Narrator voice options
Tap play on any card — only one plays at a time. Tell me the number / name.
After the boss’s note
דוד — הקול הכי עמוק וטוב, אבל מדי גרנדיוזי ולא טבעי. סטיבן — הכי ריאליסטי, אבל קצת קטן ומשעמם.
David: deepest, best voice, but too grandiose and not natural. Steven: the most realistic, but a bit small and boring. Below: what I changed, each next to the old version.
Best so far: R2·06 — David, calm and steady. It keeps David’s depth (7/10, same as before), is more natural (4 → 6) and less grandiose (6 → 4). Steven can’t get that depth from settings alone — pitch-shifting him sounds processed.
הכי טוב עד כה: דוד “רגוע ויציב”. שומר על העומק של דוד, טבעי יותר (מ-4 ל-6) וגרנדיוזי פחות (מ-6 ל-4).
David before → after
BEFORE
nowDavid — current
the current narrator: deepest, best voice — but too grandiose, not natural
AFTER
R2·06David — calm and steady★ best so far
same voice, steadier delivery: stability .85 (was .65), style 0 (was .15), speed 1.0 (was .93)
depth 7 → 7natural 4 → 6grandiose 6 → 4engaging 6 → 5
AI listener scores 1–10, same three lines, same loudness. Green = better (for “grandiose”, lower is better).
Also tried
R2·01David — toned down
same voice, looser: stability .45, style 0, speed 1.04
R2·02David — with audio tags
directed with tags ([warmly], [softly]…) — came out staged and fragmented
Steven before → after
BEFORE
r1Steven — round 1
the most realistic — but small and a bit boring
AFTER
R2·03Steven — more alive
more expression: stability .35, style .35, a touch slower
depth 4 → 4natural 6 → 6grandiose 3 → 4engaging 4 → 5
AI listener scores 1–10, same three lines, same loudness. Green = better (for “grandiose”, lower is better).
Also tried
R2·05Steven — deeper (processed)
pitch-shifted ~1.3 semitones lower + low-end body — audible processing
R2·04Steven — with audio tags
directed with tags — came out slow and disjointed
Every voice reads the same three lines from the film:
“He wrote to Reb Shaul. He called him Tatti — father. And for himself, he asked for one thing: to become a talmid chacham.”
“Conversion is not a decision you make once. He sold the house. He sold the car. And he applied. The visa was rejected.”
“He had never met a Jew.”
Round 1 — all 16 options in one go (6½ min, each voice introduced)
The current narrator, then options 01–16 in order. A neutral voice announces each one.
Round 1 — first picks (an AI listener's first filter)
07Steven★ AI pick
calm narrator · American
AI listener: natural 9 · warm 9 · grandiose 2 (lower = less announcer) — Near perfect match for the brief: intimate, deeply human, reflective, and completely conversational
10Jon★ AI pick
warm & grounded storyteller · American
AI listener: natural 9 · warm 9 · grandiose 2 (lower = less announcer) — Superb storytelling presence: weathered, deeply sincere, warm, and feels like someone sharing a personal memory
04Roger★ AI pick
laid-back, resonant · American
AI listener: natural 8 · warm 8 · grandiose 3 (lower = less announcer) — Worn, authentic mature tone; carries natural spiritual gravitas without sounding staged
01Nate★ AI pick
natural, warm podcast voice · American, casual
AI listener: natural 8 · warm 8 · grandiose 2 (lower = less announcer) — Very strong conversational feel; sounds genuine, grounded, and completely devoid of announcer pretense
12Archer★ AI pick
conversational · British
AI listener: natural 8 · warm 8 · grandiose 3 (lower = less announcer) — Gentle, thoughtful, and sensitive; offers an introspective European alternative without grandiosity
Round 1 — more options
02Leo
reflective, calm · American
AI listener: natural 5 · warm 5 · grandiose 4 (lower = less announcer) — Too mechanical and detached; does not draw the listener in across a table
03Chris
down-to-earth · American
AI listener: natural 6 · warm 6 · grandiose 5 (lower = less announcer) — Lacks spiritual weight and personal vulnerability
05George
warm storyteller · British
AI listener: natural 5 · warm 6 · grandiose 8 (lower = less announcer) — Too dramatic and performative for an intimate personal film
06Eric
smooth, trustworthy · American
AI listener: natural 6 · warm 7 · grandiose 5 (lower = less announcer) — Pleasant timbre, but too radio-polished
08Marcus
calm narrator · American
AI listener: natural 6 · warm 5 · grandiose 3 (lower = less announcer) — Crosses the line from unhurried into lethargic
09Michael
casual storyteller · British
AI listener: natural 7 · warm 8 · grandiose 5 (lower = less announcer) — Warm and sympathetic British voice, though slightly formal
11Roland
smooth storyteller · American
AI listener: natural 5 · warm 5 · grandiose 6 (lower = less announcer) — Too formal and stagey; lacks quiet interiority
13Christopher
gentle, trustworthy · British
AI listener: natural 5 · warm 6 · grandiose 7 (lower = less announcer) — Repeats the high-grandeur problem of the current reference voice
14Mark
natural conversation · American, young
AI listener: natural 7 · warm 6 · grandiose 4 (lower = less announcer) — Competent but lacks emotional vulnerability
15Rachel
calm storyteller · female, British
AI listener: natural 7 · warm 8 · grandiose 4 (lower = less announcer) — Warm, poised, and clear, but leans a bit high on formal presentation
16Jessa
easygoing · female, American
AI listener: natural 6 · warm 6 · grandiose 3 (lower = less announcer) — Too light and upbeat; does not convey the gravity of conversion and sacrifice