Battle of the AI Tools: Splash vs. SoundRaw in Music Generation

Commercials, especially those promoting fragrances, rely heavily on music to captivate audiences - it's the hook that draws viewers into the spun story. In the world of fragrance marketing, music isn't mere background noise; it's a guide, setting the mood and inviting us to feel the sophistication and allure each fragrance offers.

In my past AI commercial projects, I've used the music generative tool from SoundRaw, which I love for its ability to tailor the music's length and tempo perfectly to match the video. Recently, I found a new music generative tool from Splash called Splash Pro. It transforms text prompts into music, powered by a library of over 100,000 loops covering different instruments and genres, offering a versatile base for creativity.¹ Its innovative features not only open up new possibilities for creative projects but also enhance the art of crafting compelling commercial stories.

To test this tool out for my upcoming project, I decided to create a song for a cologne commercial. I started with the prompt "A sophisticated jazz-inspired soundtrack with a gradual build, creating an elegant and refined ambiance." After clicking the "Improve" button, it refined the prompt to "A classy Jazz track with a gradual crescendo, creating an atmosphere of elegance and refinement." There is the option to specify the desired BPM, key, and mode, but I opted to give the tool full control. 

Clicking "Create Song" generated four snippets. 

I liked sample 3 but felt it needed some adjustments. The tool allowed me to generate more variations until I found the perfect fit. 

Once I settled on a sample from the new variations, it automatically created a 30-second song based on my chosen sample.

Afterward, I could edit the song by choosing a different duration (15 seconds, 30 seconds, and 1 minute are the free options), change the structure (standard, minimalist, looping) if paying for the tool, or add vocals by inputting text. Opting for a 1-minute song, the tool automatically generated the extended version.

Although the vocals option is currently in its beta phase, I opted to give it a try. The result was a fast-paced recording on top of the beginning of the song, indicating that there are still some kinks to work out. For voice generation, I plan to stick with ElevenLabs.

While Splash Pro is undeniably impressive with its features and evident potential, it does fall short in terms of editing options. This limitation poses a challenge in realizing one's vision for a video precisely. Despite the ability to construct visuals and a narrative around the song, this approach feels somewhat like working backward. Additionally, the exported songs' quality is unfortunately lower than their performance on the platform. While I eagerly anticipate the evolution of this tool, for my upcoming project, I lean toward sticking with SoundRaw, where I have a more seamless experience in crafting the desired audio for my commercials.

Sources:

¹ https://support.splashmusic.com/hc/en-us/articles/22950699604633