM) i9-13980HX, 2200 Mhz, 24 Core(s), 32 Logical Processor(s)

Installed Memory: 64.0 GB
Display Adapter:  NVIDIA GeForce RTX 4090 Laptop GPU (16GB), Nvidia Studio Driver 566.14 Nov 2024
Overclock Off

Display:          1920x1200 240 hertz
Storage (8TB Total):
    OS Drive:       NVMe KIOXIA 4096GB
        Data Drive:     NVMe Samsung SSD 990 PRO 4TB
        Data Drive:     Glyph Blackbox Pro 14TB

Vegas Pro 22 Build 239

Cameras:
Canon R5 Mark II
Canon R3
Sony A9

farss wrote on 9/18/2011, 4:48 PM
" If you look at the 5.1 waveforms, I don't see anything that would cause a bias to the right channel."

I noticed that as well but can only take a guess as to why it may be happening. The sum of two signals also involves the phase of the two signals. If the two left channels are out of phase they cancel, if the two right channels are in phase they add.

Bob.
amendegw wrote on 9/18/2011, 5:26 PM
"I noticed that as well but can only take a guess as to why it may be happening. The sum of two signals also involves the phase of the two signals. If the two left channels are out of phase they cancel, if the two right channels are in phase they add."If anyone would like to examine the source, it's here: 00039.zip

Bonus points will be issued if you can figure out how to eliminate the crowd noise - I've failed.

...Jerry

System Model:     Alienware M18 R1
System:           Windows 11 Pro
Processor:        13th Gen Intel(R) Core(TM) i9-13980HX, 2200 Mhz, 24 Core(s), 32 Logical Processor(s)

Installed Memory: 64.0 GB
Display Adapter:  NVIDIA GeForce RTX 4090 Laptop GPU (16GB), Nvidia Studio Driver 566.14 Nov 2024
Overclock Off

Display:          1920x1200 240 hertz
Storage (8TB Total):
    OS Drive:       NVMe KIOXIA 4096GB
        Data Drive:     NVMe Samsung SSD 990 PRO 4TB
        Data Drive:     Glyph Blackbox Pro 14TB

Vegas Pro 22 Build 239

Cameras:
Canon R5 Mark II
Canon R3
Sony A9

musicvid10 wrote on 9/18/2011, 5:52 PM
Best way to get stereo is open in a 5.1 project, then downmix at the master. You still have control over the individual track levels.

"I noticed that as well but can only take a guess as to why it may be happening. The sum of two signals also involves the phase of the two signals. If the two left channels are out of phase they cancel, if the two right channels are in phase they add."

May very well be correct, and may relate to the already-documented phase shift bug on the DVDA timeline, which is stereo, and looks similar.
amendegw wrote on 9/18/2011, 6:13 PM
"Best way to get stereo is open in a 5.1 project, then downmix at the master. You still have control over the individual track levels.""downmix that the master?? I need some help here. My gut tells me to merely mute everthing but the 1/2 channels and render from there.

...Jerry

System Model:     Alienware M18 R1
System:           Windows 11 Pro
Processor:        13th Gen Intel(R) Core(TM) i9-13980HX, 2200 Mhz, 24 Core(s), 32 Logical Processor(s)

Installed Memory: 64.0 GB
Display Adapter:  NVIDIA GeForce RTX 4090 Laptop GPU (16GB), Nvidia Studio Driver 566.14 Nov 2024
Overclock Off

Display:          1920x1200 240 hertz
Storage (8TB Total):
    OS Drive:       NVMe KIOXIA 4096GB
        Data Drive:     NVMe Samsung SSD 990 PRO 4TB
        Data Drive:     Glyph Blackbox Pro 14TB

Vegas Pro 22 Build 239

Cameras:
Canon R5 Mark II
Canon R3
Sony A9

ChristoC wrote on 9/18/2011, 9:25 PM
The answer can be deduced from the file properties:
Streams
Video: 00:00:36.537, 59.940 fps progressive, 1920x1080x12, AVC
Audio 1: 00:00:36.537, 48,000 Hz, 5.1 Surround (stereo downmix), Dolby AC-3
Audio 2: 00:00:36.537, 48,000 Hz, 5.1 Surround, Dolby AC-3

The Stereo Audio Stream is NOT decoded by Vegas; therefore looks weird.
The 5.1 Audio Stream is decoded by Vegas.

