The Human Craft of AI Voice Recording


Artificial Intelligence software has recently become a hot phenomenon.

AI Voice Generation in particular has shown to be the tool that best reflects a human plus technology synergy when it comes to creating engaging content. The software acts as a helpful asset, but it is not as easy as clicking a single button. The best results require human talent to perfect voice flow in each render. By no means is this a replacement for VO artists’ work, but it can act as a strong resource when in a tight spot.

An artificial intelligence voice over application can record script lines in seconds, perfect for any last minute projects or changes that need attention quickly. With a full roster of varying voices to use, finding the specific result needed for a project is simple.

WellSaid Labs, TVB’s AI generated voice program of choice, has a number of quality of life improvements to counter any common obstacles and make the program more accessible. One such quality of life originated with the challenge of pronunciation, a common difficulty with AI voice over. You can save words via the software’s Pronunciation tab with their Phonetic Spelling so future AI generated voice over sessions deliver better results, faster. WellSaid also allows users to render their generated voices by sentence, paragraph, or all in one take (depending on the word count). This, along with the customizable project arrangement, is just another way to keep your AI voices organized. Lastly, the application has multiple video tutorials, narrated by their very own roster of voices, to guide you through the program’s features.

Having said that, it's important to note that the application doesn't always deliver the perfect line the first time around. Several words or phrases may not generate naturally. Rest assured, there are several workarounds so that you can obtain the perfect take. Commas and ellipses are a huge help in creating the pace of the voice over, or to give you enough time to edit it further. Punctuation, like hyphens, are important in making a more natural sounding phrase. Instead of awkward gaps between two different words, you can add an extra layer of ‘flow’ (correct word emphasis) with just a single hyphen. This may make the text you record look odd at times, but these simple additions can help create the right sounding VO for the project at hand.

While AI generated art is being put up for debate, AI voice over has avoided these controversies by addressing the key problem. Those that have volunteered to enlist their voice to AI programs, such as WellSaid Labs, will gain royalties for every use, similar to paying a voice artist.

However, that doesn’t mean we should leave our voice artists behind. Projects call for a certain sound or feel that can not be customized through AI generated programs easily. There may be tricks to try and surpass these obstacles, but voice artists can deliver this naturally and with no hassle. Being major advocates for hiring talent, we at The Visual Brand have found a balance between using talent and software.

Randy Herbertson

The Visual Brand (TVB) is a Metro New York based brand innovation studio, the second generation of a successful NYC based studio founded by branding veteran Randy Herbertson. TVB works with leading and emerging local, national and international brands and companies in well-established practice areas including insight development and brand and messaging foundation, and full service design from packaging, motion design, industrial and environmental design to print, video/tv and digital. Grown in the digital era, TVB leverages and builds on leading edge technology across its practice areas. TVB has a multinational presence and native bi-lingual capabilities with a close partnership in Latin America.

https://thevisualbrand.com
Previous
Previous

Creating your brand's communication code: The shorthand of icons, the choice of the right vocabulary