Have you tried to produce the same video without SVRT? I'm curious because I noticed that same thing before and have never noticed if SVRT makes a difference. I generally work with MPEG 2 and produce to MPEG 2, so I may have to play with that some and see what happens. __________________________________
CORNBLOSSOM

Hi Andrew,

I must say, I find the title of your thread a little misleading - this is a potential bug but it does not mean SVRT is unusable.

I've never encountered this paticular problem.

So basically what you're saying is that the volume levels appear to change from one clip to the next, or is it only when, say, two clips have a transition between them that needs to be re-encoded?

What camera are you using? Most of the SVRT conversations this weekend have been based around Sony and Canon!

Hope to hear from you soon,

Cheers,

Andrew

Alienware Aurora ALX R4 - Intel i7-4820 4.2 GHz - 32GB DDR3 RAM - Crucial 512GB SSD - 1TB Seagate HDD - 3TB WD Green HDD - 4TB WD Green HDD - MSI NVIDIA GTX 1070 8GB

Sony HDR-PJ810 and HDR-PJ530

Hi Andrew,

If it was just a case of each clip being at slightly different volume levels that would be ok, however the problem is that for the first and last 2 seconds of each clip the volume is slightly louder than the main body of the clip even when there are no transitions or titles. PD has to render these portions without SVRT to ensure the clips swap cleanly. I can also notice the audio boost effect if you add a title or transition.

The effect is so noticable in my case I can't use SVRT - with the option enabled it distracts the viewer with little volume increases and decreases around each clip boundary! I certainly wasn't trying to be misleading with the title of the post - it's just true in my case I'm afraid!

I do use a Canon but am using the 2508 build and as I understood it the problems with Canon where with the video encoding. Surely the audio can't be so different as to cause this effect?

Has anyone else experienced this problem?

Andrew

Andrew, what is the format of your raw video files? Providing the file type, bit rate, resolution, and camera model would be helpful in trying to duplicate your issue. I also have a Canon camera so I can give this a shot. Q9300 2.5 GHz
4 GB Ram
Nvidia 9800 GT

Hi Andrew,

Very interesting, but extremely irritating for you!

I've been working on my cousin's wedding this evening and have used multiple transitions, but have noticed no audio issues at all I'm afraid.

I've also just run a quick test just to double check, with no problems.

As Jaime has mentioned, and I agree with him, you could re-encode the entire project and bypass SVRT altogether. As you know though, you lose the speed in production and some, but not much I don't think, picture quality.

Or, you continue to use SVRT but use transitions rarely and only when the two clips are low noise clips.

James and Jeff both have Canons. Maybe they'll check if they can replicate your issue.

Sorry I can't be more help!

Welcome aboard by the way!

Cheers,

Andrew Alienware Aurora ALX R4 - Intel i7-4820 4.2 GHz - 32GB DDR3 RAM - Crucial 512GB SSD - 1TB Seagate HDD - 3TB WD Green HDD - 4TB WD Green HDD - MSI NVIDIA GTX 1070 8GB

Sony HDR-PJ810 and HDR-PJ530

Ok, I had a go with this.

I believe that Andrew may be on to something. I took a 24 Mb/s 1920x1080i AVCHD clip from my Canon HG20 and split the clip in two places. Between each split I added a transition and produced the file to AVCHD 24 Mb/s with SVRT. I played back the resulting video in Splash, VLC, and Windows Media Player (the version that comes with Windows 7) and I do believe that there is a change in volume levels. In my case the volume seemed to dip as the transition started and then returned to the normal volume at the end. It did this for both transitions. I had my wife give me her expert opinion on this as well.

Can someone else give this a try? Q9300 2.5 GHz
4 GB Ram
Nvidia 9800 GT

I have a Canon DV, standard definition 16x9 camcorder. I capture to PD in MPEG 2 and produce to MPEG 2. I am using 2508. Ok, I have a previous video still saved on my computer where the effect is a clear as day where it fades between two clips, so I imported the video to PD. I cut out 10 seconds and split that in half. I applied a fade.

Next I produced the video to MPEG 2 using SVRT, and again using hardware video encoding and again using neither. Now I can't hear the effect in any of them. I'll try some other clips, but that is exactly what I had trouble with in the past. __________________________________
CORNBLOSSOM

Andrew

I took a look and this is what I found.
I recorded a 400 Hz constant amplitude tone into my Canon and recorded at 24Mbps and 1920x1080. I then split the video at 15 seconds and added a 2 second transition. So the transition occurs from 13-15 seconds. I then looked at the audio wave form of three conditions.
1) Just dropped the raw clip with no transitions into the time line and used SVRT. The audio output of this is a constant 400 Hz amplitude sin wave. No surprise, it’s what it should be.
2) Split the video as described above and produced with SVRT off and "Allow SVRT on..." unticked in the Produce tab. (i.e. everything rendered)
3) Split the video as described above and produced with SVRT and "Allow SVRT on..." ticked in the Produce tab.

There is an obvious difference in how the audio is handled in items 2 and 3. I don't see any increase in amplitude in either case though for the downstream clip relative to the upstream clip. For case 2 above the ear may perceive that it gets louder since the audio level decreases a reasonable dB level and then increases again to my constant tone amplitude. (See pic, top chart red horizontal line) In item 3 above, it looks like they create a dead zone right at the start and end of the transition, but during the transition it is almost constant amplitude. Most ears wouldn't catch the ~.01sec lapse in sound.
The above was done with Dolby Digital 5.1 unticked. This will have an effect in overall amplitude because of the Dolby noise suppression and also an effect in characteristics as PD8 would do audio rendering. Not sure what audio you actually used.

