A piece of marketing data that may change your thoughts about the video is that I used the faces to create a "that one looks like me" vibe. This means that if someone sees something that strikes them as familiar, they will be more likely to continue viewing it.

Interesting!

I see what you're going for. In that case, though, it might be better to not have the name of each character appear. The context makes it obvious that these people are the friends the narrator is talking about. Their exact names won't mean anything to someone who hasn't read the book, but showing the names might make people less likely to relate one of the images to themselves.

I took a look at the alternate cut with black in it and somewhat shorter duration photo montage. That seems to be an improvement.

Yeah, a series of shots of people who are really good looking but aren't doing anything except having the wind blow their hair doesn't tell us much about the book. The video is definitely somewhat at odds with the audio. Also, that kind of character display is getting really common--because AI is better at that than at action scenes. For instance, there's a trailer on YouTube that explores what a Justice League trailer would have looked like in an earlier era. It's mostly famous actors standing there in superhero costumes, displayed one at a time. (In other words, it's not like any real trailer for a superhero movie would have been, ever.)

The only remotely murdery thing in the visuals is the bloody handprint at the very end. It may be impractical to do a genuine action sequence, but maybe someone lurking in the shadows or something. Or people looking actually worrying instead of looking like they're posing for a photo shoot.

The voice sounds a lot less mechanical than it would have sounded even a few years ago, but honestly, it still needs a little more emotion, particularly at the point where the speaker reveals that his friends are being framed and that he knows whodunit.


The text itself is great. I'd buy the book based on the text, which does make the good sound like a great read. But if I didn't know you, the video would make me want to take a pass.

Bill,

This is good feedback, something I can use to improve the trailer. 

A piece of marketing data that may change your thoughts about the video is that I used the faces to create a "that one looks like me" vibe. This means that if someone sees something that strikes them as familiar, they will be more likely to continue viewing it.

That given, I may add something ominous to the faces or shorten their duration. 

R.C.

Yeah, a series of shots of people who are really good looking but aren't doing anything except having the wind blow their hair doesn't tell us much about the book. The video is definitely somewhat at odds with the audio. Also, that kind of character display is getting really common--because AI is better at that than at action scenes. For instance, there's a trailer on YouTube that explores what a Justice League trailer would have looked like in an earlier era. It's mostly famous actors standing there in superhero costumes, displayed one at a time. (In other words, it's not like any real trailer for a superhero movie would have been, ever.)

The only remotely murdery thing in the visuals is the bloody handprint at the very end. It may be impractical to do a genuine action sequence, but maybe someone lurking in the shadows or something. Or people looking actually worrying instead of looking like they're posing for a photo shoot.

The voice sounds a lot less mechanical than it would have sounded even a few years ago, but honestly, it still needs a little more emotion, particularly at the point where the speaker reveals that his friends are being framed and that he knows whodunit.


The text itself is great. I'd buy the book based on the text, which does make the good sound like a great read. But if I didn't know you, the video would make me want to take a pass.

Also, if anyone is interested, I created a process document with the tools I used while learning the tools and process.  I can reprint it here if anyone is interested.

R.C.

Is the AI the voice?

It's a fun video.  Two things...without the sound on, it doesn't make much sense.  I prefer words over just names, but with the CC on it would probably be fine.

And the second is about consistency. If you show five people and one woman is African-American, then one woman should probably be African-American in the shot of the five friends.  I know that it's very difficult to find good images for making book trailers, but I would try something else to make that work.

This is one reason I quit trying to do promos for one of our series.  It focuses on three women (one is Native-American) and four generations...starting in Edwardian times.  Almost impossible to find any images, so that might be one reason I eventually try AI.