Therefore in Vegas, use the 5.1 Stream; if you want just Stereo, downmix that 5.1 stream, using your ears not guts, combining in a way that satisfies you (to my ears the intelligibility of the dialog is best in this case when downmixed to Stereo using just Front Left, Front Right & Centre).
musicvid10 wrote on 9/18/2011, 10:14 PM
Christo,
If Audio 1 reports the same from Jerry's footage as from DVD or BDAV, it is not an actual stream, just a downmix flag.

It is reasonable to assume Vegas ignores the flag.

But I think Vegas is attempting to make sense of the actual 5.1 stream, and is doing so incorrectly, "perhaps" as a result of the same phase shift bug acknowledged by Sony in DVDA. Although I have no proof of this, it is logical because they both use the same decoder . . .
john_dennis wrote on 9/18/2011, 10:32 PM
"Don't do what I do, do what I say do."

It's late now, but you might get better results if you can turn off AGC and set the audio levels manually. I seem to hear the crowd noise rising in volume between the words of the subjects. I studied the manual on my camera and "learned" how to do it. Then, last Sunday I couldn't remember how when I was shooting my grandaughter's birthday party. Lot's of crowd noise at a party for a five year old.

I rendered the two front channels and the center channel into a stereo wav file from a 5.1 project.

"merely mute everthing but the 1/2 channels and render from there."

You can use the Surround Master Mixer to set levels to suit your ear including mute. I agree that Front L&R plus center sounds as good as anything.
Red Prince wrote on 9/18/2011, 10:40 PM
I'm no guru on 5.1 but I do know in 5.1 the centre channel is there because of the greater spatial distance between the other front channels. Without it there's a hole in the soundfield.


The way I understand it, the center channel is for the dialog, so it is easier for us to separate it from all other sounds.

He who knows does not speak; he who speaks does not know.
                    — Lao Tze in Tao Te Ching

Can you imagine the silence if everyone only said what he knows?
                    — Karel Čapek (The guy who gave us the word “robot” in R.U.R.)

farss wrote on 9/19/2011, 12:40 AM
Link below is to a huge amount of information about all things "surround"

http://www.wendycarlos.com/gosurround.html

Bob.
amendegw wrote on 9/19/2011, 5:03 AM
Okay, I did a couple more tests.

First, I rendered the 5.1 project to DNxHD intermediate and created a new (stereo) Vegas project. The left/right channels are nominally equal.


Second, in the 5.1 audio project, I rendered the audio only to a stereo track. Again, the left/right channels are nominally equal volume:


So, what can I conclude from this? Either there's a bug in Sony Vegas that introduces right channel bias when dropping 5.1 audio in a stereo project. Or, there's some funkiness in the audio conversion that I don't understand.

...Jerry

Edit: If I open the mts source clip in Sound Forge 9, I see the same right channel bias.

System Model:     Alienware M18 R1
System:           Windows 11 Pro
Processor:        13th Gen Intel(R) Core(TM) i9-13980HX, 2200 Mhz, 24 Core(s), 32 Logical Processor(s)

Installed Memory: 64.0 GB
Display Adapter:  NVIDIA GeForce RTX 4090 Laptop GPU (16GB), Nvidia Studio Driver 566.14 Nov 2024
Overclock Off

Display:          1920x1200 240 hertz
Storage (8TB Total):
    OS Drive:       NVMe KIOXIA 4096GB
        Data Drive:     NVMe Samsung SSD 990 PRO 4TB
        Data Drive:     Glyph Blackbox Pro 14TB

Vegas Pro 22 Build 239

Cameras:
Canon R5 Mark II
Canon R3
Sony A9

PeterDuke wrote on 9/19/2011, 5:34 AM
If I set a 5.1 project and drag a set of clips (shot in the street) to the timeline, the left and right pairs look more or less similar. If I listen with stereo headphones, the sound is what I expect. If I now set the project to stereo and drag the same clips to the timeline, the lower channel (right?) is much higher amplitude. If I listen with stereo headphones it sounds weird - almost as if the top channel is the rear and the lower channel is the front.

The effect is the same with 9c and 10e.
Rob Franks wrote on 9/19/2011, 6:45 AM
It's a bug in Vegas.... but apparently it's isolated to the panny cam. When I do the same with my SR11 everything works fine.

Interestingly enough, I see the exact same thing in DVDa when importing 5.1 from the "studio" encoder for Blu Ray (DVDa does not display or playback 5.1.... it downmixes for preview). But when I use the PRO encoder the downmix is fine. I have always assumed that the studio encoder is kind of a 'fake' encoder and that's why the issue. Now I'm not saying the 5.1 from your cam is 'fake' but there does appear to be something different about it.