Jeff

Andrew,

It may be worth me mentioning that I have Dolby Digital 5.1 highlighted when I produce, because my camera records in that format.

From what Jeff has discovered, it seems that might be why I haven't noticed anything.

Cheers,

Andrew Alienware Aurora ALX R4 - Intel i7-4820 4.2 GHz - 32GB DDR3 RAM - Crucial 512GB SSD - 1TB Seagate HDD - 3TB WD Green HDD - 4TB WD Green HDD - MSI NVIDIA GTX 1070 8GB

Sony HDR-PJ810 and HDR-PJ530

Jeff, well done. I'm glad someone was able to quantify what was going on. I think you and I were basically observing the same thing. I'll have to try retesting with Dolby 5.1 checked. In my previous tests with SVRT I had the audio muted since I only cared about the video. I just assumed the audio would not be affected.

I guess this is one more thing for Cyberlink to look at. Q9300 2.5 GHz
4 GB Ram
Nvidia 9800 GT

It may be worth me mentioning that I have Dolby Digital 5.1 highlighted when I produce, because my camera records in that format.

Wow I'm impressed, I'd like to see your camera sound stage, Dolby 5.1, six separate audio channels, front left, front right, center, rear right, rear left, and a subwoofer. I don't have a sound stage and my Canon only has left and right channels. I do have a shotgun mic and that can be fun though.

Jeff

Jeff,

See for yourself

http://www.amazon.co.uk/Sony-HDRXR200VE-Definition-Handycam-Camcorder/dp/tech-data/B001PGXA50/ref=de_a_smtd/279-7104413-6318137

Andrew Alienware Aurora ALX R4 - Intel i7-4820 4.2 GHz - 32GB DDR3 RAM - Crucial 512GB SSD - 1TB Seagate HDD - 3TB WD Green HDD - 4TB WD Green HDD - MSI NVIDIA GTX 1070 8GB

Sony HDR-PJ810 and HDR-PJ530

Hello All

[PD8 v2508]

Yes, unfortunately I have to confirm there is an issue with SVRT audio processing right at the beginning and at the very end of each transition section. Sound gets louder when entering the transition phase, stays OK during the transition and another loudness tick when exiting the transition.

This is definitely something before releasing any patch for PD8.

On the other hand, SVRT processing works for my Canon HF200 AVCHD camcorder so I am extremely happy with the job of the RnD department at Cyberlink!

If they can fix the audio bug for SVRT as well I think the whole SVRT issue could be closed.

Keep up the good work guys at RnD ! (but pleeeeze get this little one sorted for us )

Kind regards

/z

Thanks Jeff and Andrew,

Ticking on 5.1 has solved the problem however now the audio sounds badly compressed. Not sure if compressed is the right word though - it's kind of like someone has played my audio in the bathroom and rerecorded it! I have noticed however the standard 1080 24Mbps profile uses 5.1 channel audio whereas my source clips, while using Dolby 'compression', are using the stereo channel.

I've made a custom profile matching the audio type from my source files and am trying a render now to see if that's the trick!

Andrew

For some reason my attached pic did not upload correctly, I reattached it and it appears okay now for those that tried to download and could not. It did not violate size.

As I said, wow pretty impressive that Sony has 5.1 in the camera.

Jeff

Quote:
I've made a custom profile matching the audio type from my source files and am trying a render now to see if that's the trick!
Andrew

Sorry to report that the audio 'ticks' are back with my experiment. So it seems I have a choice between audio with 'ticks' between clips or audio where the background hiss appears to be amplified and played down a cardboard tube!

Anyone any ideas for how to get rid of what I shall call the 'toilet roll' effect when making 5.1 audio from stereo source??

Andrew

You might try CLVS. I typically don't use it so can't directly comment. Might be worth a shot considering the issues you have. How you go from 2.0 channel to 5.1 can at times be an issue. It really depends what they do behind the scene to fake 5.1. Often one copies FrontL and FrontR channel to RearL and RearR and then add a slight reverb or slight delay, for the center you often add L+R and drob dB level. In PD8 I've seen outputing Dolby 2.0 camera footage to Dolby 5.1 can lead to a overall amplitude dB reduction.

Jeff

I too have never used CLVS, so I can't comment on it either. Definately worth a try!

This is an excellent discovery you've made Andrew. I know Dafydd ensures our findings go back to R&D and they do attempt to put these issues right.

What's essential, and I know this from experience, is that they can replicate your issue. Quite a few users have submitted clips, including me, after Dafydd requested we do so on here:

http://forum.cyberlink.com/forum/posts/list/8727.page

Hopefully someone with your camera will have done so already and they can use these to replicate the issue.

Cheers,

Andrew Alienware Aurora ALX R4 - Intel i7-4820 4.2 GHz - 32GB DDR3 RAM - Crucial 512GB SSD - 1TB Seagate HDD - 3TB WD Green HDD - 4TB WD Green HDD - MSI NVIDIA GTX 1070 8GB

Sony HDR-PJ810 and HDR-PJ530

Thanks guys - will try CLVS. Added camera details to the post you suggested - just waiting for FTP details to upload clips.

Andrew