ERF to TS Converter

ERF to TS: A Rangefinder Raw Inside a Broadcast Stream

.erf is one of the narrowest raw formats in existence — Epson Raw File, written by the Epson R-D1 and R-D1s, the digital rangefinders Epson built around a Leica M lens mount and then discontinued, leaving no successor camera that writes the format. .ts sits at the opposite extreme: the MPEG-2 transport stream from ISO/IEC 13818-1, the most widely deployed video container on earth, carrying every DVB and ATSC broadcast and a decade of HLS streaming segments.

Converting between them is not a normal photo workflow. It is what you do when a playout system, an IPTV encoder or a segmenter wants a still handed to it as a transport stream — a slate, a test card, an ident, a hold frame. The output is one developed photograph held motionless, with no audio. Everything on this page describes that specific job.

ERF at a Glance

TS at a Glance

How to Convert ERF to TS

1. Upload Your ERF File: Drag the .erf onto the page or click "+ Add Files". Several frames can be queued at once and will be treated with the same settings.
2. Set Merge Strategy and Image Duration: Merge strategy opens on Merge images, which packs every upload into one stream in upload order; Video per image writes a separate .ts for each photograph. Image Duration opens on 5 seconds per frame and offers everything from a single frame at 1/60s up to 10 seconds.
3. Set Background Color, Video Codec and Video resolution: Background Color opens on White and fills whatever the photograph does not cover in the output frame. Video Codec opens on H.264, with H.265, MPEG-1, MPEG-2, MPEG-4, DivX and Xvid also listed. Video resolution opens on Keep original; Fixed Resolutions gives exact width-by-height pairs and Preset Resolutions sets height only.
4. Convert and Download: Click Convert and save the .ts. Files upload over an encrypted connection, are developed and packetized on our servers, and are deleted automatically after a few hours — no sign-up, no watermark, never shared or made public.

What the Render Step Does to the Raw

The R-D1's file is a mosaic of colour-filter-array samples, not a picture. Before anything can be encoded it has to be demosaiced into ordinary RGB, and that render commits the white balance the camera recorded, producing an 8-bit sRGB frame. The R-D1 is a manual camera with no autofocus and no picture styles to reproduce, so the render is unusually close to what the body itself would have shown — but it is still a one-way step. The tonal latitude that made the raw worth keeping is spent at that moment, exactly as it would be in a JPEG.

Two things follow that are worth planning around. First, if the look matters, develop the .erf deliberately in RawTherapee, darktable or Adobe Camera Raw, export a finished image, and convert that instead. Second, archive the original regardless: with no camera still writing the format, an R-D1 file is not something you can re-shoot.

The long edge of an R-D1 frame is around 3000 pixels, which sits under the 4096-pixel clamp every still passes through on the way to the encoder. In practice that means Keep original genuinely keeps the whole developed frame here, where a modern high-resolution raw would be shrunk to fit. It also means the "original" frame size is a photograph's size, not a video raster — 3:2 and about 3000 pixels wide is not something a transport-stream decoder expects to be handed.

Choosing the Settings for a Broadcast Chain

Two details catch people out. Preset Resolutions sets the height and lets the width follow the source aspect, so a 3:2 photograph at the 1080p entry becomes 1620 x 1080 rather than 1920 x 1080 — use Fixed Resolutions when the raster has to be exact. And Background Color is pre-selected as White, which fills the pillarbox bars a 3:2 frame leaves in a 16:9 raster; for most playout contexts you want Black instead, and it is far easier to change before converting than after.

Frequently Asked Questions

Why does the clip have no sound and no movement?

Because its source is a single photograph. The raw is developed into one frame, that frame is held for the Image Duration you chose, and the encode runs at one frame per second with B-frames disabled — so a five-second clip is five identical frames. There is nothing to pan or zoom and no transition. It is silent for the same reason: a still carries no audio, so no Audio Codec control is rendered on this page at all and no audio stream is written. Add sound and motion in an editor after import.

What resolution will the TS actually be?

