News at Magic Lantern

[Chris Barcellos]

Trammel Hudson has been at it again.

He just updated and posted a Beta version of Magic Lantern, that runs with Canon Firmware 2.08.

More exciting is a post of an Alpha version of Magic Lantern that actually is attempting to control the bitrate- raising and lowering. I have have shot a few minutes of footage using fixed bitrates that are posted as being from 42 mbps to 68 mbps.

I am trying to devise some tests to determine what kind of quality difference there is.

Here is Wiki site where Trammel Hudson discusses his findings:

Bit rate - Magic Lantern Firmware Wiki

[Ben Denham]

Wow this is very cool. It's always great to see how this project progresses. Looking forward to any analysis of the the higher bit rates.

[Danny Winn]

Thats awesome!

I'm about to get the 5D and was worried because I knew ML wouldnt be compatible with the 2.08 version.

If anyone else has installed to their 2.08 version yet please give us updates!

Thanks

[Chris Barcellos]

I have run the new 2.08 version without having any issues. It seems to run very clean.

The alpha version with the bit rate change runs the same way, except for the adjustable bit rate. But jury is still out on benefits of bit rate manipulation.

[Ben Denham]

I'd be interested if anyone could chime in with a ideas about how higher bit rates might (theoretically at least) translate into noticable differences in image quality? My experience with this suggests that higher rates would possibly give greater flexibility in colour correction and post processing (i.e. you would be able to push and pull the image a bit further before it starts to fall apart) but are there other benifits that we might expect to see with the higher bit rates?

[John Vincent]

I'm curious about this as well. I mean, you'd think it would be better, but who knows? I also wonder exactly which bits of info get that extra data - is it the color scheme as you suggest, or resolution, or something else?

[Danny Winn]

Me and a buddy of mine are trying to find out if the quality of sound going into the 5d using Magic Lantern and a "JuicedLink" unit is just as good as using a Zoom H4n. Does anybody have an opinion on this?

[Peer Landa]

Originally Posted by Danny Winn

Me and a buddy of mine are trying to find out if the quality of sound going into the 5d using Magic Lantern and a "JuicedLink" unit is just as good as using a Zoom H4n. Does anybody have an opinion on this?

Well, not really an opinion but more like a fact: the H4n has 24-bit/96kHz ;^)

-- peer

[Nigel Barker]

Presumably the recording time at 62 Mbps is limited to about 8 minutes.

[Paul Cook]

All I really care about on the 5DmkII are 60p and full HD out in record for monitoring. Guessing 60p is out of the question but what about the full HD? Hasnt that been in the wings for ages now?

[Jon Fairhurst]

Originally Posted by Peer Landa

Well, not really an opinion but more like a fact: the H4n has 24-bit/96kHz ;^)

-- peer

The 24/96 thing is more of a marketing advantage than a technical advantage. I record everything at 44.1 kHz (music) or 48 kHz (film), depending on the use. The exception is when you are recording for sound design and you expect to slow the source down (like slowing a lion's roar down to a rumble.) Those inaudible high-frequencies become audible when slowed down.

24-bits is great, but in the case of the H4n, the preamps only have about 14-bits worth of dynamic range at best, so the extra eight bits are only recording noise.

Don't get me wrong. 24-bits is an awesome - but it's only worthwhile if you have more than 16-bits worth of signal at the A/D converter. Of course, all your processing should be at 24-bits or better. And the final process should be to dither back down to 16-bits for your final output.

96 kHz is more questionable - unless you are delivering at 96 kHz. Downsampling from 96Khz to 48kHz adds (very subtle) artifacts. Some high-end producers avoid downsampling in software altogether. They output through a (very expensive) 96kHz D/A and re-sample with a (very expensive) 48kHz A/D to get the best results.

The bottom line is that the H4n preamp doesn't merit 24/96 recording.

[Peer Landa]

Originally Posted by Jon Fairhurst

The bottom line is that the H4n preamp doesn't merit 24/96 recording.

I'm sure you are right. My disclaimer (or excuse ;^) is that I've never used or even seen a H4n -- the dual sys I use is a Sony PCM D1 and a D50 (which I'm quite happy with, I may add).

-- peer

[Bill Binder]

Originally Posted by Danny Winn

Me and a buddy of mine are trying to find out if the quality of sound going into the 5d using Magic Lantern and a "JuicedLink" unit is just as good as using a Zoom H4n. Does anybody have an opinion on this?

Actually, they both kind of svck, lol.

At least from a true audiophile perspective, but with the right setups they can be usable for sure.

And just because something has 24-bit doesn't mean jack, there's so many units out there with 24-bit that don't even come close to doing anything but give you a bigger file.

[Martin Guitar]

I can't get it to crash even at QS -16 with a lot of motion. It's recording with no issues whatsoever. The real issue is that i can't seem to spot any major difference in image quality / compression from the original firmware.

I'm using a Sansdisk Extreme UDMA 60Mb/s 8GB.

[John Wiley]

Martin, I'm not surprised by your findings (or lack thereof).

The Canon DSLR's already have an abnormally high h.264 bitrate to compensate for the poor encoder, and I'm sure Canon didn't just throw around a random number for the bitrate - they most likely found it to be the best performance/file size comprimise and further increasing the bitrate will do very little visually.

It's an entirely different case to the GH1, for example, which was hampered by a lower bitrate as well as poor implementation of the AVCHD codec.