Low Light Ability of the HPX300

Yes, back to the cam and sub 10Mb video files- that would be great when you get a chance. Two more points of clarification: was your client happy or unhappy with the 300 footage afterall ? Secondly, is the 300 footage less, more, or equal to the 700's noise?

High end client was happy and is settling invoice today. Phew!!!

The noise on the 300 is far greater than the 700 and the lower the IRE value (luminance level) the greater the noise.

On the weekend I test bread and butter DV mode. This will be MY footage so can do/post what I want with it! I can also properly monitor this on a CRT at full res rather than using pesky LCD panels. It is however a valid comparable using same LCD panel as I view the 700 material on.

BD
 
I can't remember in which sample I saw it but there it is again, the noise in the mids which is so visible. Not that I mind it, but I thought it was supposed to be much cleaner.

I'm seeing the same thing throughout all of the posted video, the noise in the mid levels is the most conspicuous.

As an aside, a local retailer is having a 300 demo day today here in town. If time permits I'll be attending. I'm going to try and pop in a card and replicate some of these issues for myself.

Is there anything I can check on for folks here ? I'll head over to the demo in about six hours time, so post a request in the mean time if so.
 
Last edited:
Hotch - glad the stuff I sent helped.

On your test you have to assign the DRS off/on to a user button - so dig deep in the menus!

Here is still with DRS on, 0Db, blacks -11 from a SD DV setting. I could see the hair area crawling in the viewfinder, but I have no noise issues with the still...
face_0dB_DRSHIGH_black-12_stdSDDV.jpg


I'm nailing this cam on the weekend and getting definitive answers on all the stuff we worry about. Again it will be my material and I'll upload freely.

BD
 
I'm seeing the same thing throughout all of the posted video, the noise in the mid levels is the most conspicuous.

As an aside, a local retailer is having a 300 demo day today here in town. If time permits I'll be attending. I'm going to try and pop in a card and replicate some of these issues for myself.

Is there anything I can check on for folks here ? I'll head over to the demo in about six hours time, so post a request in the mean time if so.

Looks you and I went to the same demo - at EAR in Phx, correct?

I had them set up a 300 and a 170 up, side by side, shot a static of lit flowers against a green screen using 1080i 60i with both cameras and created 3 short clips: an AVC-Intra100 and a DVCPRO HD both on the 300, and a DVCPRO HD on the 170. I set both cams to cinegama (Scene File F6) and matched the master ped of -2, and detail coring at +5. I couldn't determine if the DRS was on on the 300 as I was running out of time (everyone wanted to go home - I could have spent the entire day there). I do believe the 170 menu reported DRS off but I'm not sure it is the same function on both cameras. I don't recall seeing a DRS setting on a 200 on which I have most of my experience using a Panny.

The most important thing I was looking for was the crawling noise I have always noticed in the low mids on a 200 for instance and seems to defy defining in a still because, I would agree with someone who posted somewhere here not long ago, that the distraction seems more the Movement of the "noise" than the noise itself, so to speak. There was another shooter with me there and was quite vocal about how one should evaluate images on a great monitor as a client would, not magnified 3x or something, even though that kind of scrutiny might provide clues to an image that seems "off" if it can't be categorized otherwise.

So what I saw surprised me and yet made sense. The most striking thing among the three shots was the amount of clarity and depth of color one saw with either 300 shot compared to the 170's shot - that's what got my attention, not noise in the 30 IRE or so. And although at the time of the test, we had thought that it seemed obvious that we were seeing the difference (among several factors I'm sure) between 10 bit's quantizing output of 1024 values versus 8 Bit's 256, in retrospect, looking at the 300's spec sheet later this evening, I read that 10 bit processing is only used with the AVCIntra format, so the almost identical dynamic, gorgeous shots from the 300 seem to be endemic to the camera overall, at least in 1080 mode. If there had been time, I really wanted to go to 720 as that has been the mode I've shot in the last couple of years.

As to noise, it was not evident full screen, as a client would see it, in pause or running from any of the cam's shots. It's possible that it existed and would be evident if compared to a 30k and upwards camera (which we didn't do) side by side but as it was viewed straight out of the camera on a $6k Panny monitor, none of us saw objectionable noise in any part of the picture and instead commented on the obvious picture quality differences. Now, this was an extremely limited test and other shooting situations might create very different results. My attempt today was simply to try and compare one little aspect - get a handle on what one can expect if one is used to shooting on a 170 or 200.

