We’ve just added some optimization to the H264 encoding. The video looks much better now, especially at low bitrate. See the image below for the comparison (click to view the original version):

This is done automatically for all the H264 encodings.

There is a weird behavior with the ffmpeg watermarking option (vhook): the watermark is applied before the video is resized. It means the watermark can have a different size depending on your input video resolution. Pretty annoying since you make the image according to your output format and not the inverse. This is now fixed.

This example shows a 1280×720 video encoded in 640×0 before and after the fix:

We always tried to respect the video aspect by adding black bands at top and bottom of the video. In fact, that was true when your video is 16:9 and your format is 4:3.

Now, 16:9 is becoming the new standard because of HD format, HD TV, and even the web. You may already know that YouTube changed their flash player to be displayed in 16:9 instead of 4:3.

This is why, we added the side padding in our internal encoder, so you can now convert 4:3 videos into true 16:9 ones. To have the good aspect, the video will have black bands at left and right like in the screenshot above.

Hi People,

Tomorrow, the biggest European web event will start: LeWeb’08.

Particles will be there to show Hey!Watch (Video Encodind Web Service) and Hey!Spread (Video Mass-Distributing and Tracking Web Service). We’d love to meet you.

So, feel free to contact our VP Business, Eric Fontaine, to meet him at any time:
Phone: +33 676 399 518
Email: eric.fontaine(at)particle-s.com

Woody on the forums pointed out that he can’t find the difference between 1-pass encoding and 2-pass encoding. So I did a screenshot comparison to show him the difference:

Like I said, 2-pass encoding is better, but you need to change your format settings. Two things to take into account: video bitrate and video resolution. The raw video quality is also important, but you already know that an encoder can’t improve the video quality.

Let’s say you want to encode a video in 1280×720 @400kbps, you’ll get the worst quality ever because the bitrate is too low according to the resolution (>= 2000kbps is more appropriate). And you’ll notice that the 2-pass encoding will be totally useless for the same reason. So increase the video bitrate according to the video resolution.

Also, by using the 2-pass encoding, you’ll get another benefit since the encoder knows exactly when to reduce the video bitrate: smaller file size.

To conclude, Tweak your format until you have the best quality / size you need.

2-pass encoding

By using 2-pass encoding, you’ll have the video quality improved significantly. You just need to edit your custom format and check the 2-pass checkbox. Nothing more. It will cost the number of credits for the video * 2, since 2-pass means 2 encode jobs.

MOV, MP4, 3GP Hint

We now hint 3GP files, so they can also be streamed like we did with MOV and MP4 videos months ago. It’s done automatically if the video format is mov, mp4 or 3gp.

FLV Metadata Injection

Here are the injected metadata:

• creator
• metadatacreator
• hasKeyframes
• hasVideo
• hasAudio
• hasMetaData
• canSeekToEnd
• duration
• datasize
• videosize
• videocodecid
• audiosize
• audiocodecid
• audiosamplerate
• audiosamplesize
• stereo
• filesize
• lasttimestamp
• lastkeyframetimestamp
• lastkeyframelocation
• keyframes (filepositions, times)
• width
• height
• framerate
• videodatarate
• audiodatarate

It’s done automatically after the FLV encode.

Don’t forget you can follow us on twitter to give some feedbacks and to get updates about our services.