I tried it with DALL-E 2 and had a well generated female with worse spelling and wrong number of letter groups (I can’t call them words), so it isn’t a new thing. It has to be said I thought regurgitating words I gave it would have been easier than generating a new person. Who knew?

The very new and free Dall-E 3 has been highly praised for it's ability to create words within images. Prior to this, Dall-E - and also others like MidJourney - could only display gooble-de-gook scribbles. My prompt was: photo of a pretty blond woman in her twenties wearing a white t-shirt which says "The Flat White Album" written in large black capital letters. The resulting lettering IS remarkably good and it follows contours of both the body AND the cloth folds, but not one of the three logos presented was spelled correctly!! No doubt this will be fixed within weeks ... and then MidJourney will likely frog leap Dall-E with many new features. Progress in AI image generation is absolutely astounding.

Music visualizers are great way to present music! 👍 Though speaking as someone who prefers building their own tailored visualizers from scratch, I would like to briefly touch on similar vein to what @GregB said. Custom music visualizers which are exclusively tailored to the content of you're music can be created absolutely for free by using open source software like Blender. I could also recommend paid software or other free software if anyone was interested. Granted like with all software there's is a learning curve, but there are also plenty of work-around's that make things easier, not to mention, there are plenty of free model catalogs to pick and choose from. Which, if combined creatively, can produce impressive results which can be tailored to you're specific musical vision and creative needs. Though to be fair, I would also like to add that Videoblot does seem nice and I like how user friendly it is. I think its a very useful tool for people that are only looking for templates to use in their videos, and with minimal hassle.

I think the main thing here is that we are now finally on the road to AI video. It's still early days and everything I've seen so far (Sep '2023) LOOKS experimental or proof of concept, cartoony, ill-defined edges, and very SHORT clips. But considering the steep J-curve improvements seen in the last 12 months with image generation, video (and music) are perhaps just 6-months away from stunning 'can't tell the difference from real footage' results, along with consistent characterisation.

... though personally I'm not a fan of arbitrary images which have no relationship to the song's story.

My latest (and likely my last) music video. It had to wait almost 3 years after the track's release in 2020 before the very recent advances in AI allowed me to generate the images I had had in my head when creating the song. In the past I've used a lot of stock video, largely from https://artgrid.io/. The main benefits have been: high quality - framing, content, colour, resolution multiple 'story' clips - allowing for consistency of character/location videos clips, provided they are relevant, can fill a LOT of time in a music video!! price (unlimited downloads and total freedom of usage) HOWEVER, if given the choice, I much PREFER to use still images. I think photography is far more impactful and the eye is free to examine and absorb an unchanging image. This remains so even with a slow steady zoom/pan of an image. And recent AI can produce exquisite stills of tremendous surprise, beauty, detail and 'imagination'. It'd be nice to delve deeper into the issue of stills vs film if anyone wishes to contribute their thoughts. Greg

Hi. I've just posted a new music video in "Showcase" with more use of AI imagery. All the images of the guy were AI. E.g. <link to previously generate image of a guy> a handsome man in his 30s lost in thought, soft crumpled white shirt with collar, realistic, writing a letter, moody lighting [though these are square because I forgot to use "--ar 16:9" for normal landscape video!! It was early days for me ]

Thanks for the heads-up ... I had assumed the lack of a linked thumbnail was just processing lag. Tried to fix now ... the thumbnail appeared but has now gone, leaving the relevant amount of space!! Tried again ... SEEMS to be OK this time.

Hi They get pretty good results. It’s a quick way for a creative to generate a number of pro quality lyric videos with minimum hassle for a reasonable cost. Really, I can see a big reduction in time invested to get quality results.

Hi This time I thought I would post up something about using AI to generate the full video animation. The methodology is a bit clunky. I would definitely add the audio in Adobe Premiere in preference to the method they use. Still, interesting to see some alternatives.