As to the diff between the 300's AVCIntra 100 shot and the DVCPRO HD shot: the AVC-Intra was slightly smoother but oddly enough, the DVCPRO HD lack of "smoothness" made it seem more dynamic, 3D looking to me than the smoother version - but no one else really felt that way. Whatever the difference was, one REALLY had to look hard to see it.

So what made sense, aside from any creature features or the user interface, camera form factor, etc, the 300's imagery, as tested in our little setup at 1080 justified a camera that costs roughly twice as much as the 170. Noise, if indeed present, would not have been a deciding purchasing issue in our evaluation of the camera based on this test.

Notice that I didn't mention skew - we/I weren't testing for that, and although I did bring it up at one point today, the rep wouldn't say anything more than one imager is scanned (MOS), the other is not, hence skew issues for the 300. I mentioned the faster chip read mod someone else here had posted as a fixit down the road, but the rep didn't touch that comment. The only fixit mentioned was something to remove still flash lighting glitches where part of the frame is lit by a flash going off in mid scan, a software post-process using a frame before or after to correct the bad frame.

So the owner of the store has said he is going to publish the three clips on their blog. And depending on the format of what he ends up publishing (H.264 or something too compressed to be useful), you all may want these three clips to be also available on Giuseppe's FTP in native format. Now I know people had asked for the clips that have been sent to date to remain in P2 format, but that has made it difficult for folks who don't have the ability to convert AVC-Intra clips (like me) because they have PPC and not Intel Macs. But assuming for the moment, that people are mostly using FCP to edit, it makes sense to me for me to publish the AVCIntra 100 clip as a converted ProRes file - that's what would be used for the edit ultimately anyway. Even if there is a ProRes "footprint" as has been suggested, if one is editing in FCP, one is going to have to live with it (unless I've missed something in the latest workflow paradigms offered of late). So unless someone has a better idea, objection, or general lack of interest, sometime early next week I'll send a ProRes file and 2 DVCPro HD files to Giuseppe's FTP. The three files are all 10 secs or less - shouldn't be too big - waaay shorter than this post anyway :).

whew...
 
Last edited:
Ok, since I don't know where it was mentioned what can and can not be posted I'll just ask. Could you post the three frames? I would love to compare the 300 with the 170.
 
Great post renhinks :)

Its probably good for you to do the Pro Res transcode/re-wrap for FCP boys.

I'm doing AVC as 'native' P2 MXF this weekend as a test and this just flies into Avid without any transcodes/tweaks. Unusual for Avid to be ahead, but they had to catch up at some point - the workflow is beautiful! I edit ON the card for mega quick turnarounds.

BD
 
renhinks,

Yes it was the EAR demo. ;-)

AVCintra 100 1080 24PN

I shot some down and dirty tests, requesting that the DRS be turned off by the Panny rep, so if you had the same camera as I , it was probably still off late in the afternoon.

I haven't had a chance to download the footage yet so I can't comment fully on the footage, but I can chime in via the Readers Digest version FWIW. My primary concerns were the noise complaints and skew concerns.

In regards to noise, and evaluating through the monitors they had set-up, it's definitely not a low light wonder cam, and I would avoid cranking the gain up past +3db unless (ENG stuff) needed. Yeah it's not as clean as the 500 or cameras costing twice what the 300 does, no question. However, at it's price point and considering it's feature set and form factor it's a deal. Ultimately, I'm going to reserve judgement on the noise until I'm able to view more comprehensive tests from the likes of Ben D. and others. Like you, my main work station is a G5 (with FCP 5.1.4) , but I do have an Intel laptop that I can view the clips I shot in P2CMS.

As for skew I went out on the street and did some panning shots of cars going by. The posted speed limit is 35 mph, with most cars going 40-45 mph. I shot long lens shots trying to fill the frame and then zoomed out as they went by. They weren't race cars or otherwise traveling super fast, but there did not appear to be any objectionable skewing. That's not to say that the camera doesn't skew, but for my purposes it most likely will not be an issue. The truth will be in the viewing of course.

I pointedly asked about a skew fix as well, and got a similar answer. The flash band fix is very likely the only modification they will be making to improve the cmos issues. A skew fix is not in the cards. Of course we could always be surprised.

I'll post back after viewing the goods.
 
Last edited:
Great post renhinks :)

Its probably good for you to do the Pro Res transcode/re-wrap for FCP boys.

