No, it’s fine I’m just letting you know that I tagged you in that thread in case you want to leave feedback on it or report any issues then you can go to that thread as well. If you do have issues or want to give feedback then the developers will see it faster in the beta thread.
Dear @Austin
I am totally overwhelmed, but your very well understandable help shows me a viable way. What I like best is that you talk about listening. So you are not just an engineer but also understand something about listening, audio and sound. This is really great for me.
My experience in hearing is still very limited. An audio specialist recommended the values for my spoken voice. I think he doesn’t mean Q, but dB.
- Hi-pass 100 Hz -12 dB Low Shelf = Q???
- 350 Hz -5,0 dB, Slope 3-4 dB = Q???
- 2.640 Hz +5,0 dB, Slope 3,0 dB = Q ???
- Low-pass 11.000 Hz -12 dB Hi Shelf = Q???
Yes, I understand: I have to listen for myself, which Q value suits me. And I understand that a dB value for low frequencies is quite different than for high frequencies, sounds logical when it comes to a ratio.
So, my microphone is pretty good, Rode NTG, the distance is ok, my voice too, and if I reduce at 350 Hz and increase at 2.640 Hz, it sounds better. The friend had given me the values in dB. Which Q do you think I should start with?
mickae
There’s a saying I really like…
“If it measures bad but sounds good, it’s still good. If it measures good but sounds bad, then somebody measured the wrong thing.” 
It reminds me that my ears have to be the final authority, not numbers. Except when it comes to hitting loudness targets so I don’t get rejected by the Quality Assurance department… but I digress…
I think you’re right about the audio specialist recommending dB/octave for you. I updated my previous post to talk about dB/octave. It finally occurred to me what you were describing by Flankensteilheit, which is definitely a new term to me.
Shotcut uses filters that are provided by other audio processing libraries. Those libraries chose to specify the bandwidth with Q rather than dB/octave. Shotcut has to live with their choice. If I had to guess, I would guess that the math and code to implement Q are easier to write than dB/octave, and that’s why it was written first. Maybe another filter will come along some day to implement dB/octave.
EDIT: Just saw your edit. I can try to figure out the math between dB/octave and Q, but I need to wait until evening to dive into a project like that. What I can say is that the numbers you provided are not very sharp in terms of the spike they create, so values in the area of 0.5-1.0 would be a good start point. It’s all going to come down to the sound you like hearing the best, though. Based on the -5/+5 bend in that EQ curve and the response of an NTG, it sounds like you’re going for a radio commercial sound. If that’s true, then bumping the Q as high as 1.5 might accentuate it. Values above 2 or lower than 0.5 will probably not be what you’re after.
Hallo @Austin,
Oh, your help is great. No I don’t do radio spots, they are purely artistic very short videos with philosophical content. And I only want one thing: my language to be understood as well as possible. And the settings make it less dull and therefore clearer to understand.
Before, when I was a normal person and made video editing on Windows with Vegas Pro, everything was normal, everything in dB. But I really wanted Linux and it’s hard to do video editing there. In Kdenlive it is unfortunately also like this: Q from 0.0 to 1.0. Here is an explanation that you may understand, I unfortunately do not.
Eq4p 4-band parametric shelving equaliser
http://quitte.de/dsp/caps.html#Eq4p
Can you do something with it?
That’s an interesting link. However, it calculates EQ differently than the filter in Shotcut. Any setting that works great for it, will not be optimal in Shotcut. Same problem if going the other direction. Getting any new audio tool (like this EQ filter in Shotcut) usually means starting over from scratch with new measurements or new listening tests.
If you reach the point of ear fatigue (which is super easy to do after testing 100 settings in a row), it may be time to pick your five favorite sounds so far, and then play them for somebody that hasn’t heard any of them. Let them do a blind listen, and see which they thought had the highest speech clarity. You can let your audience make certain decisions for you when feeling overwhelmed. Then you also get confirmation that the decision is actually the best decision, because the audience chose it and you didn’t have to guess.
1 Like
Thank you Austin,
it looks like work, ok, I will feel my way forward piece by piece. It’s not that hard, it’s just about my spoken voice, it should be clear and distinct.
mickae
Hello Austin,
wishful thinking shattered: I had hoped that the evaluation and settings that improve my voice could be easily transferred from a professional audio software of a Mac computer to Linux and Shotcut or Kdenlive does not work. Thanks to your skepticism I have now started to listen more and not just transfer values: And it suddenly sounds much better, because the transfers of the abstract numbers (db in Q) did not work sensibly. And it is much more fun to find results by myself and not only to take over given values blindly (i.e. numbly).
I determine the Q-values via my hearing, and what I don’t hear now, I don’t have to hear. Maybe later, then I do it again, if it seems necessary to me.
Thank you very much for all your help.
micka
If you ever want to translate something, try this out: Translated with www.DeepL.com/Translator (free version)
1 Like
Glad you found some settings that worked! Brian should be happy to know his filter is working as expected based on your experience.
1 Like
Hello @brian,
the equalizer: Parametic works great. I am impressed how fast you have implemented it.
I can make all the necessary settings and everything responds to the Effekts as you would expect. Great. And even if I change a value often, sometimes back and forth, Shotcut remains stable and does not crash. That’s also remarkable.
It all works out so well that I almost get a little sad. Why? Because since I left Windows and Vegas Pro I now work on Linux and do all my videos with Kdenlive - but this good forum here and the contact with the developers, makes me think that maybe I would be better off here at Shotcut…
No question, Kdenlive can do more, but the bugs won’t be fixed forever and now Shotcut has not only the better compressor, but also the better EQ. In addition, “Color Grading” works very well with S., with K. “Lift/gamma/gain” has a bug for ages and it is not fixed. God knows why, because this effect is one of the most important. I wonder a lot about this, have said this in the Kdenlive forum at least 5 times.