Hey I found this pretty in-depth walk through of the creation of a music video, on a budget of $200. The video uses AI to create the animations, uses stock footage as seed clips, uses a tool to create the lyric animations and more. I am pretty sure there are tools that could take even more of the man/woman hours out of it. The results are pretty good, though I think there are some great generation tools out there that might have enhanced the results further. Enjoy…

Did you mean that to be an embedded video, a clickable link or plain text (as it is now)?

It is with great relief that I have finally made and released a video to accompany “Agincourt” from “The Flat White Album”, my 30-track opus. It is my longest ‘music video’ to date (11:33) and is almost 3 years ‘late’ due to lack of available imagery, but the very recent change of affairs is explained in the description below the video. It uses 150 AI-generated images (approx 99% of the running time content!).

Yep, About time we started using it!

Hi I am looking for a tool to help me create music videos. Ideally something that might take photo input and then generate video output, where photos are extrapolated using AI guided by my text input. For example: I upload a photo and provide the input text: Use the uploaded photograph as a starting point. Create a music video segment of 4 bars in length for a piece of music of 132 bpm. Music is upbeat, lyrics are about loss and are contemplative in nature. Include at least 2 camera perspectives. The video should be creative and inspired visually by the works of Monet and Degas and the movie A Clockwork Orange. Assume a frame rate of 30fps and an aspect ratio of 16:9 in full HD quality. This video segment is for the first 2 lines of the chorus section, for the following lyrics: ”There’s not much now that we can do For the feelings between us and you” Any recommendations for the tools and how to use them would be greatly appreciated. Cheers John @GregB

Hi. I've just posted this elsewhere (I had forgotten about this Club ... which I started!). But I'll post the videos HERE very soon, along with some details about how I created them. E.g. issues, actual prompts, transferring the images from the web app to my local machine, and two installable apps that are free and quick to batch-convert PNG to JPG as well as batch upscale x4. Greg

Music seems a ludicrously facile expression of the terrible Ukrainian situation, but I was struck that humanity keeps repeating the micro and macro failures of the past. We all thought Russia's aggression in Europe post-WWII had finally run its course after the fall of the Berlin Wall, that there was a blossoming of peace, and a real feeling of optimism was spreading everywhere. Now, Russia has resumed its old ways. China too has increased its belligerence. (North Korea goes without saying!) The music release itself was posted and discussed in "Showcase" (Oct 16, 2022) https://forums.songstuff.com/topic/57145-running-from-red-extended-mix/. Only now, four months later, and feeling a little edgy due to not having any other creative projects simmering, I decided to have a crack at the music video for the track. As mentioned elsewhere, I've always found difficulty in find enough quality/relevant imagery from my handful of limited cleared-to-publish sources to tell any visual story that I imagine. Now I had to find roughly double the quantity for a track which is 8:20 long! However, I learned long ago to accept 'make-do' with what I have. The final visuals are good enough. A few are compelling/emotional. There is no attempt in the story to be self-righteous and preachy. The 'Red' here can also be equally applied to the flags of the USA and its 'running dog' allies, including the UK (my birth) and Australia (my home) who, with only immediate political expediency in mind, have shown they will invade, bomb (indiscriminately killing civilians), bully and coerce governments, and assassinate individuals in other countries, all in the pursuit of power, influence, resources, etc. ... usually under the supposed guise of 'self-defence'. Isn't that exactly the same as what both Russia and China do? If you have any questions about sources, editing, etc., I enjoy discussing the technology and decision making. PS. The original video (which I actually prefer) is below and in which I created a video prologue to give it historical context, i.e. the sounds were NOT part of the original album release. I decided to change its YouTube status to UNLISTED (only those with the link can find and see it) as, back in 2015, I could not be 100% certain that everything was cleared for use. This problem made me become really anal regarding my future documentation of sources, creators, and usage rights. Cheers, Greg

