How to Convert PDF to Audio: Step-by-Step Guide
Convert any PDF to listenable audio in 3 steps with SpeechGeneration AI. This tutorial covers standard PDFs, scanned PDF troubleshooting, long document strategies, and a cost-per-textbook calculator.
Note: SG.ai converts TEXT to audio. You paste the text content from your PDF. For the product overview, see PDF to Audio. For bulk conversion (50+ PDFs), see Batch Processing.
Quick answer: (1) Copy text from your PDF, (2) paste into SG.ai and select a voice, (3) generate and download MP3. Takes under 5 minutes per chapter. Free tier covers ~1 chapter (10K chars). Starter plan ($5/mo) covers ~6-10 chapters.
Important: Scanned PDFs (images of text) require OCR first. See troubleshooting below.
Contents
Before You Start
You need two things: your PDF open on screen and a SpeechGeneration AI account (free, no credit card).
Standard Digital PDF
You can select and copy text. Works immediately — go to Step 1.
Scanned PDF
You CANNOT select text (it's an image). Needs OCR first — see troubleshooting.
Encrypted/DRM PDF
Copy-protected. Cannot extract text. No workaround available.
Quick test: Try to select and copy one sentence from your PDF. If you can → proceed. If you can't → it's scanned or encrypted.
Step 1 — Extract Text from Your PDF
Open your PDF in any viewer (browser, Adobe, Preview). Select the text you want to convert — typically one chapter or section at a time.
For a single chapter: Scroll to the chapter start. Click at the beginning, hold Shift, click at the chapter end. Or use Ctrl+A (select all) if your PDF is a single chapter. Copy with Ctrl+C.
For multi-column PDFs: Select column-by-column, not across columns. Left column first (copy), then right column. This preserves reading order.
Pro tip: Open the PDF in Chrome's built-in PDF viewer for the cleanest copy-paste experience. Chrome handles most PDF text selection better than dedicated PDF readers.
Skip these when copying: Page numbers, running headers/footers, figure captions (unless you want them read), and table of contents.
Step 2 — Paste into SG.ai + Select Voice
Go to speechgeneration.ai. Paste your copied text (Ctrl+V) into the text input area.
Select a voice: Preview at least 3 voices with a representative sentence from your text. For textbooks, a clear, measured voice works best. For fiction, try voices with more expressiveness. For non-English PDFs, select a voice in the matching language.
Select quality tier:
- •Economy (0.1×): Quick scan — hear the content, evaluate if it's worth deeper study. Lowest cost.
- •Studio (1×): Study sessions — clear, professional narration for focused listening. Recommended default.
- •Studio+ (2×): Engaging narration with emotion tags — for fiction or content where engagement matters.
Adjust speed if needed: 0.85× for accessibility, 1× default, 1.25× for review. See our speed workflow guide for recommendations by use case.
Step 3 — Generate + Download MP3
Click Generate. Wait 30-60 seconds (varies by text length). The audio preview plays automatically.
Quick QA check: Listen to the first 10-15 seconds. Check that proper nouns are pronounced correctly. If a name sounds wrong, add a phonetic hint in your text: "Aethon (say: EE-thon) walked into the room" — then regenerate. The phonetic hint won't appear in the audio; it guides pronunciation.
Download: Click the download button. The MP3 is saved to your device. Name it descriptively: psychology101_ch03.mp3
Listen: Transfer to your phone via cloud storage, USB, or email. Play in any MP3 app. For podcast-style listening, import into Pocket Casts or Overcast as a local file.
Troubleshooting: Scanned PDFs
If your PDF is a scan (photograph of pages), you can't select or copy text. You need OCR (Optical Character Recognition) to extract text from the images first.
Option 1: Google Docs (Free, Easiest)
Upload your PDF to Google Drive. Right-click → Open with Google Docs. Google automatically runs OCR. The text is now selectable. Copy from Google Docs into SG.ai. Accuracy: 90-95% for clean scans.
Option 2: Adobe Acrobat
Open in Adobe Acrobat (paid). Click Edit PDF — OCR runs automatically. Text becomes selectable. Copy into SG.ai. Accuracy: 92-98% (best for complex layouts).
Option 3: Tesseract (Free, Technical)
Open-source OCR tool. Requires command-line usage. Best for batch OCR of many scanned PDFs. Accuracy: 85-95% depending on scan quality.
Always proofread OCR output before generating audio. OCR errors (especially on proper nouns and numbers) become pronunciation errors in the audio. A 2-minute proofread saves a wasted generation. For a deeper accuracy analysis, see Is PDF to Audio Accurate?
Troubleshooting: Long Documents
SG.ai has a 5,000 character limit per generation (~700-800 words). A full textbook chapter is typically 15,000-30,000 characters. Strategy:
- •Break at natural boundaries: Section headings, paragraph breaks, or topic transitions. Don't cut mid-sentence.
- •Name files sequentially:
ch03_part1.mp3,ch03_part2.mp3, etc. - •Use same voice + tier + speed across all parts for consistency.
- •For 50+ PDFs: See our batch processing guide for automation strategies.
Cost Per Textbook
| What | Characters | Economy | Studio | Studio+ |
|---|---|---|---|---|
| Single chapter | ~25K | ~$0.17 | ~$1.70 | ~$3.40 |
| Full textbook (80K words) | ~400K | ~$2.70 | ~$27 | ~$54 |
| Semester (5 textbooks) | ~2M | ~$13.50 | ~$135 | ~$270 |
Free tier (10K chars) covers ~40% of one chapter. Starter plan ($5/mo, 100K chars) covers ~4 chapters at Studio. For full pricing details: Pricing comparison.
Frequently Asked Questions
Can I convert a scanned PDF to audio?
Not directly — scanned PDFs contain images of text, not actual text. You need OCR (Optical Character Recognition) first. Upload to Google Docs (automatic OCR), use Adobe Acrobat's Edit PDF function, or run Tesseract (free, open source). After OCR, copy the extracted text into SG.ai.
How long does it take to convert one chapter?
Under 5 minutes total: ~1 minute to copy text from PDF, ~30 seconds to paste and select voice in SG.ai, ~30-60 seconds for generation. The bottleneck is copy-pasting, not generation.
What's the character limit per generation?
5,000 characters per generation. A typical textbook chapter is 15,000-30,000 characters, so you'll paste in 3-6 segments per chapter. Name your downloads sequentially: chapter01_part1.mp3, chapter01_part2.mp3.
Can I convert the entire textbook at once?
Not in one paste — the 5,000 character limit requires chapter-by-chapter conversion. For converting 50+ PDFs at scale, see our batch processing workflow which covers automation strategies.
How much does it cost to convert a full textbook?
An 80,000-word textbook (~400K characters): ~$2.70 on Economy tier, ~$27 on Studio tier, ~$54 on Studio+. Compare to $3,000-5,000 for human narration. The Starter plan ($5/mo, 100K chars) covers roughly 2 chapters at Studio quality.
Does it handle footnotes and references?
Footnotes and references will be read aloud if included in your copied text. For academic papers, consider skipping the references section — paste only abstract + body text. Footnote numbers will be read as 'one,' 'two,' etc.
Can I convert PDFs in other languages?
Yes. SG.ai supports 70+ languages on Studio+ tier. Select the matching language voice before generating. For multilingual PDFs, convert each language section separately with the appropriate voice.
Where can I listen to the downloaded MP3?
Any device that plays MP3: your phone's default music app (Apple Music, Samsung Music), Spotify (import local files), VLC, podcast apps that support local files (Pocket Casts), or your computer's audio player. Transfer via USB, cloud storage, or email.