A comparison of the Sony EX1 and the Canon XL-H1
 

Hello all,

So I have been researching a lot about the Sony EX as a good friend of mine has one on preorder for our production business. I have the Canon XL-H1 (and absolutely love it) and have begun contemplating the differences between the cams. I plan on doing a feature film around the summer of next year, and am looking at purchasing certain "upgrades" for my camera soon, including the Flash XDR HD-SDI recorder and the P+S tehcnik mini35 compact or the upcoming new LetusXL Extreme for the XL-H1, along with Primes. I've shot with this setup before, capturing fully uncompressed, and the footage is absolutely amazing.

But now this new Sony camera is very intriguing and I can't wait to test it side-by-side with my XL-H1. How would you think they would compare? The specs do seem compelling, but I can't quantify how much (if any) better overall it would be against the H1.

I guess I'm just thinking out loud. Any thoughts between the 2 cams?

Does the XL-H1 have embedded audio in its SDI/HD-SDI? I used it to feed a Sony Anycast via SDI but I had to use a SDI audio embedder and loop the audio from the XL-H1 line out to the SDI embedder. So that might be a problem for you.

Sony PMW-EX1 uses embedded audio in the HD-SDI (I guess in SDI also but haven't tried).

Unfortunately no it does not.

The XL-H1 doesn't have embedded audio in HD-SDI. We have always recorded audio separately and married in post. But the new Flash XDR will solve all this.

I just know that making use of the HD-SDI output with the XL-H1, along with a great adapter like the P+S technik and great glass produces stunning results. I have stills of the setup we used on my website, and will be posting some RAW clips of the footage there as well soon to show what the camera is capable of (this is the same footage that B&H is using in their store to demo the camera with). I'm curious how the HD-SDI looks like out of the Sony EX. Early tests from Barlow Elton have shown to him that results weren't nearly as good, but I would think more testing needs to be done.

Please share your footage with us. Really looking forward to see it!

Alright, I've compressed an H.264 for my website of some quick clips that you guys can check out. It is in the portfolio section, under "Uncompressed Footage."

I will try to get a short couple second clip up for download so you can see the video in full quality.

The clip posted is from a film trailer I shot back last spring. It should give an indication of what the camera is capable of when taking advantage of its capabilities.

Well if you were just shooting 1080i material I would say your H1 would be fine. If you are talking 24p however then I think the EX1 has a big advantage in clean details. High detail isn't as important as clean detail and so far from what I have seen the image is very clean and natural looking. Dump down the edge enhancement slightly and you will have some mini Red looking footage.

I have always loved the look and style of Canon video cameras but I find myself getting very tempted by the EX1.

Yes, I too am getting tempted. Though I'm a huge fan of the XL-H1, the Sony is getting very tempting. Even though I love Canon cameras, I have no brand loyalty, and will use whatver is best. I'm most curious as how the 1080p will compare from the EX to the 1080p of the XL-H1.

The 1080p out of the H1 is very tack sharp, especially out of the HD-SDI. I recently did a shoot with both the XL-H1 in 1080F24 and the HVX200 in 1080P24, and I ended up not combining the footage due to very noticeable softness of the HVX200 footage. But it has lower res chips. With the EX being full raster, it'll be interesting to compare.

My gut tells me it really won't be much of a difference, if any though. Though there have been tests showing that the F modes degrade resolution, the image remains visually as sharp as 60i footage I've shot with the camera. Don't know to sure what Canon is doing here, but it works, and it works well.

I'm going to see if any film houses here in NYC are interested in taking a quick RAW clip from this shoot and output it to film. I know I talked with someone at DuART who said they were interested.

But back to the EX and XL-H1 :)

My guess would be that the most noticeable differences will be in terms of depth of field and low-light response. I'll have an EX1 for testing soon but haven't been able to line up anyone with a Canon HDV camera for comparison - if you have one in the Sacramento area send me a PM.

Thanks for the clips - I too am at a crossroads as to what camera to invest in. Can I ask with the canon uncompressed footage on your site - did you color, grade or do anything to this or is it straight from the camera and compressed to H.264?

I would love to get my hands on a few seconds of uncompressed footage to throw in and grade to see how far you can push it...for me that’s the ultimate test, how much leeway you end up with in post.