Hi. I've created another music video for which I needed to make some unique decisions. I rejoined the stock video site www.artgrid.io (after a 6-month lapse in subscription) as I had a few videos to do, this being the third. I love this site for its cost (the 'Basic' is good enough for me), the quality, simple T&Cs, and the multiple clips available for every 'story'. There were approx 150 clips in this 'story' of a Blues Band. 70 were colour and 80 B&W (of which some were simply greyscale duplicates). The biggest problem, and one I've encountered before but skirted around, is that of people in clips either PLAYING an instrument or MOVING rhythmically. Either of these can clash horribly with your own track. Instruments in particular may be absent on the track, or totally wrong (e.g. an acoustic shown instead of the electric playing), or the 'playing' is faked and not believable, or the performance detail shown is nothing like what's in the track (e.g. strumming vs picking). But in this case, because the clips I found were so good MOOD-wise for this video, I was determined to make them work. HOPEFULLY this has been achieved by: slowing the clips to a dream-like state that is obviously NOT playing to or synchronised with the soundtrack showing as little performance DETAIL as possible ... e.g. the pianists hands are never shown, the guitarist's rarely so. avoiding the drummer completely using extra-arty shots of the singer. There's no female voice on the track, but she is the quintessential core of the band in the visual story. there is no saxophonist on the track, and I've kept his clips to a minimum I DID focus on ONE clip (in colour, of them around the piano) to get her clicking her fingers in time with the track. I think this helps re-focus the eyes/ears. the above criteria meant that only about 25 clips (out of the 150) were suitable but, fortunately, having to slow their speed really helped pad out the running time. The Artgrid clips are presented and downloaded in random order and have random names (grrr!). So I then have to do the usual grouping for 'scenes' and, continuity (e.g. opening shots ... the first clip shows her putting IN earrings, the second shot has them fully displayed ... so these clips must be used in THAT order!), and also 'story' arc (e.g. preparing to play, and THEN playing). I also grouped clips according to being B&W or colour ... it would look crazy to flipflop between the two colour schemes. Once clips were placed on the timeline and invariably trimmed down to allow changes on the beat, I ended up being short of material. I did find clips for a white guy walking with a guitar case ... a great analogy for 'going down to the crossroad' BUT ... the scenery was wintery, and the case looked odd - like a banjo. I eventually stumbled across the end sequence - just four clips suitable out of a 'caravanning' story of 60 clips. The sunset vibe felt good, as did the hippy dancing (Cream's "Crossroads" was in 1968). I used the shortest of these (the hair flick) to end and speed-ramped it down, capturing the last frame as a still to support a long fade-out. Not one of my best/favourite or more innovative music videos, but it turned out 'good enough'. Greg

Hi Logan. I had seen The Swan & The Ram elsewhere on the site. Well done for taking the rare tongue-in-cheek visual approach to music pioneered in Hard Days Night and The Monkees. It really is great fun. Music videos are peppered throughout Songstuff, e.g. Showcase, Performance, Videos & Images, Blogs etc. The Music Video Club is also a 'please watch this' platform but is specifically background/explanation about the creative and technical side of the filmmaking. The concept, planning, shooting, editing, equipment, problems, decisions, etc.. You did provide a little about the location shoot (buried in the middle of your interview video) which was interesting and, by the way, exactly why I don't do location or performance shoots! I look forward to reading something about your video journey. A write-up (mini article) is never wasted effort as it can always be re-used elsewhere, e.g. sent to journalists, or posted on your own online platforms. Cheers, Greg

Hello! My name is Logan Kono and I am the frontman/owner of Kono, a hard rock art collective from Santa Barbara, California. The reason why I say art collective is because I have an amazing team of friends and fellow creatives that help create the music videos and live format for Kono!! I write all the songs and play all the parts of the records, but it wouldn't be what it is without my team, and music videos are a huge part of all that. I am currently recording the second Kono full-length album, while the Kono Klub and I are working on a couple more official music videos to wrap up phase 1 of Kono visually! . T Please check us out if you enjoy hard rock and new original heavy music!! Much more from Kono coming soon!!! Thank You

