Sound Checker

×

Quick start

Just what you need to know — a 30-second refresher.

1. Paste your text in panel 1 — a word list, or a whole reading passage. The tool pulls out the words and breaks each into its sounds.
2. In panel 2, mark the sounds you’ve taught. Use the taught through group row to follow a scope & sequence, or click individual chips. Everything starts selected.
3. Read the report in panel 3. For each sound it shows three things: ⚠ appears but you haven’t taught it (the one to double-check), ✓ taught and appears, and ◯ taught but missing from this text.
4. Click any sound in the report to see the words that contain it.

Leave every sound selected and the report is just a plain inventory of what’s in your text. Your choices are saved on this computer.

×

How the Sound Checker works

This tool takes a word list or a reading and breaks every word into its sounds — the same way Blending Practice and the Word List Builder do — then checks each sound against the ones you say you’ve taught. It answers two questions at a glance: did anything I haven’t taught yet slip into this text? and are the sounds I’ve taught actually showing up?

1 · Your text. Paste a list of words (one per line, or separated by spaces or commas) or a full reading passage — punctuation and numbers are ignored, so prose works fine. Repeated words are counted once. Inappropriate words are screened out. The box holds about 1,700 words — three or four pages, more than any one lesson’s reading — and a longer paste is cut off at that point.

2 · Sounds you’ve taught. Every sound starts selected. Switch off the ones you haven’t taught, or use the taught through group row to select everything up through a point in a common structured-phonics order in one click; the add / remove one group row flips a single group on top of your current choices. Fine-tune anything with the chips below, or use Select all / Clear all. Not sure which sounds a numbered group holds? Open What’s in each group? for the full list — it’s a common teaching order, and your program may number or sequence the sounds differently.

What counts as a sound. A digraph is two letters for one sound (sh, ch, ck, ng); a trigraph is three (igh, tch, dge); doubles are the floss letters (ll, ss, ff, zz). The tool treats each of these as one sound, so catch is c · a · tch. Other doubled letters (rabbit, happy) are just the single consonant sound, so they count as b and p.

Magic e. cake is c · a_e · k: the split vowel a_e is one sound, and the silent e belongs to it, so a passage full of made and like shows up under Magic e, not as short a and e. Only the strict pattern counts — one vowel, one consonant, then e — so house and paste are split the ordinary way, and a vowel before r (care, here) stays r-controlled. Exceptions such as have and come still count as magic e; treat those as tricky words yourself.

Two-sound teams. oo, ow, and ea each spell two common sounds, so each has two chips: oo (moon) / oo (book), ow (cow) / ow (snow), ea (bead) / ea (bread). Spelling alone can’t tell them apart, so the tool recognises the second sound from a short word list (book, snow, bread, head … plus their common endings) and gives everything else the first sound. A rarer word with the second sound may land on the first chip — click the chip to see its words if the count looks off.

Glued / welded sounds. In king, bank, and ball the vowel is “glued” to what follows and taught as one chunk. With Weld glued sounds on (the default) each counts as one sound and appears in its own Glued / welded group. The ng/nk families always weld; all/am/an weld only when a consonant follows, so name and shallow stay unwelded. Turn the toggle off to break them into single letters instead.

Blends. A blend is two or three consonants you still hear separately (bl, str at the start; nd, mp at the end). Blends aren’t usually taught as separate items — you blend sounds you already know — so the report lists each blend it finds and marks it ✓ decodable when every letter in it is taught, or ⚠ when one of its sounds isn’t taught yet. Blends are detected where consonants cluster around the edges of a word; an unusual cluster in the middle of a long word may not be flagged.

3 · The report. A summary up top calls out anything that appears but isn’t in your taught set — the sounds worth a second look in a lesson — and anything you’ve taught but that’s missing from this text. Below that, each sound family is broken into:

• ⚠ Appears, not yet taught — in your text but switched off in panel 2.
• ✓ Taught & appears — in your text and already taught.
• ◯ Taught, but absent — taught, but no word here uses it.

Click any sound to list the words that contain it. The number on a chip is how many different words use that sound. A few less-common spellings (for example ph, au, ei) sit outside the numbered groups — reach them with the chips or Select all.

Everything is remembered on this computer. Use Clear saved settings at the bottom to wipe your choices and start fresh. Copy settings (next to it) copies your taught-sound choices as a block of text you can paste into a document — handy for keeping a record per student, or for moving a set-up to another computer. Paste settings reads that text back in and applies it. The pasted reading text itself is not included.