LS300 for narrative filmmaking?

[Mike Halper]

I have been considering the JVC LS300 for some time. It's really between the Ursa Mini 4.6K and the LS300. I'm very concerned about the image quality, reliability and QC issues with the Ursa Mini (FPN, magenta, crosshatching, black lines, audio, etc), so am reluctant to go that route. And at about half the price the LS300 is very attractive.

I would mainly be using it for narrative projects (shorts, commercials, music videos, maybe even a micro-budget feature). The problem is that it seems very rare that anyone is shooting with the LS300 in narrative situations, so I don't really know what to expect. I can only find 3 shorts online, the 2 Heist shorts and another. And no reviews about using the LS300 in narrative shooting.

I'm sure the LS300 is great in documentary and corporate situations as I've seen plenty examples of those. But I think narrative situations can push a camera to different extremes and in different ways than a documentary or corporate shoot can, especially when doing grading and VFX work in post.

Does anyone have any narrative projects shot with the LS300 that they could point me to view? It would be a great help in my decision-making. And any feedback and reports of your experiences of shooting with the LS300 in narrative projects would be helpful as well. Thank you.

[Alex Humphrey]

Mike,

5:00 am thoughts before coffee.. I'm not sure yet but here is my rambling thoughts. My apologies if it doesn't make a lot of sense.

I just got my JVC LS300 2 weeks ago for narrative indie work straight to DVD/BluRay and 4K downloading and maybe if I'm lucky send to film festivals in 1-2 years from now.

My first thoughts is that my shot in the dark for a new camera will be good enough. I had intended to purchase the Digital Bolex D16 Cinema 2K camera about the time they stopped making cameras. I as well looked at the Black Magic Ursa Mini. I chose the JVC due to MFT to Nikon and other lens adaptability (I have a dozen Nikon lenses laying around) instead of the Ursa Mini that was PL and Canon only. Also the JVC 4K at 150mbps is about 50 gigs per hour on cheap SDUHS cards. 64g for $32 for instance and edits fairly easy on 1 external spinning HDD on my macbook. Cinema 4K/2K/1080p raw well not so much. far more data per second and far more CPU/GPU intensive to edit raw. Probably easiest at the moment to use Adobe Premier (lease per month) for editing raw and faster external HDD's (RAID or fast SSD?). JVC 4K 150 MB/s, Ursa mini 4K CinemaDNG Raw265 MB/s (3:1) 125MB/s. Or easier to edit (not needing to deal with raw) Apple ProRes 444XQ 312 MB/s all the way down to ProRes Proxy (yuck) 22.4 MB/s. Also Ursa Mini shooting on more expensive CF2 cards. So about twice the data per second with cards that cost 5x as much per gig. So the cost of running the Ursa mini is more. But all of that you probably already know from doing your research on B&H and Adorama.

So thoughts on using the JVC?

1. The rumors of it feeling CHEAP in your hands is NOT exaggerated. It is plastic housing and care should be used when handling it. If you were working with a crew for months I would buy a few of them for the the grip smacks into it with a C stand. Then again we lost a Arriflex overboard one day in Mexico while on Titanic so anything can happen.

2. J-Log I'm experimenting with takes the JVC from being a ho-hum 8 bit camera to pushing the limits of 8 bit image. dropping in the LUT in FCPX (with addition of Color Finale plug in $99) to enable FCPX to use Luts ballparks the image nicely. Issues is enhanced gain in shadows makes noticeable noise in dark areas. This usually looks like grain, but if I don't like I can use Color Finale plug in to use curves adjustment and crus the blacks just a little so the digital grain look goes away in the shadows.

3. Easy to use, light and small compared to my OLD JVC HD110 with IDX battery and HDD recorder.

4. Post work in FCPX? Easy, fairly fast with my 15" MBP 2014 model. Mini display port to Dell 27" 4K monitor both at 60hz pretty good. The color balance and gama looks like the older the JVC ProHD cameras (which I like) not nearly as video as Sony or as garish as Canon.

5. Image quality? Much more latitude than my Nikon DSLR's, more data per second than my friends GH4's (so should be cleaner?) and I swear more DR than either DSLR's.

