---
title: "Elevation localization and head-related transfer function analysis at low frequencies"
authors: ["V. Ralph Algazi", "Carlos Avendano", "Richard O. Duda"]
year: 2001
venue: "The Journal of the Acoustical Society of America"
arxiv: null
doi: "10.1121/1.1349185"
source: "https://doi.org/10.1121/1.1349185"
topics: [hrtf, personalization, anthropometry]
seed_rank: 1720
seed_batch: "audio-harvest-2026-09-09"
reviewed: "2026-09-09"
pool: "realtime"
relevance_score: 9
lineage: hrtf-spatial-audio
cites:
  - title: "The CIPIC HRTF database"
    url: "https://doi.org/10.1109/aspaa.2001.969552"
    year: 2001
    arxiv: null
    doi: "10.1109/aspaa.2001.969552"
see:
  - "1684-the-cipic-hrtf-database"
  - "1721-approximating-the-head-related-transfer-function-using-simple-geometric"
---

# Elevation localization and head-related transfer function analysis at low frequencies

## One-sentence takeaway

Low-frequency elevation cues are mostly torso/head-shadow, not pinna notches — Algazi/Avendano/Duda 2001 JASA.

## Why it matters here

Cheap LF HRTF personalization can use head/torso size (CIPIC anthropometry) without a pinna scan. This paper justifies snowman/ellipsoid models (1721) at low frequency and why generic KEMAR bass isn't crazy.

## Key ideas

- **Elevation at LF.** Torso reflection / head diffraction dominate.
- **JASA 2001.** Same CIPIC authors, companion to the database paper.
- **Personalization hint.** Anthropometry → LF HRTF.
- **~153 cites.** Harvest personalization/anthropometry.

## Caveats

High-frequency elevation is still pinna and *does* need individual or SADIE-class data. Not a full HRTF model. Not near-field.

## Links

- DOI: https://doi.org/10.1121/1.1349185