75Mbps GH1 Peak Performance Patch

To get a tripod is certainly a good idea - and even a cheap one is better than none. Why don't you try a Canon FD lens - such as a 50mm f1.4 - one can get those for a couple of bucks on e-bay and they a great for filming with the GH13 - in particular at low light.

Thanks for the recommendation. I've never bought anything off of E-bay. But, if I can get a Canon f1.4 lens for a few bucks, it might be time to roll the dice and try my luck on E-Bay? But, surely I would also need some type of lens adapter to mount between the GH1 and the Canon lens? I only see a few adapters listed in my GH1 manual, and none seems related to Canon FD lenses. Do you know what type of adapter is required to mate the Canon lens to the GH1?

BTW: I assume that even after using the adapter, the x2 factor stills applies, so, this would make the 50 mm effectively a '100 mm' in terms of old 35mm film cameras? That's getting to be a bit too narrow of a FOV then I would like.

What does the 'FD' in the Canon lens mean? Would the auto focus function still work with the GH1 with this Canon FD lens, or would it be strictly a manual focus lens when mounted on a GH1? And, don't worry about answering these very basic questions. I'm just thinking out loud. I can Google around and find out.

A new lens for a couple bucks? Hmmm...I really do like the sound of that.
 
They're not really a couple of bucks - more like $100. Yes, 100mm field of view in 35mm stills terms. Adapter from Ciecio7 on eBay is he one - FD to m43. FD doesn't stand for anything much important. They are old 1960s to 80s Canon lenses. THey are manual focus only.
 
There are many old lenses you can get for $20-40 that will give you superb optical quality. I like the Rokkor 50s, but there are so many to choose from. Color is just as important as sharpness.
 
There are many old lenses you can get for $20-40 that will give you superb optical quality. I like the Rokkor 50s, but there are so many to choose from. Color is just as important as sharpness.

Go with Canon FD lenses on Craigslist. I got a 28mm and 50m 1.7 for $25 each.
 
Once you have tried the FDs you will love them. The 50 1.4 is absolutely amazing - although already in the tele range. I have also a 28 2.8 and a 85 1.8 - they are all great. Of course - you have to focus manually - it is a competely different experience compared to the 14-140 - but very enjoyable.

I also registered with e-bay just to get the FD lenses.
 
Once you have tried the FDs you will love them. The 50 1.4 is absolutely amazing - although already in the tele range. I have also a 28 2.8 and a 85 1.8 - they are all great. Of course - you have to focus manually - it is a competely different experience compared to the 14-140 - but very enjoyable.

I also registered with e-bay just to get the FD lenses.

Thanks to all for the advice. I'll have to start scouring CraigsList and E-Bay for some of these lenses. For that price they sound great. I noticed that the adapters for the Canon FD lenses run ~ $30 - $100 on EBay, and new adapters cost > $200 (Novoflex). Are the cheap adapters available on E-Bay any good? My one complaint about the GH1 is the slop in the lens mount that causes the "Check if Lens is attached" errors. Is that issue aggravated by using an adapter, or, maybe it gets better?

I sometimes use the manual focus ring on the kit 14-140 mm, and love the huge magnification when focusing. I don't suppose there is any such thing withe Canon FD lenses? That is: do you get the great zooming while focusing effect that you get on the 14-140? Do the Canon FD lenses have that dohickey in the ceneter that lets you know when you are accurately focused? I don't know what its called but my dad used to have a 35 mm film camera, and as you focused a little blurry outer circle would become invisible when you focused. But, that was a real SLR. Do the Canon FD lenses have something like that to indicate 'in focus' or not?

These questions probably belong in another thread (like GH1 hardware and lenses thread)?

But...I'm kinda hoping lpowell will weigh in with an answer to my question of which patch he thinks would give the 'best' results for 1080/24: the peak performance patch or the new blackout-powell patch for native 1080/24p?

I'm very pleased with the 1080/24 in the 60i wrapper from the peak performance patch, and...I only get a chance to 'play' with the camera on weekends, so...I'm lazy and relying on the work of others to determine the best resolution patch so I can *use* the patch when I get to play with the camera. Too busy with work thriugh the week to even touch the camera.
 
But...I'm kinda hoping lpowell will weigh in with an answer to my question of which patch he thinks would give the 'best' results for 1080/24: the peak performance patch or the new blackout-powell patch for native 1080/24p?
Actually, I'm hoping someone will do an independent shoot-out for us to evaluate. Most of the shots I've taken with the Blackout-Powell Native 24p Patch have been designed to break it rather than demonstrate how excellent it looks.
 
Regarding GOP size, LPowell (I wish I knew your first name :-), the standard settings set by Panasonic are 15 for NTSC and 13 for PAL, to achieve 30fps for pre-pulldown NTSC, and 13 which is close to the 12.5 GOP for half a second of 25p.

So, going native, I see you've set the 24pN to a GOP of 12 instead of 15, which makes some sense. For PAL and 25pN, setting the GOP is harder, right? We ideally want a GOP of 12.5, and that's not possible, so we choose between 12 and 13. Why would you think about setting it to 15? Does it not matter if they are more than 0.5 secs?