6. It's 8 bit. It's not 12 bit raw or even 10 bit QT. No 10 bit output via connectors. However uncompressed 4K live output via the HDMI to say a Blackmagic 4K video assist or Atomos Ronin recorder/monitor (I hear the Atomos now has a JVC LUT so you can see what your footage will look similar too after grading live on the monitor as your recording 10 bit less compressed from the 8 bit uncompressed live image.

7. Being 8 bit does it scream 8 bit? Not so far, at least not too bad. Some of my DLSR work when it was perfect comes close, but most of the time my Nikon's DSLR D5200 just cold NOT handle any contrast and post work with Nikon footage had very little latitude before falling apart. I nearly always get a usable image from the JVC LS300 vs my Nikon DSLR that is hard to use.

8. Both JVC and Blackmagic have sample files for you to download and edit with on your home system. (that's great news) JVC did not at the time have any JLog footage to download but even so I was impressed enough to risk $3,000 on a camera I hadn't seen in person.

9. JVC does NOT list their camera as a cinema camera. Other companies with cameras not as good do list their cameras as a cinema camera. I would place the JVC in between a DSLR like the Panasonic GH4 and a Blackmagic Ursa Mini. Slightly more expensive than a GH4 when prepped for video and better in nearly every way. Same price as a Ursa mini, but half the price of a Ursa Mini when prepped with batteries and recording media for 8 hours of shooting on a 12 to 15 hour day. Image quality in between the two.

I'll try mixing some of the downloaded Digital Bolex 1080p raw footage with my JVC LS300 and perhaps some Blackmagic footage to see what I really think side by side.

For me getting back into video work it was probably the right choice due to cost of operation and quicker to edit and work with than raw, though the BM does shoot QT ProRes as well but at even higher MB/s than cinemaDNGraw.

Now if you are in an area to be hired out for bigger projects or have been working a lot recently on your own projects you may want the benefits of the 12 bit raw or at least 10 bit pro res for broadcast and paid indie work. If your like me who is going more indie route to get back working again, the JVC may make more sense for the film festival, broadcast and straight to DVD/BluRay amazon/itunes download as well as 4K youtube etc.

As far as real world Narrative work? I'll let you know this winter when I have time to work on my own projects, but for me for right now, it was probably the right choice.

[William Hohauser]

It's as fully functional as any modestly priced large sensor camera out there. The feature set is tilted towards a self-contained camera where many of the other options out there work better with extra audio equipment and at least one more crew member. The image is great and sure 10bit is better but so is 6K, the question is whether that makes any difference in the end. Does 8bit 4K look better than 8bit 1080 when downconverted? I don't know. I have projected 720HD (from a HD100) on a big screen and it looked great, nobody went up to me and said "That's not 1080, I hate it!".

[Mike Halper]

Alex, thank you for taking the time to give a thorough response and insightful feedback. Yes, I'm well aware of the data rate differences between Ursa Mini and the LS300. That's another reason why I'm still going back and forth. Love the lower data rate of the LS300, while at the same time would like to have at least the 10-bit ProRes and/or 12-bit compressed raw of the Ursa Mini. Both aspects definitely have their pluses and minuses. My computer is power enough to handle the raw of the Ursa Mini, so that's not a concern at all.

I'm not too concerned with the cost though. If there weren't so many complaints about the Ursa Mini (magenta, FPN, cross-hatching, horizontal black lines, audio issues) I would have bought one already. I've been waiting months for these issues to be resolved by BMD, but they haven't so I'm now considering the LS300, which actually has been on my radar for over a year. I have everything I'd need to get an Ursa Mini up and running (except for the camera and CF2 cards). For the LS300 I'd need to figure out a battery plate situation for my Gold Mount batteries and probably a more robust handle and maybe hack the XLR ports/handle connector to work. I would maybe buy an extra handle and take it apart to make my own XLR input box or something. And also the SD cards to record to.

My absolute main concern is image quality and how well the camera performs overall. I've heard the complaint about it feeling cheap, but that's really the only complaint I've heard. Does it feel cheaper than a GH4 or any other plastic camera? I haven't ready anywhere on the net about anyone having theirs break, but there's at least 2 reports of an Ursa Mini breaking, so maybe it's actually not an issue at all?