I'm doing AVC as 'native' P2 MXF this weekend as a test and this just flies into Avid without any transcodes/tweaks. Unusual for Avid to be ahead, but they had to catch up at some point - the workflow is beautiful! I edit ON the card for mega quick turnarounds.

BD

Ahh, I was wondering if there was an NLE out there able to use AVCIntra natively. And AVID no less?? - you're right, I would have not expected them to be out in front. I can see now why people would want samples posted in the original wrappers. What I had thought prior to your post here was that it might be some obscure editing box that hardly anyone used. I'll have the dealer unwrap the AVCIntra example into ProRes then - maybe post both versions.

Looks like I jumped ship to FCP too soon :) - I was an AVID editor for 10 years until 03. Your workflow sounds slick.

All of this brings to a head for me if, or how, I'm going to field offload P2 cards in the near term (now that I'm buying a cam and won't be renting a P2Store) and how much I should take into account the ability of a laptop (I have both a PC and a Macbook Pro sans FCP), if that's what I use, to transcode as well. But my musings on this are another thread so I'll stop here.
 
We spent the $75 bucks to try Calibrated Software's plugin that allows FCP to work with the AVC-Intra files in their native state without transcode.

The good - works fine.
The bad - There are no FCP presets for AVC-Intra, so you do need to make sure you set your sequence manually as the timeline won't just adjust to the clip. Also, the clip comes in with a red bar and therefore needs to be rendered in the timeline in order to play back smoothly.

So, not a bad way to handle the AVC-Intra without transcode. Not perfect though.

Here is the link: http://www.calibratedsoftware.com/

I have no affiliation with the company.
 
Yes it was the EAR demo. ;-)

AVCintra 100 1080 24PN

I shot some down and dirty tests, requesting that the DRS be turned off by the Panny rep, so if you had the same camera as I , it was probably still off late in the afternoon.

There were two 300s there until the rep took off with his. What I used was EAR's demo cam so I'm not sure if we used the same bird. I don't usually shoot 60i and I was hoping to work my way down to 24pN in both 1080 and 720 (but ran out ofd time), so it's conceivable noise could creep up in other formats and certainly a broader range of shooting scenarios than what I spent time on.

What is tweaking me (and my meager budget) is that now that I've seen comparable images from both the 170 and the 300, my plans to buy a 170 this month are being re-evaluated - constantly - I think I'm even dreamed from inside a viewfinder last night! However, there are lots of times in my style of shooting where the smaller form factor of the 170 makes it almost a must-have. I have lived with the rented 200's noise the past two years mainly because the clients I have had appreciated my low budgets more than they objected to (if they even noticed it at all) the occasional squirmy shot. I have become quite intimate with FCP's 3-way color corrector, it occasionally competing with my girlfriend's Jammie'd glare and oft repeated "When are you coming to bed?", most often met with me muttering slurs about contemporary production standards and wondering why I ever moved from LA into the sticks.

The plan at the moment is that EAR is going to let me have the 170 demo out for a shoot next weekend. Then, depending on my tolerance for mental anguish, I thought I should also take the 300 out for a test drive before I plunk down my money, although, based on some posts here, I'm not sure I should be shooting the 300 for a client when I do. But, given what I've seen to date, aside from the larger size, the superior imagery and good value of the 300 is sure giving me pause.
 
So unless someone has a better idea, objection, or general lack of interest, sometime early next week I'll send a ProRes file and 2 DVCPro HD files to Giuseppe's FTP. The three files are all 10 secs or less - shouldn't be too big - waaay shorter than this post anyway :).

whew...

I will gladly post up the ProRes files, please contact me when they are done.
 
Ahh, I was wondering if there was an NLE out there able to use AVCIntra natively.
Sure there are. EDIUS 5 has native AVC-Intra support, and Premiere Pro CS4.1 has it; they were demonstrating it at NAB. And Avid. I'm sure there are lots more at the higher levels, but as of right now those are prosumer-priced and they have native support.

As mentioned elsewhere in this thread, FCP can do it on an Intel mac if you have installed the AVC-Intra codec and the Calibrated MXF plug-in.
 
What is tweaking me (and my meager budget) is that now that I've seen comparable images from both the 170 and the 300, my plans to buy a 170 this month are being re-evaluated - constantly
I know what you're saying. I go back and forth on whether I should sell my 170 and put the dough towards a 300. The image is much nicer. However, the form factor of the 170 is so fantastic! And no CMOS issues... When we came back from Africa both Kevin and I were torn on this same issue, because we used both cameras there, and frankly I think I appreciated my 170 more and more as I became more exposed to the 300.