Of course when you throw in what appears to be a massive % of faulty EX1's with the vignette issue - the canon may well end up as the only choice unless sony fix the issue ASAP!

Tonight we shot a bit under the lights at a high school soccer game just to see how the EX1 would work in low light.

Amazing well.

We played the footage back on an JVC Reference monitor via SDI off of the EX1. Very sharp. You definately can see the difference between 1280x720 footage and the 1920x1980 stuff. The 720 60P was very fluid. the 1080P stuff has so much detail. Simply amazing.

I also played back the footage running a parallel connection off of the SDI out on the JVC monitor over to my HDLINK>DVI to my 24" Dell 2405 1920x1080 monitor.

As you know, it's not a reference broadcast monitor, but looked good via SDI off of the EX1. The black level was better on the reference monitor. The reference monitor had a very clean professional look.

One issue noted with the SDI>HDLINK>Dell 2405 was the interlaced signal was showing itself (. So any time there was vertical image movement I could see the 1080i lines.

Does anyone know why this is happening with the Dell? I do not see any way of deinterlacing it. The JVC worked fine.

Well, according to my preorder, I should have my EX within the next week or so, then I can do direct side-by-side comparisons between the XL-H1 and the EX1. Hopefully, I will post soon in regards to that.

The footage is RAW from the camera. I have done some grading on it before, and it has amazing latitude to push the image. I will try to get a quick couple second clip up for download soon.

I really am pleased with the Canon and it can provide results that are more than acceptable for a film transfer ... just as long as you shoot knowing what it's limitations are ... I guess that goes with any camera.

Though I think resolution will be very close between the two cams, I am interested in the dynamic range and low light differences. Also, what HD-SDI capture will look like compared to the Canon. Has anyone confirmed that the HD-SDI from the Sony EX is 10-bit? That would be very interesting ...

Re: Soccer footage...

Sounds good - would luv to see some footage, world REALLY luv to get a small sample of untouched straight from the camera footage of such a shoot.

One thing I thought the EX1 was 1080P - why would interlacing be an issue?

Well I would think the reason being that 1080P is sent down through the HD-SDi as 1080i60 with the progressive embedded with a 3:2 pulldown. I don't think the HDLink does any sort of reverse telecine, so you are probably seeing the footage as interlaced on your progressive monitor.

So I went to Able Cine Tech here in NYC yesterday to check out the EX 1 and all I can say is WOW. Sony has really gotten their act together on this cam. It was hooked up to a 1080p monitor and the 108024 out of this camera is gorgeous ... and very very sharp. An HVX200 and the XH-A1 were also next to it, and while the XH-A1 looked close resolution wise, the HVX just looked soft.

Of some notes:

1. The LCD screen is in a league of its own ... it certainly is much much sharper and more vibrant than anything I've seen from these types of camcorders. Blows the LCD screen from my XL-H1 out of the water. It blows the LCD screen from the HVX200 even farther.

2. Menu system seems so much more easier to navigate through than the HVX. Seems very intuitive ... has image adjustments on the same level as the Canons ...

3. Really, really nice lens. Great controls and feel.

4. Image quality is just superb. I believe the 1080p out of this camera is the sharpest of the bunch. The XH-A1 looked really close in sharpness, but I have to assess it further ... but no doubt the EX 1 is probably sharper.

5. Wow ... the 1/2 sensors make a difference!!! I exposed both the EX 1 and the XH-A1 properly, and the XH-A1 needed the gain boosted +6 in addition to a full open iris. The EX-1 was at 0db gain, and the iris wasn't even full open!!!

6. And of final note, the guys there confirmed that the HD-SDI is indeed 10-bit. Looks like capturing this thing to a 10-bit codec like ProRess 422 HQ with an AJA IO-HD would yield incredible results.

I am going to bring my XL-H1 there next week and do side by side comparisons along with captures from both cameras. I really really love the XL-H1, but this Sony Cam is getting more and more compelling.

Michael you are right about the HDLINK HD-SDi as 1080i60 with the progressive embedded with a 3:2 pulldown.

I was hoping there may be another solution short of buy a different monitor.
The JVC reference LCD does not show the jagged vertical motion artifacts.
But, I ran straight SDI out to the JVC; although, this should also be 1080i60 with the progressive embedded with a 3:2 pulldown.