In the past I've owned a Panasonic HVX200, GH1, GH2, AF100 and a Red Epic Dragon. I don't want to be an owner of a really expensive camera anymore. I want a camera to shoot some of my own projects and maybe some projects for others. If they're happy with using my camera, whether it's an LS300 or Ursa Mini, then great. But if they want something better like an Alexa or a Red, then it would need to be rented. Here in LA though, there are so many Red owner/ops that it's difficult to get those jobs. And I've never seen or heard of anyone shooting anywhere around here with an LS300, so that could hurt me. I don't see a lot of BMD shoots happening, but they do happen. So it's a conundrum. The JVC seems more reliable and the Ursa Mini seems like a crapshoot.

William, that's exactly one of my thoughts. In the end, the audience that you really care about watching the film won't know the difference between 8-bit, 10-bit, 12-bit, or AVCHD, ProRes, raw, or digital vs film, or in some cases 4K or 1080p when they watch it streaming over the internet. I totally get that. Many people prefer 1080p from Alexa over 6K from Red Epic Dragon. That's why I'm not hooked on the resolution, codec, etc. I will say I see a difference in quality with footage from the Ursa Mini and footage I've seen from the LS300. I like a lot of the LS300 footage I've seen, but the Ursa Mini footage at it's best, looks better to me than the best LS300 footage I've seen. There's bad footage from both, just as there's bad footage from any camera. The problem with the Ursa Mini is all the issues I've read about that can creep up at the wrong time and ruin the shoot or project, and that's a big problem.

So does anyone have any narrative work they could share with me so I can see some more narrative projects shot with the LS300? Preferably at 4K? Most of the JVC footage I see online is only 1080p, even if it was shot 4K. Has anyone tried VFX work with LS300 footage? Any extreme grading cases? (I'll download the footage from JVC's website, but I don't think it's J-Log) I would love to see what everyone is doing with these to see if it's something that will really work for me. I don't really hear anyone complaining or saying they're dissatisfied with the camera, so I'm certainly leaning in the direction of the LS300 right now. I think I just need something that gives me that final push.

Thank you.

[Noa Put]

and probably a more robust handle

The handle on the jvc is very solid and robust, that together with the lensmount are the strongest parts of the camera, the weakest part by far is the lcd screen as it feels very flimsy and I have seen reports of it starting to develop play in the hinges after longer use.

[Stephen Brenner]

BMD just released beta 4.0 update for the Ursa Mini 4.6. It addresses some of the remaining issues
of the camera.

Originally Posted by Mike Halper

I have been considering the JVC LS300 for some time. It's really between the Ursa Mini 4.6K and the LS300. I'm very concerned about the image quality, reliability and QC issues with the Ursa Mini (FPN, magenta, crosshatching, black lines, audio, etc), so am reluctant to go that route. And at about half the price the LS300 is very attractive.

I would mainly be using it for narrative projects (shorts, commercials, music videos, maybe even a micro-budget feature). The problem is that it seems very rare that anyone is shooting with the LS300 in narrative situations, so I don't really know what to expect. I can only find 3 shorts online, the 2 Heist shorts and another. And no reviews about using the LS300 in narrative shooting.

I'm sure the LS300 is great in documentary and corporate situations as I've seen plenty examples of those. But I think narrative situations can push a camera to different extremes and in different ways than a documentary or corporate shoot can, especially when doing grading and VFX work in post.

Does anyone have any narrative projects shot with the LS300 that they could point me to view? It would be a great help in my decision-making. And any feedback and reports of your experiences of shooting with the LS300 in narrative projects would be helpful as well. Thank you.

[Lee Powell]

Originally Posted by Mike Halper

In the past I've owned a Panasonic HVX200, GH1, GH2, AF100 and a Red Epic Dragon.

I produced some of the most popular hacks for the GH1 and GH2, and helped develop the Leeming LUT for J-Log1 on the JVC LS300. I've also owned an AF100 and have made extensive use of high bitrate hacks on Nikon D5200 and D800 cameras. From that perspective, I can make some comparative observations on image quality.