Also, might it not make more sense for us PAL users to forget native as we don't have pulldown issues to deal with, and just shoot in a 50i wrapper? What would be the negative in that? It can be made fully progressive in post very easily I believe, though I've never done it myself.

Thanks
Lucas
 
I think that I already commented to similar post once before, so don't blame me if I repeat it once again.
So: why GOP should be 0.5 seconds long? Is there any physical reason? I believe that it's wise to have GOP which is a multiple of 5 for NTSC (due to 3:2 pattern) or multiple of 2 for PAL (upper & lower field), but it is not necessary. What does SECOND have to do with it?

Let's imagine that world leaders agree that a new second should be 2/3 times the old (today's) second. Does it mean that we would have to change GOP size as well from 15 to 10? Of course not. Note that second by itself is not directly related to GOP size. It is just indirectly related since we should not have too long GOPs (let's say equivalent to [today's] several seconds).

However, the stability could depend on relation between GOP size and frequency of writing data to SD card by the camera (although probably not so much).
 
Last weekend it happened that 75Mbps patch resulted in a speed error (PAL version, 25p-wrapper). I am attaching the clip (fortunately it was taken up to the error message) if somebody is willing to experiment with the clip. The clip was taken with 14-42mm Panasonic lenses, but with locked focus and exposure. The clip was taken simultaneously with two cameras (through half-mirror setup) and speed error appeared on both cameras simultaneously (the resulted clip length on both cameras was exactly the same = 42 frames). At first sight the clip doesn't seem so problematic for me (when compared with a pure grass, larger tree, etc). I was using Transcend 16GB, Class 10 card.

 
Last edited:
The clip was taken with 14-42mm Panasonic lenses, but with locked focus and exposure. The clip was taken simultaneously with two cameras (through half-mirror setup) and speed error appeared on both cameras simultaneously (the resulted clip length on both cameras was exactly the same = 42 frames).
That's remarkable how you got precisely identical errors on two cameras at the same time! I downloaded the clip from Vimeo and examined it in Stream Parser, but unfortunately, it doesn't appear to be corrupted or contain any clues as to what may have caused it to fail...
 
So, going native, I see you've set the 24pN to a GOP of 12 instead of 15, which makes some sense. For PAL and 25pN, setting the GOP is harder, right? We ideally want a GOP of 12.5, and that's not possible, so we choose between 12 and 13. Why would you think about setting it to 15? Does it not matter if they are more than 0.5 secs?

Also, might it not make more sense for us PAL users to forget native as we don't have pulldown issues to deal with, and just shoot in a 50i wrapper?
From my MPEG-4 development work, I suspect the prevalent 0.5-second GOP-length is no more than a working rule of thumb. I don't believe I've ever seen any technical requirements for AVCHD encoding that specify minimum or maximum key frame rates. Highly compressed Blu-ray encodings commonly use much longer gaps between key frames.

The MPEG-4 container that AVCHD files use was adopted from Apple's Quicktime file format and inherited many of its legacy encoding practices. In these files, video data is divided into "chunks" which are typically about 0.5 seconds in duration. This turned out to be a good data size for interleaving audio and video streams in a manner that enables playback to be synchronized without audio dropouts. I suspect this may be where the AVCHD 0.5-second GOP-length arose.

(BTW, you can call me LP if you like :)
 
Here are some high detailed shots I got using this patch of the blue mountains in australia. Absolutely no problems using a class 10 card
http://vimeo.com/21580195

Im still pretty new at this, so please let me know if you have any suggestion on how to improve
 
Here are some high detailed shots I got using this patch of the blue mountains in australia. Absolutely no problems using a class 10 card
http://vimeo.com/21580195

Im still pretty new at this, so please let me know if you have any suggestion on how to improve


Nice video. One little suggestion: try enabling auto exposure / autofocus lock on back of camera just before you start recording. That way the lens won't continually hunt for proper focus and exposure will remain constant. Have fun!
 
In the last week it happened three times that I've got speed errors on Transcend 16GB Class 10 with 75Mbps patch.

This time I've made the following:
After focusing and using AFL and AEL (like I always do), I placed my hand in front of both cameras (3D shooting) so that hand covers more than half of the scene. Then I started shooting. After about 3-5s I removed my hand and the cameras continued shooting in all three cases without any problem. I think that I read something similar somewhere on that forum. So, in some cases it might work and save your shot. Please, try it on problematic shots and tell us if it works. :thumbsup:

By the way, when I checked the first problematic "shot with a hand" with Stream parser, I found out that camera needs several seconds to increase the bit rate AFTER the hand is already removed from the picture. Strange.
 
...After focusing and using AFL and AEL, I placed my hand in front of both cameras (3D shooting) so that hand covers more than half of the scene. Then I started shooting. After about 3-5s I removed my hand and the cameras continued shooting in all three cases without any problem.
In many high bitrate streams, Stream Parser displays an extremely high peak data rate for the second key frame in the video. This occurs about 0.5 seconds after the video starts and likely explains what happens in cases where a recording fails immediately after starting. The Peak Performance Patch is probably more prone to this type of failure, as it permits a peak bitrate of nearly 75Mbps, significantly higher than any of my other patches. As crunchy's experience shows, starting a problematic shot with the lens partially covered may help the encoder to pass safely through this initial rough spot.
 
Back
Top