The JVC LCD has faster response for interlace, or is it deinterlacing, I'm not sure?

I also checked out the camera at Abel in NYC yesterday and second Michael Galvan's observations. I tried an on-the-spot vignette test throughout the lens range and didn't see any dark corners on the monitor, but some reports say it can only be seen on playback.

On an off-topic note, they also had the soon-to-be-released Litepanels Micro on display, which runs off four double-A batteries. It's got a very small form factor, and if it's got sufficient throw for run & gun interviews and B-Cam work, it seems well worth the $330.

You won't see it unless your monitor shows the overscan region.
Try SDI out for full raster.

I love the parroting that salespeople and marketing reps do. These people don't do any testing at all. Sure, it's probably 10bit from a standards compatibility perspective, but I've tested the output with the Kona and the SheerVideo 10bit codec, (basically uncompressed 10bit) and it doesn't have the extra smooth gradations you would expect with true 10bit output. It looks just like the H1's output, which is quantized to 8bit and sent out with two bits set to zero.

That's interesting, because I specifically pointed that out to them and asked them if the HD-SDI of the EX 1 does the same thing, but the sales guy (who's a cool guy there .. they're all cool and nice) went to ask their Sony rep and they confirmed it as true 10-bit.

Well, based on what Barlow has said, I'm questioning it is filling all 10 bits.
If it's not showing as a benefit, it's probably filling only 8 bits.

Is there some sort of software that can extract the stream information and display if it's filling all 10 bits?

Good to know that you've done some definitive comparison, Barlow. This question was raised early on about the EX1's HDSDI pipe.

-gb-

Sounds exciting Michael-we will be eager to hear of your findings. You had some really impressive footage from your H1 on your site- if you can top that, well....

Thank you Hayes ...

I'll try to see what I can do in terms of what I can record. There is a setup with a lot of colors, but everything will be mostly static.

Mike, do you have access to Canon manual 16x or 14x lens, if you do, please take same shots, from my testing 14x produces clearly better color, with more color contrast and similar resolution compared to the stock lens on H1.
Regards.

Hey ... sorry, currently I only have the 20x HD stock lens. Maybe they may have one for me to test ...

Barlow did you try to color correct the footage to see how much each version would band? They may in fact both look the same when looking at it straight on but it is when to push the CC to extreme levels where the 10bit would shine. Remember 10 bit video still has to display on your monitor as 8 bit.

No, I haven't had the time yet. Will give it a try but I did view it on a CRT VGA display and the video card supports 32 bit color.

That isn't 32 bits per channel but the total bits.

10 bit video has 30 bits total. If it had an alpha channel it would be 40 bits.

The 32 bits on computer monitors is 8 bits per RGB and Alpha. so it is only 24bits for the actual color or 8 bit per channel.

The video display graphics card is also a limiting factor here since they usually only work in a 8 bit color space. This is why having a HD-SDI based editing system and a good 10 bit HD monitor are so important.

Yes, I would like to have a real 10 bit HD-SDI montior. I wil say that the banding was much less noticeable on the CRT, but when viewing higher bit depth photos on my LCD, the banding is still way less noticeable than a regual 8 bit pic. And btw, it wasn't bad banding at all, just something I'd expect from 8 bit video.

Thomas, do you have a way to view 10 bit video properly? I'd be happy to send you a couple clips to peruse.

Barton,

I'd love to see some footage out of the HD-SDI and have a look at it, even a single uncompressed frame would be fantastic...

cheers
paul

The only monitor I've used that displays 10 bit color is the Sony BVM L230. There are other manufacturers that manage this- but I've never seen their monitors for real.

The BVM L230 is MSRP $25000 USD. That's a "base" unit. I think you'd have to add an HD SDI i/o board.

I don't think that the earlier BVM CRT monitors displayed 10 bits per channel, but I've never really used or considered one, so I don't know for sure.

No computer monitor displays 10 bit color. In fact a lot of HD LCD computer monitor panels don't even display 8 bit per channel.

(I have two Gateway FPD2485SW monitors and I can assure you they display dithered 6 bit color. There is a clear difference when comparing to my Apple Cinema Display (aluminum 23"))

So... unless you've examined the image on a high end monitor it is hard to be definitive.

I just so happens I have looked at the image from the EX1's SDI port on the BVM L230- but I was looking at a "straight" uncorrected image live from the camera. I simply can't draw definitive conclusions based on that.

