DV to BMP Converter

Grabbing a Still From DV Tape: What This Tutorial Covers

Pulling a frame out of captured DV footage and saving it as a Windows bitmap sounds like the simplest conversion on the site. It has two traps in it, and both of them are invisible until you look at the output closely.

The first is geometry: DV does not store square pixels, so a widescreen frame comes out horizontally squashed unless you tell it otherwise. The second is that none of the three BMP quality presets writes a 24-bit bitmap — the one most people assume they are getting. This page walks through both, with the numbers we measured.

How to Convert DV to BMP

1. Upload Your DV File: Drag the .dv onto the drop zone or click "Add Files". DV is a fixed-rate format at roughly 3.6 MB per second of footage, so trim before upload if the capture is long.
2. Choose the Frame Under Frame Selection: The default is the first frame. Pick Specific Frame and give a time in seconds for a later moment, or Multiple Screenshots with a capture rate to get a run of stills in one pass.
3. Pick a Quality Preset: The dropdown here offers High, Medium and Low, and they are three genuinely different bitmap encodings rather than three points on one scale — see the table below.
4. Convert and Download: Click Convert. Files upload over an encrypted connection, are processed on our servers, and are deleted automatically after a few hours. No sign-up, no watermark.

Walk-through: What Each BMP Preset Actually Writes

Behind the three friendly names sit three completely different encoders. We ran a real 720 × 480 DV frame through each and read the BMP headers back:

The bottom row is the negative control and it is the important one. The source frame contained 18,364 distinct colours. High keeps only 3,974 of them, because RGB555 gives each channel five bits — 32 levels instead of 256. It is uncompressed and therefore large, but it is not full colour. Medium and Low quantise to a palette and run-length encode it, which is why they are so much smaller.

So the choice is not "how good do you want it" so much as "which bitmap flavour do you need":

• High when you want an uncompressed bitmap that any BMP reader in existence will open, and 32 levels per channel is enough.
• Medium when you want a compact palettised BMP with reasonable colour — 254 colours with Floyd–Steinberg dithering, RLE compressed.
• Low when file size dominates and 86 colours will do.
• None of them if you actually wanted a full-colour 24-bit still. DV to PNG gives you that in a smaller file than any of these, provided you set its Quality Preset to Highest — the other presets quantise to a palette.

Walk-through: The Anamorphic Squash

DV stores 16:9 widescreen on a 4:3 pixel grid using non-square pixels. Our test file probes as 720 × 480 with sample_aspect_ratio=32:27 and display_aspect_ratio=16:9: the raster is narrow and a player stretches it back out at display time.

A still image has no equivalent metadata that viewers reliably honour, and the extracted frame we measured came out at 720 × 480 — the storage dimensions, not the display dimensions. On 16:9 footage the picture therefore needs stretching horizontally by about 19% before it looks right, and nothing in a bitmap will do that for you.

The fix lives in the Image resolution group. Set Width × Height explicitly:

If your footage is 4:3 to begin with, the 720 × 480 raster is already close enough that most people never notice, and Keep original is fine.

Common Errors and How to Fix Them

• "Everyone in the frame is too thin" — anamorphic squash. Set Width × Height to 854 × 480 for 16:9 NTSC.
• "The still has horizontal comb lines through anything moving" — SD DV is interlaced and there is no deinterlacing anywhere in this pipeline. The frame holds two fields captured milliseconds apart. We confirmed this with a test whose fields were solid blue and solid red: the extracted still alternates blue, blue, red, red down the scan lines. Pick a static moment with Frame Selection, or deinterlace before uploading.
• "The colours look banded or posterised" — you are seeing the preset. High gives 32 levels per channel; Medium gives a 254-colour palette; Low gives 86. Switch to PNG if the banding is unacceptable.
• "The BMP is bigger than the whole video clip" — uncompressed bitmaps are like that. A 720 × 480 High preset file is 691,254 bytes for a single frame. Use Medium or Low, or a compressed image format.
• "I got dozens of files instead of one" — Frame Selection is set to Multiple Screenshots. Switch it back, or set the capture rate to a longer interval.

Frequently Asked Questions

Why is the High preset only 16-bit instead of 24-bit?

Because that is what the preset is defined to write: the header reports 16 bits per pixel with the BI_RGB compression field, which by the BMP specification means 5 bits each for red, green and blue. The arithmetic confirms it — 720 × 480 × 2 bytes plus a 54-byte header is exactly the 691,254 bytes we measured. If you need 24-bit truecolour, BMP on this page will not give it to you; PNG will.

Which preset preserves the most colour?

High, by a wide margin — 3,974 distinct colours survived from a source frame containing 18,364, against 254 for Medium and 86 for Low. It is also by far the largest file, because it is uncompressed while the other two are run-length encoded.

How do I grab a frame from the middle of the clip?

Set Frame Selection to Specific Frame and type the time in seconds into the Time field. If you would rather pick from several candidates, choose Multiple Screenshots and a capture rate — one frame every second, every five seconds, and so on — then keep the one you want.

Will the still be 720 × 480 or 854 × 480?

720 × 480 unless you ask for something else, because that is what is physically stored. NTSC DV is always a 720 × 480 raster; whether it represents 4:3 or 16:9 is metadata, and a bitmap has nowhere to put that metadata. Set Width × Height to 854 × 480 to bake widescreen geometry into the pixels.

Does the extracted frame lose quality before it reaches the BMP encoder?

No. The frame is pulled out of the DV stream and written as a lossless PNG first, then handed to the BMP encoder. Everything you lose is lost in the preset you chose, not in the extraction. That is also why the Quality Preset is the only lever that changes image fidelity here.

Why does my BMP have comb artefacts when the video looked fine?

Because your player deinterlaced the video on the fly and this conversion does not. There is no deinterlacing filter in the pipeline at all, so both fields land in the same still. Any frame containing motion will show the artefact; a frame from a locked-off shot will not.

Can I extract frames from an entire tape capture at once?

Yes — Frame Selection's Multiple Screenshots mode takes a capture rate from one frame every tenth of a second up to one every ten seconds and returns the whole set. Be aware of the size implications: at the High preset each 720 × 480 frame is roughly 691 KB, so a long capture at a fast rate produces a great deal of data.

Is BMP a sensible archival format for video stills?

Not really. It is uncompressed or crudely run-length encoded, has no metadata to speak of, and the presets here all reduce colour depth. It earns its place when a piece of software specifically demands a .bmp — some industrial, medical and older Windows applications do. For anything else, DV to PNG is smaller and keeps full colour at its Highest preset, and DV to JPG is smaller still.

Does resizing the still change which preset I should pick?

It changes the size arithmetic but not the colour trade-off. High stays 16-bit regardless of dimensions, so if you scale a frame down to a thumbnail the uncompressed penalty largely disappears and High becomes the obvious choice. At full frame size the 691 KB versus 110 KB gap between High and Medium is what pushes people toward the palettised presets.