It's hard to believe that a multicam production is SO easy to do and leaves you totally in control. It is offered by all the front runners - Premiere, Resolve and Final Cut. It is especially easy to contemplate when we all have multiple high quality phones, either from friends/family, or the devices we've recently replaced and are simply lying unused in a drawer! Full HD, 1920x1080, has been standard in phones for close to 10 years! Here's an example done in Premiere (before I switched to Resolve). It used: My Samsung phone for the close-up on me (unfortunately the lens wasn't clean!) Martin's iPhone for his close-up My Canon 70D DSLR on a tripod for the wide shot A Zoom stereo field recorder on the corner of the table between us (camera audio is still poor by comparison). You might prefer to record directly into your DAW via one or more mics. NOTE - ensure all cameras are set for the same frame rate. 25fps works well as this is the YouTube standard, but anything is fine as long as common to all. I used the traditional hand-clap to provide a sync point all four audio stream being recorded ... the spike in the waveform is easy to identify visually in the editor and line up manually if needs be. This is essential when there may be several seconds between each device being placed in record-mode. But note that all editing software has the ability to sync the videos if they contain the same audio content, and it seems to do a good job. The hand-clap is cheap insurance! Once the 3 video clips were sync'd, I MUTED the camera audios and just used the file from my Zoom device. Depending on your taste, you may wish to process this audio (e.g. some EQ/Compression) in your DAW before importing into the video editor. (My example is the raw recording.) When selecting 'Multicam' in the editor, you are shown a window with (in my case) the 3 different camera views. Whilst playing in real time, or scrubbing through to specific points, you simply click the camera angle you want at that moment. The editor creates the new timeline with the clip/cut in place. You can go back at any time and edit those cut points, e.g. to change the chosen camera angle. More often, a cut point needs a dynamic roll to extend one clip and shorten the other. If I performed more, I would take this approach this all the time. It makes for a far more dynamic product. It also helps to avoid visual problems (e.g. an unwanted excessive flare in glasses, fiddling with a cheat sheet, or even a dog walking past) simply by switching the view.

Strange - I only saw this when visiting to start my own new post ... i.e. I wasn't notified of your new post!! The dolls in the "Fairytale" music video (https://www.youtube.com/watch?v=qvCkyuT6Dgc) were posed on and against a dark blue cloth and, because they were stills, the background was removed in Photoshop before being imported as PNGs into the video editor. I've only ever done video 'solid colour' background removal as concept demo-examples for friends and clients. The main things I discovered to save HEAPS of time are: the background has no sags/folds. Sheets CAN work but need to be pulled taught. the background has its own even lighting. If only using lights in front of the subject, then shadows will be cast on the background (requiring extra work to set the masking colour range, or require manual masking) be careful of colour reflections from the background onto the subject's hair/clothes, causing odd changes in subject outline. John, from memory, you use Premiere? If consistent background colour is an issue, Resolve has a very quick/easy way of selecting area masks. Find the most problematic frame, draw a line through an area (and expand with other lines if not complete), and then tracking the mask back and forward. Unsure if this is available in the free version.

I’ve been getting ready to use the black and green screens for the Songstuff Vocal Coaching Academy work that I’ve been working on. Somethings to keep in mind Lighting - if there’s a gradient to the green because of shadows, it’s gonna affect the end result. Good lighting is key. You could make do with at least two to three spot lights with portable stands. Creases - having a lot of folds and creases on the green screen an affect the end product too. So keeping it well ironed and folded when not in use can be helpful.

Hey Gang I'm soon to be starting shooting for some videos for my songs. I expect to use a bit of green screen. Do you have any tips you can give me for green screen productions? Tips for shooting, editing & post production are very welcome. Cheers John

Congratulations, Greg!