The best way to determine the nature of the footage is to capture some test patterns and gradations using a 10 bit or better codec, ideally uncompressed, and then analyze it using the scopes in Color or Lustre.

Sorry to reply to me.

I should say that even if you use Color or Lustre you need to make sure that they are set up for high precision color.

Also, the color correction tools in the NLE (Final Cut or Avid or whatever) aren't likely to give you good results on a test like this. Although they'd certainly break the footage for you- 10 bit or 8 bit, they won't indicate why or how the footage broke.

It also occurs to me that compositors like Shake should be able to show you the differences readily as well. In fact if you really know what you are doing you might be able to get more information out of Shake than Color. Color will give you information much more readily though.

Finally... I mentioned monitoring on the BVM series production monitors, but I neglected to mention Digital Cinema projectors that manage DCI spec color. Sony showed such a beast working at 4K for the paltry sum of $100000 USD.

So.. if anyone has BVM-L230, an ioHD and an EX1 to loan me I'd be happy to put this one to rest forever.

Unless your material is going to be projected in a DCI spec theater then 10 bit or not is just a matter of headroom, not necessity.

You'll get a lot of that headroom just by using a 10 bit codec for post and the right tools correctly configured.

(By the way FCP is set up by default to use 8 bit color for everything- so change it if this matters to you.)

Not yet I don't. I just started a job where we use the F900, Red and a SONY Xpri editing system but I'm not if they will let me bring in material to goof around with yet. The sad thing is I think they are also using 8 bit monitors so even with all this massive gear I'm still not sure if I can help you there.

One way to tell of course is to take two clips or even two still images and use a levels filter to really crush the min and max levels. If you have After Effects this test should be pretty easy.

1. Load the 8 bit and 10 bit video into AE.
2. Set AE to work in 16 bits/channel. If you do not even the 10 bit video will work as 8 bit.
3. Load one clip into one layer and the second clip into another layer.
4. Put a levels filter on the top clip. Really bring in the min and max points to push the colors to extreme levels.
5. Copy the filter and paste it onto the bottom layer.
6. Now turn the top layer off and on to view the difference between the two.

The 10 bit video if it really is 10 bits should have held up much better while the 8 bit video will now start to look like a 256 color image.

If you do not have AE (I beleive you use FCP if I remember correctly) You can do the same thing in FCP or pretty much any other program including photoshop for stills.

I did capture in true 10 bit uncompressed (lossless Sheer codec--still huge data rate--about 62 MBs), but that doesn't open in Color so I transcoded to ProRes HQ. I shot the sky at dusk (almost dark, actually) and the sky had a smooth but deep ramp from warm glow of the sun just dipped below the mountains to dark blue towards the sky above.

If you want to see it shoot me an email.

Might be a problem there...

:)

I'd love to see the footage. Drop me an email.

If you can I'd love to get OpenEXR float images (16 or 32bit) Motion will output those for you from any footage FCP can read. I don't need much, 1 frame would work.

Or, you could post a 2-3 frames as a 10 bit uncompressed quicktime or about 4 seconds as ProRes HQ quicktime to the forum.

In a program like Shake, you can subtract one image from the other. If you get anything other than pure black, you know the images are different at least.

Ok, here is a 3 second 10 bit ProRes capture. (transcoded from raw 10bit SheerVideo..essentially uncompressed)

Please work your post magic and see if you can discern its bit depth. And I really really hope I'm wrong about it. :)

I can only leave the clip up for a few days.

http://homepage.mac.com/mrbarlowelton/FileSharing1.html

Sorry, been super busy the last couple days, but I will try to post some of the footage I shot at AbleCineTech when I can from both cameras.

I copied the contents of the card onto my HD from the EX1, but can't seem to open the movie files. I can only read them in the reader software that Sony provides. Can anyone tell me what I need to open them in the latest Final Cut Pro 6.02?

Hey Greggor ... I'm gonna make you an internet sensation just like we talked about ... lol

I am not a mac guy, so I don't know what you would need for FCP. But I have a guess. The files from the EX1 are mp4 files. You use the clip browser to export the footage as an mxf file. It is just a simple re-wrapping function that is extremely fast. If FCP will edit mxf files, then you are set.