Is the 5.1 sound on the panny cam a registered Dolby system as in the SR11 (is there an official Dolby logo)?

EDIT;
I checked the Dolby and it does appear to be registered so I have NO idea why it works with the SR11 and not with your cam
PeterDuke wrote on 9/19/2011, 7:25 AM
"but apparently it's isolated to the panny cam"

My camera is a Sony.
dalemccl wrote on 9/19/2011, 7:45 PM
>>>"but apparently it's isolated to the panny cam"<<<

>>My camera is a Sony.<<

Mine is also a Sony and I see the same issue with 5.1 audio in a stereo project.
john_dennis wrote on 9/19/2011, 8:24 PM
Bump the hit counter by one. Here is the Sony HDR-SR1

I almost always start with a 5.1 project when I work with these files but the tell-tale signs are there.
PeterDuke wrote on 9/19/2011, 8:26 PM
Can anyone else check whether my theory is correct? Upper channel = rear, lower channel = front.
ChristoC wrote on 9/19/2011, 8:50 PM
The 'stereo' track as observed in Vegas looks suspiciously like a AC-3 Studio 2-track stream encode, as evidenced in snapshot below;



upper trace pair is from DVDArch and is a 2-track AC3 Studio file produced by render in Vegas of the original file's 5.1 stem.
(This AC3 file decodes to a perfectly good 5.1 balance in WinDVD, similar to the original)

lower trace pair is the original 2-track stream as imported into a Vegas Stereo session.

AC-3 2-track encoded streams are NOT the same as Stereo Stream.
It is NOT Upper channel = rear, lower channel = front.

I stand by my original post.

Why is Octoberfest being held in September?????
Rob Franks wrote on 9/19/2011, 9:49 PM
Well it's clear that what ever vegas is doing (or not doing) is the same as what DVDa does when the studio encoder is used. I wonder if the studio encoder has anything to do with decoding for the vegas time line?
musicvid10 wrote on 9/19/2011, 10:49 PM
Here are the original threads on the subject; still seems to be directly related to the AC3 phase shift flag, as I've mentioned a couple of times in this thread. Please read the following threads because they may spare some speculation . . .

2009
http://www.sonycreativesoftware.com/forums/ShowMessage.asp?ForumID=22&MessageID=655195

2008
http://www.sonycreativesoftware.com/forums/ShowMessage.asp?ForumID=19&MessageID=612261

Following the history back, it seems that the stereo timelines in DVDA and Vegas handle 5.1 correctly when the AC3 90 deg. phase shift flag is enabled, but not with the flag disabled.

It would seem plausible, but would bear some further investigation to confirm that the 90 deg. phase shift flag is indeed disabled in the Panasonic 5.1 footage.

It is important to note that this is a Stereo PREVIEW bug, and that renders come back at correctly balanced levels. Perhaps that is why Sony has done nothing about it. In the same sense, I welcome more testing and verification that what Jerry is documenting is indeed the same bug.

Nevertheless, it is important to note that Sony has acknowledged and held this anomaly in their hopper for at least three years, and they have either forgotten about it, or else tacitly chosen to do nothing . . .

PeterDuke wrote on 9/19/2011, 11:09 PM
I rendered a clip with a "clunk" sound to six different outputs each with a different single channel at 0 dB, the others at -inf dB.

I loaded them all into a 5.1 project and listened to them with stereo headphones. The left, centre and left rear appeared in the left headphone, while the right, LFE and right rear appeared in the right phone.

The centre and LFE are strange, but not the subject of this thread.

I then loaded the clips into a stereo project and examined the audio waveforms in an audio editor. Designating the source channels as L, C, R, Lfe, Lr, and Rr, I got for the stereo left channel: L+C-Lr-Rr and for the right channel R+C+Lr+Rr. The source channels may have a scaling constant that I did not check for. The LFE channel was not used. In a diffuse sound field, all channels will pick up much the same signal, so you can see that the right channel will have greater amplitude, which is what we get. I saw no evidence of 90 degree phase shift, only 180 degrees.

Post script

I rendered out the stereo project to a stereo file and loaded it back in again. Nothing got fixed during the render.

Post post script

I checked that all channels (except LFE) in the source clip were in phase with one another (more or less).
musicvid10 wrote on 9/19/2011, 11:36 PM
This will be easy enough to test, but later in the week . . .