But no doubt, the 300's image is sharper and cleaner, especially in AVC-I, but even in DVCPRO-HD it's still noticeably better.

And less noisy. Not a lot less, but a little.
 
Last edited:
As mentioned elsewhere in this thread, FCP can do it on an Intel mac if you have installed the AVC-Intra codec and the Calibrated MXF plug-in.

Me still PPC. And even if I ponied up for a Mac Pro, the constant re-rendering in FCP using the Calibrated MXF plug-in would be enough to put me off my feed.

But when the 300 becomes a reality for me (as well as an upgrade to FCP 6.0.5 or the Next Generation - 7.0 ? - that (hopefully) incorporates AVC-Intra natively) - and I really think you sent me off on the right track about this with your Africa article - by end of year, I'd like to own both cams, although the intercutting, based on my little test yesterday, would have to be shots that don't have to visually dovetail - I saw too much difference between the 2 images.

Now, my test was in 1080, which I've not been shooting to date - 720 24pN is where I've been living the past year or so and that is what I need to compare next time I'm up to my dealer. The 170 CCD images might come somewhat closer to being able to compete with the 300's 2.2mp 1920x1080 images if one levels the playing ground by shooting in a resolution closer to native for the 170 - 720 or even SD. I may be (way) over-simplifying the tech behind my understanding of this, but I'm rushing to get out the door and act like a normal person on a Friday night - eeehah.
 
Renhinks,

Good to hear that even the 300's DVCPRO HD image is a step forward from it's predecessors.

The heavy CPU requirements of the 300 are a bummer , for field use in particular. My latest thinking is that I'll shoot DVCPROHD on the 300 in the interim until adequate field dumping/viewing gear for AVC-i becomes necessary or paid for by a job that requires it.

In the mean time I have the capabilities of the 300. MY impetus for buying this cam comes from a client who I've been renting a 500 for ( and shooting 1080/30P DVCPRO HD). I have a built in ROI with this camera. I do have a 200 and 170, but this client require a large form factor P2 camera.

I was going to sell the 200 to help off set the potential 300 purchase, but then I land a client that wants to shoot SD DV on tape- go figure.

So...buy the camera for DVCPRO HD now and utilize the AVC-i down the road.

Just thinking out loud.
 
Now, my test was in 1080, which I've not been shooting to date - 720 24pN is where I've been living the past year or so and that is what I need to compare next time I'm up to my dealer. The 170 CCD images might come somewhat closer to being able to compete with the 300's 2.2mp 1920x1080 images if one levels the playing ground by shooting in a resolution closer to native for the 170 - 720 or even SD.
Well... yes and no.

The 170 can generate an image that's about 1440 x 810. So it's more than 720 can handle, and most particularly it's a lot more than DVCPRO-HD 720p can handle. It's a pretty good match for DVCPRO-HD in the horizontal; DVCPRO-HD 1080 stores 1280x1080. So I've always recommended that if you want to get the most out of a 170 or 200, you should be using 1080 mode.

720 mode might seem like it's a better match, but you then lose 33% of the horizontal resolution due to the codec's prefilter down to 960, so -- 720p mode isn't optimal for an HVX or HPX170. 1080 mode is sharper, and noticeably less compressed.

Either way, a 170 can't match a 300 for sharpness in either mode, unless you're comparing dVCPRO-HD 720pN mode against dvcpRO-HD 720pN mode. If you're comparing 720p AVC-I on a 300 against 720p DVCPRO-HD on a 170, the 300 is gonna smoke the 170 in all ways.

But dvcpRO-HD to DVCPRO-HD, 720 to 720, that would be the closest those two will look to each other.
 
Hotch - i still feel your pain and you make me laugh - dreaming from inside the viewfinder indeed! I do relate! Y mi amigo en Buenos Aires Damian - tranquilo por un poco mas tiempo!!!

I have a much better handle on the camera now after tests this weekend and will try and upload/post tomorr.

I have made major changes to the settings and, although noise can be a big issue, I have also achieved some truly stunning results too (and that is feedback based on watching them at usual half res on the P2 viewer). If only I had an HDSDI CRT - I would probably be dancing in the streets!!!

More later...

BD
 
Back
Top