In short, the forum and the team of Shotcut are for me several classes better…
So I’m rethinking what I’m going to do. The thing I would still miss are the titles, texts. Unfortunately, Shotcut doesn’t let you set the scroll width (text running distance), or I haven’t found it yet. And unfortunately I can’t make shadows on the text - letters.
If that were possible, I would have few reasons not to switch.
I hope that I have not written too much. But video editing is the most important thing for me and unfortunately it is most difficult with Linux.
Thank you all very much, you have helped me a lot.
Translated with www.DeepL.com/Translator (free version)
Micha
2 Likes
Is it possible to create just the titles in Kdenlive, export that video in a format that preserves transparency, then overlay those title videos in a Shotcut project? Then you get the best of both worlds.
Check my tutorial on making shadows, a little more work but no so hard
2 Likes
If you want full control over your titles or other overlay graphic, then grap a frame from your video and use inkscape to make the overlay and save it as a .png or .svg and use that in your video.
It works great in both Linux and Windows.
Very true. However, the title component in Kdenlive offers a lot of animation features, which I assumed (perhaps incorrectly) would be in use. If animated titles are involved, then the Inkscape/GIMP/Krita overlay method turns into a lot of manual labor.
Hello @TimLau,
Very good, in your video is great to see that you can also make very good shadows with Shotcut. When you have learned how to do it, you need maybe 2-3 minutes, but with Kdenlive only 15 sec.
My question: If both programs are based on MLT and are very similar in many details, why can’t one take over from the other what works well? Or is it not so easy to integrate e.g. the text tool of Kdenlive?
It will not take you 2-3 minutes to copy and paste the text filter, change the color of the copied text font to black and offset the copied text filter a little to see the “shadow”.
I can do it in less than 15 seconds. 
And adding the blur part is optional but that wouldn’t even take 5 extra seconds.
They are not based on MLT. They use MLT. How one program decides to use MLT is different than the other. In any case, shadows being added inside the Text filters themselves is on the Shotcut roadmap. You can read the roadmap here.
Many people don’t like how text is done in Kdenlive. Not even Kdenlive. They’ve been working to revamp it.
Hello @Austin,
This certainly works well, but when I start with a program, then I would like to make my videos also preferably with this software to the end. Especially because I often need to make a small change at the end of the job, such as the color of the background or text - then, with an almost finished project, I’d have to change the text again in the other program, and somehow pick the color out of Shotcut’s video and paste it into the title in Kdenlive to then render the text as a video or PNG to paste it back into Shotcut. That’s certainly no fun.
For this reason, I also consider audio editing essential in the video editor, even if Audacity can do everything better. But one more adjustment just before the end? Crazy effort to insert the generated wav then again fit exactly…
I prefer to do without one or the other quality of a specialized program, but I have everything in one.
Just like I now redo all my 40 videos I made on Linux with Kdenlive, because I found a better EQ in Kdenlive and now adjust the correction by ear, thanks to you, dear Austin.
Even though I am very used to Kdenlive since 15 months, I find some things in Shotcut (Colour Correction, Compressor and EQ Parametirc) very attractive and especially the forum. I suspect I will find a way to do everything with the right software for me. Maybe Shotcut soon.
Hello @DRM
ok, text with shadow in 15 sec.
Oh, very good, I saw the roadmap and am very happy to see the improvements in the text.
An almost more important wish would be the possibility of “letter spacing”. This improves the effect of titles enormously.
I agree.
1 Like
Yes, those are all excellent points for doing the entire production inside a single program. I agree.
Some channels lock in specific colors for their text or titles just to enhance their branding and recognizability. If you had a color palette that identified your brand, then there is a chance there wouldn’t be so many last-minute color changes on titles. But if your videos are each unique creations not tied to the others, then yeah, being able to make last-minute changes easily is important.
Not sure what to tell you. It’s common for me to use six or seven tools in the creation of a video, simply because they’re each better at a specialized task. It’s inconvenient sometimes, but worth it to me. But I also have the luxury of working from storyboards and scripts, with a look and style that is defined before editing even starts. If I had to post a daily vlog, or if my clients were short-term contracts with wildly diverse styles, I would not be able to use this workflow.
1 Like
But the animation features in Kdenlive are limited, and Inkscape has no animation. Animation in GIMP and Krita are rather laborious. That is why I am excited about glaxnimate, which you will be hearing more about from me. Recently, I contributed a change, which is now in its “experimental” builds, to export as video with alpha channel if you enter a .webm filename extension. (Due to how it currently works, it does not pick a codec and lets FFmpeg libavformat choose the codec from the format based on the extension, and only WebM chooses a default codec with pixel format that includes alpha.) I am also sponsoring its development. Glaxnimate does not yet include support for shadow or blur. Otherwise, for animating text and shapes it is nice and simple. Now, let’s imagine if instead of a checkboard background to show transparency while in the tool there is the option to show your Shotcut video project… 
We can take a look at reading kdenlivetitle files, but not editing them. Its editor uses KDE libraries, and I will not add that dependency. Glaxnimate mostly uses the same libraries as Shotcut. 
A visual editor is the bigger problem in my opinion with any animation capability. For example, we used to have a HTML feature, and you can make text and even shapes in HTML, but how to make the animation in a manner that does not require a special engine that is not open source or has no seek API? I examined many tools over the years, and the only tool that came close (Tumult Hype) was macOS only and still required special manual steps to make it work in Shotcut.
1 Like