Panasonic AG-HMC 150 discussion

[Bob Diaz]

After running several tests and timing it this time, I'm finding that the way the MPEG-4 H.264 file is encoded will have an impact on the time to transcode to AIC. This is only a VERY rough estimate... Keep in mind that the AVCHD files from Panasonic may be very different than this. The clips I'm testing are about 20 seconds long.

720/25p = 3.5x, That is 1 minute takes 3.5 minutes

1080/30p = 10x, That is 1 minute takes 10 minutes

The number you get is +/- 20% (Answer x 0.8 or Answer x 1.2)

With the limited files I've tested, it seems that the higher the compression, the longer the time and the lower the compression, the shorter the time.

My MAC is a 2.4 GHz dual core iMac with 2 GB of RAM.


Bob Diaz

[Jonathan Schwartz]

Bob,

How is the time needed for converting to an intermediate codec worthwhile. I love everything about the 150 except for this. Do you think that Apple will come up with a better workflow for AVCHD or is this just going to be the norm?

Jon Schwartz

[Ron Evans]

The SR7 and SR11 that I have are only interlaced but the SR7 is 1440x1080i and the SR11 is 1920x1080i. I see no real differences in times for conversion to Canopus HQ. It takes between 45 and 50 mins for an hour of Sony AVCHD.

Ron Evans

[Bob Diaz]

Originally Posted by Jonathan Schwartz

Bob,

How is the time needed for converting to an intermediate codec worthwhile. I love everything about the 150 except for this. Do you think that Apple will come up with a better workflow for AVCHD or is this just going to be the norm?

Jon Schwartz

When I play a clip in the Preview window, I can play it back without a problem, even the 1080/30p footage. There are some dropped frames, but at least I can see what the video is doing. At this point, selecting the In/Out Points is very easy, because the response of the pointer is reasonable.

However, once I drop the clip into the timeline, things slow down a bit. Now with the 720/25p footage, I can stand the delays, BUT the delays I find with the 1080/30p footage are getting in the way of being able to work properly. In the timeline, neither 1080 nor 720 can play back with audio at normal speed and the video has dropped frames.

Once something is rendered, it is converted into AIC (Apple Intermediate Codec) and all the delays are gone. Now I can play back with audio and there are no delays in moving the pointer.

So rendering allows the system to respond faster and makes editing a lot easier.


In terms of work flow with the HMC-150, I don't know if importing clips allows for rendering at that point. (See below.) However, I know that a rough cut can be done on the clips in the preview window. Once the selected clips are dropped to the timeline, rendering would make sense.

Because rendering could take a while and the computer does not need my help, I would let the computer do it's thing what I do something else. After it's done, I would finish the edit...


(Below...) IF I rendered the files using Quicktime to convert them to AIC, the AIC files could be read direct to Final Cut Express and there was no need to do additional rendering. The response of the editing pointer is very good when pre-rendered.


Bob Diaz

[Jonathan Schwartz]

Bob,

I shoot multicamera events with anywhere from two to four cameras. The thought of having 4 cameras filming an 8 hour event and then having to ingest, convert and render the footage seems daunting. Seems to me that the advantage of using solid state goes out the window when you take all of this into account. If you are spending the same amount of time converting your footage as you would if you were importing tape, why loose the advantage of having a tape as a backup? Just curious what you see as the advantage to this set up vs. say the Canon XH-A1.

Thanks for your thoughts.

Jon Schwartz

[Ron Evans]

I also shoot multicam and have used my SR7 and SR11 with my FX1. The advantage of the SR's is not needing any tape for the whole event and cost less than a Firestore etc to attach to a HDV cam. The Sony Motion Browser software will backup to disc for file sizes that fit on a DVD9. But now I will just backup all the files to a Bluray disc. Still more expensive than tape but coming down quickly. Yes its an extra step but I wanted a cam for family use rather than lugging around the FX1 and the SR11 is much better for these tasks. For the last event it did an excellent job intercutting with the FX1 on closeups with the SR7 fixed wide. The event was 2 hours and 20 mins, used two 80mins HDV tapes in the FX1. Capture time was realtime for the HDV and about 5 hours total for the SR's to HQ. So instead of a tape based being 7 hours capture it turned out to be 7 hour and about 35 min in the end. So 35 min longer in 7 hours or so. Not really significant for me anyway in that most of the time I didn't have to do anything!!!!! As I mentioned earlier I edit with Edius 4.6 and the time to get to a HQ file for each of the cameras was not really a lot longer than using just HDV. I used to save finished projects to standard size DV tapes which are actually more expensive than 50G double layer Bluray discs!!!!!! I too like the security of tape knowing that even if I mess up a piece I can cut and splice and get the tape going again. Not possible with disc based systems. However with discs it is fairly easy to just make a few copies.