In H.264 video encoding, there's always a trade-off between single-frame image quality and inter-frame motion picture quality. We were able to explore this at 1080p bitrates up to 150Mbps on the GH2. The best All-I-frame hacks require massive amounts of bitrate to encode each frame individually at the highest possible quality. With inter-frame motion tracking, you can greatly reduce the bitrate used in B-frames, and devote more bits to encoding higher quality I-frames at any given bitrate. Consumer grade H.264 encoders exploit this too far, with gaps between I-frames of a half-second or more. With my GH2 hacks, I used a very short GOP of just two B-frames between each I-frame, optimizing for both image quality and motion tracking at the same time. Where you can see the difference is in smooth inter-frame transitions of sharply detailed moving objects. I have yet to see another H.264 camera produce better 1080p motion picture quality than a well-hacked GH2.

Where the GH2 suffers, however, is in its limited useful ISO range and crappy shaddow details. You can't alter its REC709 gamma curve or adjust its black level pedestal, leaving you stuck with crude chroma encoding at the bottom of its dynamic range. Nikon DSLR's are significantly better in this regard, with higher ISO ranges and customizable in-camera tone curves. In practice, I use the GH2 for sunlit outdoor shooting and the D5200 for dimly lit indoor scenes. They are both virtually free of moire and the high bitrate hacks eliminate compression artifacts. The REC709 color palettes also look very good straight out of the camera.

The AF100 was disappointing in many ways, and serves mostly as a starting point for evaluating the LS300. While it did have an adjustable pedestal, its REC709 tone curves were non-customizable and its super-white highlight handling was gross. Of course, it couldn't be hacked, and its H.264 compression artifacts were horrendous with fast moving subjects. But it had all the nice pro features the DSLR's lacked, like ND filters, XLR audio jacks, and a waveform monitor. But compared to hacked GH2 and Nikon cameras, the image quality was pathetic and its dynamics were straight out of the camcoder era.

Now we have the LS300, a true successor to the AF100, lacking only its waveform monitor. While it's not hackable, its 50Mbps @ 1080p and 150Mbps @ 4K bitrates are high enough to handle all but the most highly detailed scenery. ISO range is pretty good and in REC709 mode, both pedestal and gamma can be adjusted. The great leap forward, however, is J-Log1 mode, which eliminates all in-camera settings other than white balance and highlight exposure. It produces well-balanced dynamics that are pale and modestly flat, not as extreme as the log curves used in other cameras. It also makes Panasonic's GH4 V-log L debacle look like a self-inflicted trainwreck. OTOH, while JVC nailed its log firmware update, it dropped the ball with its funky LUT package, leaving users to eyeball tweak their J-Log1 footage. That's why I requested an LS300 LUT from Paul Leeming that matches the look of the GH4 in Cinema-D mode. (It also intercuts well with the GH2.)

With LS300 J-Log1 footage, cinematic vs video quality is almost entirely a matter of post-production esthetics. Using the footage straight out of the camera is no longer an option, even with a LUT you still need to sharpen it to bring out the details. If you shoot in 1080p, I recommend the 422 codec, it does make the gradients smoother, avoiding a major pitfall of the GH4. With 4K footage, you need a very sharp lens, because the camera adds no sharpening at all and the encoder will neglect edges that don't appear sharp to begin with. Most of the edge contrast you see in consumer grade video is courtesy of artificial sharpening and compression artifacts. Also, with J-Log1 there is no longer any excuse for overexposing your highlights, you have zebras and can see the highlights spike right in the LS300's histogram.

One final observation - the LS300's VSM is literally a dream come true, a perfectly parfocal, fixed aperture electronic zoom using any prime lens! Too bad the optical image stabilization works only on the latest Panasonic lenses...

[B.J. Adams]

@Mike Halper I had your same dilemma, choosing between the Ursa 4K and LS300, as they were in the same price range. I ended up getting the LS300 because I too heard a lot of issues with QC of the Ursa. Not to mention the added overhead cost of batteries and media with the Ursa. The Ls300 is not perfect, but it has advantages... I just had a shoot last weekend where the action was happening in a very small room and all was tight and crampy. Managed to get all the angles I wanted as the LS300 is rather lightweight and small. Don't think I could have done it with the Ursa

[B.J. Adams]

@Alex Humphrey if you are NOT doing work for film festivals, but your main medium of showing off your work is youtube/vimeo, does the 8bit limit matter after all? I did some experimentation myself, and find that even using 150mb/s compression is overkill. I'm getting good results in my eyes using the 70mb/s compression. What's your opinion on that?