Whatever the developed frame is, unless you change it. On Keep original an R-D1 file arrives at the encoder at roughly its native size — around 3000 pixels on the long edge, under the 4096-pixel clamp that every still passes through — and that is what the transport stream gets. It is legal H.264 but wider and oddly shaped for a broadcast decoder. Switch Video resolution to Fixed Resolutions and take 1920 x 1080 (Full HD) if the stream is going anywhere near a playout chain.

Why did the 1080p preset give me a 1620-pixel-wide frame?

Because there are two different resolution lists on this page and they behave differently. Preset Resolutions sets the height only and derives the width from the source's aspect ratio, so a 3:2 photograph at "1080p" comes out 1620 x 1080. Fixed Resolutions hands you the exact pairs — 1920 x 1080, 1280 x 720, 3840 x 2160 and so on. For anything that must match a broadcast raster, use Fixed Resolutions.

Is anything left of the raw data after conversion?

No. A transport stream has no concept of undeveloped sensor data, so the colour-filter-array samples in the .erf are demosaiced into RGB and a white balance, exposure and tone curve are baked in before a single packet is written. What survives is a finished 8-bit picture. Keep the .erf — with Epson out of the camera business it is the only full-quality version of that exposure you will ever have. For a full-resolution still master you can actually open in an editor, ERF to TIFF is the right target and ERF to JPG is the shareable one.

Which codec should I leave selected inside the TS?

H.264, which is what the dropdown opens on and what transport-stream players handle most reliably. The container itself is codec-agnostic — this page will also write MPEG-1, MPEG-2, MPEG-4, DivX, Xvid and H.265 into it — but only MPEG-2 and H.264 are routinely expected inside a .ts. Pick MPEG-2 if a classic DVB or ATSC chain demands it. Note that on the non-CRF codecs (MPEG-1, MPEG-2, MPEG-4, DivX, Xvid) the Quality Preset dropdown has nothing to work with for a still source, because there is no source bitrate to scale from; use Constant Quality there instead.

Will my R-D1 file open at all, given how rare the format is?

It should. ERF is a TIFF/EP file with the sensor data in a SubIFD, which is the same shape as most raw formats of its era, so raw decoders that support the R-D1 read it without special handling. Where files fail, the usual causes are a truncated copy off the card or a renamed file that is not really an ERF. If a decode fails here, opening the file in RawTherapee or darktable and exporting a TIFF or JPEG first will get you past it — and gives you control over the develop step as a bonus.

Can I merge several ERF frames into one transport stream?

Yes, and that is the pre-selected behaviour. Queue the photographs in order, leave Merge strategy on Merge images, and each developed frame is held for the Image Duration before cutting straight to the next. Total length is the duration multiplied by the file count. There are no transitions, no motion within a frame, and no audio anywhere. Switch to Video per image if you would rather have one .ts per photograph to place independently.

Is a .ts the same as the .m2ts on a Blu-ray?

They carry the same kind of encoded video but are packetized differently, and the extension decides which framing you get. A plain .ts is 188-byte packets with the sync byte 0x47 at offset 0. A .m2ts uses BDAV framing: every packet is prefixed by a four-byte arrival timestamp, giving 192-byte packets with the sync byte at offset 4. Renaming a file does not re-frame it. Generate the extension your receiving tool actually checks for.

Would MP4 be a better target than TS for what I am doing?

For almost everyone, yes. A bare transport stream is built for lossy delivery channels, not for handing to a person: many consumer players open it, but MP4 opens everywhere, indexes cleanly for seeking, and is what phones, browsers, televisions and current editors expect. Reach for .ts only when something specific in your chain — a DVB or ATSC playout system, an IPTV encoder, a set-top capture tool, an older HLS segmenter — asks for the standard 188-byte stream by name. Otherwise ERF to MP4 is the friendlier clip.

How are my files handled during conversion?

Your .erf is uploaded over an encrypted connection, developed and packetized on our servers, and deleted automatically after a few hours along with the transport stream it produced. There is no sign-up, no watermark, and nothing is shared or made public. Six-megapixel raws are small by modern standards, so upload time is rarely a factor even when you queue a batch.