Ron Evans

[Wacharapong Chiowanich]

Originally Posted by Jonathan Schwartz

Bob,

The thought of having 4 cameras filming an 8 hour event and then having to ingest, convert and render the footage seems daunting. Seems to me that the advantage of using solid state goes out the window when you take all of this into account. If you are spending the same amount of time converting your footage as you would if you were importing tape, why loose the advantage of having a tape as a backup?
Jon Schwartz

This is possibly why no camera makers have adopted AVCHD at any bit rate for their professional systems so far. My guess is they have foreseen that the technologies would not likely be adequately developed on the computing systems' hardware and NLE side to allow editing to be done efficiently enough (i.e. comparably to the long-GoP HDV, VBR Mpeg2 or I-Frame only DVCPro HD etc.) on most reasonably-equipped NLE platforms. Most users of their consumer AVCHD cams who are their main target market probably don't have any ideas what we have been complaining about but they sure love what they see on their LCD TVs.

Panasonic will be the first to try but I'm afraid their scheduled launch sometime in the 2nd half of this year maybe a little to early considering the initial specs for their new cams include the yet-to-be-tried bit rates of 21-24 Mbps. The file sizes will definitely be larger and require some intermediate codecs to expand to 200-300+ Mbps, putting unprecedented strain on even today's high-end computer systems.

Looking at it purely on commercial terms, tape-based HDV, DVCPro HD or VBR Mpeg2 (for higher quality output) are perhaps more cost and definitely time-effective in most applications for several months to come until more advanced CPUs, GPUs and on-camera IPs are available.

Wacharapong

[Bob Diaz]

Originally Posted by Jonathan Schwartz

Bob,

I shoot multicamera events with anywhere from two to four cameras. The thought of having 4 cameras filming an 8 hour event and then having to ingest, convert and render the footage seems daunting. Seems to me that the advantage of using solid state goes out the window when you take all of this into account. If you are spending the same amount of time converting your footage as you would if you were importing tape, why loose the advantage of having a tape as a backup? Just curious what you see as the advantage to this set up vs. say the Canon XH-A1.

Thanks for your thoughts.

Jon Schwartz

While it looks like the higher data rates convert to AIC or ProRes faster than the lower data rates, it will be another 3 months (October) before the HMC-150 comes out and we can measure the time to convert. The time one takes for conversion will depend on the speed of the computer; so faster systems have an advantage.

Because no single solution is perfect for everyone, we enter into a trade-off between different systems. For example, P2 may cost more, but is in a format (DVCPRO HD) that can be edited directly on some editing systems without conversions delays. To compare solid state recording to tape, there's a plus side and a minus side.

Plus side to solid state over tape:
Pre-record function, you don't have to burn up storage waiting for something to happen, the video captured 3 seconds before you press record is captured to the file.

No tape drop-outs.

Can mark and delete bad clips in the field, thus freeing up space.

Can mark good or important clips in the field, thus making it easier to find the right clip.

Ability to review any clip and not worry about rewinding or fast forwarding to continue recording.

It a lot easier to make a backup copy of the file than with tape. (I've had older tapes fail on me.)

It's easier to pull a single thing from solid state then with tape. (If you had to find one shot in the middle of the tape, you will have to fast forward and rewind to find the clip.)

IF working in P2, DVCPRO HD, one can start to edit right away without conversion or copying from tape to HD. (AVDHD could do that, but it would take a powerful computer to edit native. As processing power increases, in 3 to 4 years, this will not be a big issue.)

If working with P2, features like Time lapse, slow motion, and fast motion are possible.


Minus side to solid state over tape:
Saving the files requires using hard drives or optical media. (Its best to have more than a single copy of the file.)

With AVCHD, you must either use a very powerful computer OR have a delay time for transcoding.

If using P2, the cost of the cards adds up. However, the price per GB has been dropping over the years.



I'm sure there are other reasons to consider both pro and con, but I just can't think of them right now. Because there's a trade-off, depending on each person's needs, some will find AVCHD works for them, some will find that P2 works for them, and some will find that tape works for them.

Bob Diaz

[Noa Put]

Originally Posted by Wacharapong Chiowanich

My guess is they have foreseen that the technologies would not likely be adequately developed on the computing systems' hardware and NLE side to allow editing to be done efficiently enough

Think that depends on the workaround solutions there are, look at this article: http://kb.adobe.com/selfservice/view...nalId=kb403297
The Blackmagic Intensity Pro card's seem to offer real good value at low cost with real-time capture and giving Premiere Cs3 an easier format to edit.
At least, that's what I understand from it, I"m still looking for more user experiences but if this card has the potential they are saying, I don't see any reason why not to work with AVCHD.

[Kevin Shaw]

Originally Posted by Wacharapong Chiowanich

The file sizes will definitely be larger and require some intermediate codecs to expand to 200-300+ Mbps, putting unprecedented strain on even today's high-end computer systems.

I don't see why AVCHD would require anything different for an editing intermediate than HDV, which works fine when converted to an intraframe codec at ~80-100 Mbps. Given how large and cheap hard drives have gotten, this isn't a big deal for anything short of hundreds of hours of footage - and even then it wouldn't be a huge problem.

