Hakim “Nour” (Levantine) vs Piper “kareem” (Jordanian)
The same six words from both. Nothing has been changed in the app yet.
What to listen for
Palestinian drops the ق to a glottal stop. If Nour says qahwa rather than
’ahwe, she is reading Modern Standard and the dialect claim does not hold —
which is the same problem Piper has, and the reason 40 words carry a
“how it’s said here” note.
WordHakim · NourPiper · kareem
قَهْوَة
want
ق initial
’ahweق initial
Nour · 3.84s
kareem
سُوق
want
ق final
soo’ق final
Nour · 1.75s
kareem
وَقِت
want
ق medial
wa’itق medial
Nour · 1.13s
kareem
ثُوم
want
ث → t
toomث → t
Nour · 1.75s
kareem
أُسْتَاذ
want
ذ → z
ustaazذ → z
Nour · 2.23s
kareem
مَبْسُوط
baseline
nothing shifts
nothing shifts
Nour · 1.44s
kareem
One thing I noticed without being able to hear it.
قَهْوَة came back at 3.84 seconds — 2.7× longer than مَبْسُوط at 1.44s, for a word
of similar length. Piper’s is about 1s. That is either a long silence, a pause
inserted mid-word, or the letters being read out. Worth checking on that clip
specifically.
Quota. 86 of 10,000 characters used so far. All 186 words plus their 186 example sentences come to 5,337 characters, so a full pass fits inside this month with 4,663 to spare. Overage is billed rather than blocked, so a second pass with a different voice is possible but not free.
Nour is the only Levantine voice on the account — 14 Arabic voices, the rest are MSA, Gulf, Egyptian, Iraqi, Maghrebi, Sudanese and Saudi. Her register is tagged “upbeat for social-first”, which is not obviously right for a learning app, but dialect beats register here.
Quota. 86 of 10,000 characters used so far. All 186 words plus their 186 example sentences come to 5,337 characters, so a full pass fits inside this month with 4,663 to spare. Overage is billed rather than blocked, so a second pass with a different voice is possible but not free.
Nour is the only Levantine voice on the account — 14 Arabic voices, the rest are MSA, Gulf, Egyptian, Iraqi, Maghrebi, Sudanese and Saudi. Her register is tagged “upbeat for social-first”, which is not obviously right for a learning app, but dialect beats register here.