[William Hohauser]

I'll jump in here with my opinions and experiences.

The final result of bandwidth and compression rates is so open to interpretation and needs that it's really hard to quantify it. What we see and what is seen by the audience (and I'm including sub-conscious effects) can be two different things. Yes a shot of a duck on moving water looks worse at 24mbps than 50mbps, than 70 mbps than 150 or 4K ProRes. Are you making a sit down interview doc or a handheld action film? The doc might work perfectly fine at 24mbps even when transferred to the big screen. A handheld film could work at that rate but would probably be better at 70 and above. With the difference between 10 and 8 bit recording, a panorama of the skyline with a clear blue sky might work better at 10bit than at 8bit but sometimes it is the camera itself that could cause a banding problem.

[Earl Thurston]

Originally Posted by Mike Halper

If there weren't so many complaints about the Ursa Mini (magenta, FPN, cross-hatching, horizontal black lines, audio issues) I would have bought one already.

It seems a bunch of us are in the same evaluation mode between these cameras. I echo that statement about the URSA Mini -- it would be a no-brainer if there weren't so many complaints. The LS300 sounds like a safer choice. The biggest problem for me is that I've been unable to find one to demo in the Vancouver, BC area, whereas the URSAs are everywhere.

[B.J. Adams]

BlackMagic need a couple more years to finetune the Ursa and make a good name with indie makers... just like RED did all those years back... in the meantime, if i start making thousands of $ from video production, I will get one :) :)

[Alex Humphrey]

Originally Posted by B.J. Adams

@Alex Humphrey if you are NOT doing work for film festivals, but your main medium of showing off your work is youtube/vimeo, does the 8bit limit matter after all? I did some experimentation myself, and find that even using 150mb/s compression is overkill. I'm getting good results in my eyes using the 70mb/s compression. What's your opinion on that?

I would say 8 bit is fine for youtube and vimeo without question. IF you plan to go to Bluray or DVD or higher bitrate (paid download 1080p/4K) and have done much color grading/effects etc, you may wish you had shot at 150mbs to start with. But I only say that not having tried the 70mb/s settings. Since I was planning on the Digital Bolex or Ursa Mini by comparison the JVC's 150Mb/s on cheap cards and edits easily with a lot of wiggle room in post (not as much as raw of course but still tons more than HDV or Nikon DSLR) I just shoot the 150 Mb/s. The Digital Bolex for instance was 1080p and 512g was about 1.5 hours of recording. Looks phenomenal by the way. I still may track down a used Digital Bolex in a 1T in a year, but I missed my opportunity to custom order one plus their monochrome one (BW sensor). So temper everything I say with knowing that I was looking at 12 bit raw 444 art house indie cameras.

Oh and I do occasionally see 8 bit sky banding on my JVC LS300. Not terrible, I doubt anyone could see it after Vimeo and Youtube step all over it with their compression.

[Noa Put]

I shoot 70mbs 4K all the time now for speeches and church ceremonies at weddings, there is not much movement and the codec is not stressed so I"ll gladly take the benefit of getting more then twice the recording time on the same cards while visually retaining the same IQ. Pixelpeepers might see a difference compared to the 150mbs bitrate but I don't. Ofcourse I hardly do any grading on my footage but if I would I certainly would choose 150mbs over 70mbs, the same applies for footage with lot's of movement and fine detail, as soon as I would shoot a scene that would stress the codec or when I would push the footage in post it's an obvious choice to use the highest available bitrate.

[B.J. Adams]

I used 70mb/s in my first shoots as I was still experimenting with SDXC cards, and was buying the cheapest options, so the LS300 refused to record in 150mb/s! I did no have much problems as the final output will be youtube, so I was lucky! But hey, I'm still learning, so I have a good excuse!

Now I found some inexpensive cards that work at 150mb/s so I'm stocking on those (and will use the other cards for audio)