As far as tape versus solid state is concerned, that will only be fully resolved when good video cameras use memory which is cheap enough to buy in bulk without taking out a loan, and the corresponding editing process is painless. Panasonic could solve that today by shipping an affordable camera using AVC-intra on SD memory cards, but they don't seem to be inclined to do that. Or Sony could start approving standard SSD cards for basic recording on the EX1, but they don't seem to be in a hurry to do that either.

I'd say Bob nailed it with his comment that we have a series of trade-offs now with different resulting preferences for different users. His list of pluses and minuses are a bit biased toward his preference, but suffice it to say that solid state memory has some workflow advantages while tape is still a good medium for many circumstances. And most tape-based cameras can now be used with a hard drive recorder to get some of the advantages of solid-state recording, but in a less convenient form factor.

[Bob Diaz]

I'd say Bob nailed it with his comment that we have a series of trade-offs now with different resulting preferences for different users. His list of pluses and minuses are a bit biased toward his preference, but suffice it to say that solid state memory has some workflow advantages while tape is still a good medium for many circumstances. And most tape-based cameras can now be used with a hard drive recorder to get some of the advantages of solid-state recording, but in a less convenient form factor.

I'll be the first to admit that I have a bias toward solid state recording and the HMC-150.

I guess I should have added to the list of the plus for tape is it's a proven technology. I should have also spent more time thinking about other things that are a plus for tape...


I'll try to attach photos of the copy of the early specification sheet from Panasonic on the HMC-150. I only scanned in the key parts...


Bob Diaz

[Ethan Cooper]

I really like what I see out of those specs. Here's some food for thought:

32GB of 21Mbps AVCHD storage (SDHC class 6 card) will cost you $430
32GB of P2 storage will cost you $1500

1920 x 1080 and 1280 x 720 take up the same amount of space on the card according to the chart. Why is this?

It shoots 29.97p and 23.98p (native) which is good, but I don't see 29.97i listed. Does this mean no 1080i? Or does it mean 1080i but only at 59.94? I'm confused. Not that I'm a huge fan of 1080i, but some people might want to mix and match with their 1080i HDV cams and this would make it easier.

Overall, this thing seems impressive on the surface. I can't wait to see some footage from it, especially in low light since I do weddings. I've been hoping this is the camera I've been waiting for, in the price range I need it to be in. I can't afford an EX1 or camera in that range.

[Bob Diaz]

I can understand 1920x1080/30p and 1280x720/60p taking up about the same amount of space on the card, but you would think that 1280x720/30p and 1280x720/24p would take up less space.

Either:

(1) This is just a mistake on the early flier for the HMC-150.
.......OR
(2) Both 1280x720/30p and 1280x720/24p have much lower compression ratios than 1920x1080/30p.

IF it's case 2, then the 1280x720/30p (& 24p) have a higher bit rate than the JVC's HDV video (19 Mbps). In that case, the compression noise would be a lot lower than HDV.


If you look closely at the specification, every mode has to be at the maximum data rate, except the 1080/60i. See Recording Format. The 1080/60i can be at the maximum data rate, but can also be lower. Not that I'd ever want to record at a lower data rate.

DV Expo West comes to the LA area in early November and I can't wait to check out the HMC-150. It should be shipping by them, so this will be the final product and not hte prototype.


Bob Diaz

[David Heath]

Originally Posted by Bob Diaz

I'm sure there are other reasons to consider both pro and con, but I just can't think of them right now. Because there's a trade-off, depending on each person's needs, some will find AVCHD works for them, some will find that P2 works for them, and some will find that tape works for them.

Bob - I broadly agree with the pros and cons you mention, but doesn't that all make the case for a dual approach the stronger? Maybe sometimes you CAN have cake and eat it?

The obvious example is the Sony Z7 - tape AND/OR solid state? (And cheap Compact Flash at that.) But it doesn't end there, in the pro market the PDW700 is going to have an SxS add-on option (OK, that's disc and solid state, but......), and such as the Firestore make a dual harddrive/tape option viable for other tape cameras.

And solid state doesn't uniquely equal P2 any more. SxS is the obvious competitor, but CF is being used more and more in recording from the Z7 to the XDR - the latter proving you don't need P2 or SxS for very high quality. And SD in the consumer sector. As regards one of your solid-state cons (cost) I agree when Kevin says:

Panasonic could solve that today by shipping an affordable camera using AVC-intra on SD memory cards, but they don't seem to be inclined to do that.

What a shame the HVX170 isn't AVC-I 50 with SD cards.

[Barry Green]

Originally Posted by Seun Osewa

Great question! Because I don't understand how an extra conversion step in the workflow could possibly save time.

Not all editors require the footage to be converted. EDIUS Broadcast and EDIUS Neo can play AVC-HD files directly, as can editors like Ulead and Pinnacle Studio. Now, that doesn't make them efficient; AVC-HD is extremely processor intensive, so until CPUs catch up it's likely that you'd actually want to convert the files. But it's not an absolute requirement, unless you're using an editor like FCP or Premiere Pro CS3 which don't have native AVC-HD file playback.

On a dual-core 2.4GHz system, EDIUS plays native AVC-HD files at about 12fps (my guess, not a scientific calculation!)