Hey Eugen!

Thank you for your great insights.

Yes, I use a 59.9FPS 1080p video profile and I am almost positive that the 
entire audio is 44100 Hz. So you are saying that this is a known issue under 
these specific conditions? (I will do my research right now, thank you!)

I always try to cut audio when there is silence, to be more specific, I make 
MMA Youtube videos, and mostly edit interviews. I make my cuts exclusively 
during pauses or outro sections.

But the pops also occur randomly with clips that I did not cut/edit at all or 
of which I positively know that they are cut correctly without any lingering 
sound pieces at the beginning, but you have already given a good explanation 
now.

I will compile the latest ffmpeg, ffplay and ffprobe and report back.
Is there anything else I should consider regarding this out of synch issue?

Maybe increase the Hz of the audio that I use?

Regards

John

Sent with [ProtonMail](https://protonmail.com) Secure Email.

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
On August 21, 2018 8:53 PM, Eugen Mohr <eugen.m...@gmx.net> wrote:

> Hi John
>
> It comes 2 things to my mind:
> -    Do you cut the audio at the zero point? In Audition (for example) is a 
> function that you can cut only at that point (when the wave goes through 
> zero) to avoid clicks.
> -    When audio frequency (i.e. 44.1kHz) and video frequency (i.e. 60fps) 
> interfere each other after a time and from there you get clicks. Video 
> rendering is always video frequency driven. This would be a FFMPEG problem.
>
> Merlimau
>
> Am 21.08.2018 um 19:55 schrieb johnar1:
>
>> Hey Jean, thank you for that valuable info.
>>
>> I have tested the 18.08 version that you linked and unfortunately the pops 
>> at the beginning of audio clips still occur, but I have made an interesting 
>> discovery.
>> When I render the entire project, there are three pops at the beginning of 3 
>> clips, let's say at 01:23, 02:02 and 03:15 in the time line.
>> These audio pops only come up in the rendered output, not during playback.
>> When I only render a small selected zone surrounding the area where I know 
>> the pop will be, it does not occur.
>> Only when I render the full project, they appear.
>> I have tried this with all three known audio pops that I have spotted in the 
>> fully rendered project, and each time they disappear when I only render a 
>> selected area around them.
>>
>> The next thing I have noticed is that switching the playback settings betwen 
>> SDL and RTAudio  can affect this issue as well.
>>
>> As you know me now, I will now enter full testing mode and report back to 
>> you with a more detailed analysis.
>> I will compile the latest melt, ffmpeg and co in order to provide a full 
>> report, but I will use the AppImage versio 18.8.0 which you have suggested.
>>
>> Anything else I should consider?
>>
>> Best regards,
>>
>> John
>>
>> Sent with [ProtonMail](https://protonmail.com) Secure Email.
>>
>> ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
>> On August 20, 2018 10:23 AM, Jean-Baptiste Mardelle 
>> [<j...@kdenlive.org>](mailto:j...@kdenlive.org) wrote:
>>
>>> On 18.08.2018 14:55, johnar1 wrote:
>>>
>>>> Hey Jean and guys!
>>>
>>> Hello Johnar!
>>>
>>>> I have uncovered a new problem with the latest stable AppImage 18.04.1 
>>>> release of kdenlive which is very troublesome.
>>>>
>>>> Randomly, kdenlive will cause a "click" or "pop" sound at the beginning of 
>>>> an audio track. I have observed that this happens the most when two audio 
>>>> tracks are overlapping or a cut has been made at the beginning of the a 
>>>> clip.
>>>>
>>>> I have also found that audio pops disappear in the same footage disappear 
>>>> when switching from SDL to RTAudio in the settings.
>>>>
>>>> I am using a fresh install of Kubuntu 18.04 with updates and
>>>> Mlt and ffmpeg versions all from the Appimage, so this has to be a thing 
>>>> in kdenlive or maybe the latest Alsa / Pulseaudio package?
>>>>
>>>> I have tried some stuff suggested in forums such as exporting only the 
>>>> audio in wav, but without luck.
>>>
>>> I have just uploaded the 18.08.0 AppImage containing the latest MLT. Could 
>>> you check if the problem still happens ?
>>> It is available here:
>>>
>>> https://files.kde.org/kdenlive/release/kdenlive-18.08.0-x86_64.AppImage.mirrorlist
>>>
>>> Thanks for your feedback.
>>>
>>> Jean-Baptiste
>>>
>>>> Any ideas?
>>>>
>>>> Regards
>>>>
>>>> ‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
>>>> On July 26, 2018 9:20 AM, Jean-Baptiste Mardelle 
>>>> [<j...@kdenlive.org>](mailto:j...@kdenlive.org) wrote:
>>>>
>>>>> On 23.07.2018 00:21, Juku Trump wrote:
>>>>>
>>>>>> Hi,
>>>>>
>>>>> Hello Juku!
>>>>>
>>>>> Thanks a lot for your contribution and sorry for my late reply. Your help 
>>>>> is very welcome. I have now merged your contribution. Regarding the 
>>>>> tests, they need some work and currently don't all pass and crash. We 
>>>>> need to work on that.
>>>>>
>>>>> Regarding other tasks, we really need to make a list of UI improvements 
>>>>> that could be done without too much knowledge of the whole code. I will 
>>>>> try to spend some time on it tomorrow, help is welcome. But basically, if 
>>>>> there is any UI related thing that you want to improve, feel free to ask 
>>>>> and we can guide you if needed.
>>>>>
>>>>> Best regards,
>>>>>
>>>>> Jean-Baptiste
>>>>>
>>>>>> My name is Juku. I am a film hobbyist and I have been using Kdenlive for 
>>>>>> about 8 years. As I'm also a developer, I thought I would try my hand on 
>>>>>> contributing to Kdenlive.
>>>>>>
>>>>>> I have mostly been a web developer so far, so I do not have much 
>>>>>> experience with C++ (except for some tutorials) and I have never done 
>>>>>> any Qt development.
>>>>>>
>>>>>> I viewed the "Junior Jobs" section in Kdenlive Development Information 
>>>>>> page and tried to fix #384511 for start ("New project window does not 
>>>>>> fit to laptop screen (1366x768)"). Thanks to a hint by Christoph Feck in 
>>>>>> the comments below that bug, I got it fixed and submitted the patch 
>>>>>> ([D14281](https://phabricator.kde.org/D14281)) for review.
>>>>>>
>>>>>> Currently I'm having some trouble with the automated tests. Should they 
>>>>>> succeed in their current state or are there some known problems with 
>>>>>> them?
>>>>>>
>>>>>> It would be helpful if someone would suggest some other tasks which I 
>>>>>> could start with.
>>>>>>
>>>>>> Thanks,
>>>>>> Juku

Reply via email to