Sheraton Wild Horse Pass (Sony RX10 Firmware V2.0 XAVC-S) on Vimeo

Sheraton Wild Horse Pass and Rawhide, Phoenix Arizona. First project after V2.0 firmware update.

The RX10 AVCHD codex is exceptional and holds up well in post. The new XAVC-S is equally impressive and may perform better at high ISO. This short project includes several clips (blacksmith shop) at ISO 6400.

Looking forward to seeing what other improvements the RX10 XAVC-S codec provides.

Sony DSC-RX10
Benro A38TB Fluid Monopod & S2 Head
Marumi Polarizer

may perform better at high ISO

It doesn't, at least not on mine, 50mbs xavc-s is actually a little more noisy then 28mbs avchd. From a normal viewing distance to a tv I don't see any difference but when I blow up framegrabs in photoshop 28mbs avchd looks cleaner, just a little bit but I do see a difference.

It doesn't, at least not on mine, 50mbs xavc-s is actually a little more noisy then 28mbs avchd. From a normal viewing distance to a tv I don't see any difference but when I blow up framegrabs in photoshop 28mbs avchd looks cleaner, just a little bit but I do see a difference.

Yours is a much more scientific test Noa. After shooting the blacksmith scene, I noticed my reference photo was ISO 6400, and thought "That's gonna get cut.", but ended up happy with the result.

After shooting the GH2/3 for years, usable ISO 6400 is difficult to accept.

I downloaded the original m4v file. Nice, as always Dan. I did notice macro-blocking particularly at 23 to 28 sec. and I also noted that the data rate was only 10181 kbps.

Curiosity got the better of me so I headed out and shot some comparative shots of poorly lit forest away in the distance across water where I have had problems with the RX10 AVCHD codec previously. The day was clear and sunny but the forest in the shaded valleys had a distinctly bluish cast as it does most often. I set the RX10 to movie mode, 60p, f4 (ND filter), 125th and ISO 125. I also shot the scene with my EX1 knowing these areas of shaded forest would be quite noisy, in order that I had an independent known standard.

When rendered off using the Sony Internet AVCHD Vegas Pro 12 template with no post manipulation the EX1 footage was as expected; noisy and blocky however the AVCHD and the XVC-S footage was not noisy nor was there any visible macro-blocking.

I also rendered off to the Sony Vegas Pro 12 xdcam 35mbps 29.97 frame rate and as expected all three samples were clean and essentially noise free.

I then applied sharpening at the default setting in Vegas Pro 12 and rendered off to xdcam again. The EX1 footage was quite noisy as expected, while the AVCHD and the XAVC-S footage was just fine.

The sharpened footage when rendered off to the Sony Internet 30p template in Vegas Pro 12 the EX1 footage was a mess while the XAVC-S footage was remarkably good and the AVCHD almost as clean but some slight macro-blocking appearing.

I conclude that the XAVC-S is as expected, more robust than the AVCHD when manipulated in post and that when rendered off to XDCAM VBR 35 mbps 29.97p there is no visible difference; they all look just great.

The difference between the XAVC-S and AVCHD footage was not as great as I expected. I wonder if Sony also did something else with the firmware update they are not telling us about. Wouldn't surprise me in the least. In any event I shall continue shooting using the XAVC-S codec, for now.

After shooting the GH2/3 for years, usable ISO 6400 is difficult to accept.

Usability of high iso levels was not what I was referring to, my reaction was because you said it may perform better at high iso and it doesn't. I saw another user mentioning he saw much improvement in less noise on high iso and that's why I did a test, from what I see on my camera that is not the case but it's just the opposite, avchd had less visible noise. I"m sure the higher bitrate on the xavc-s codec will not break up as easily but in terms of low light performance and noise levels it's not better then avchd.

I downloaded the original m4v file. Nice, as always Dan. I did notice macro-blocking particularly at 23 to 28 sec. and I also noted that the data rate was only 10181 kbps.

Curiosity got the better of me so I headed out and shot some comparative shots of poorly lit forest away in the distance across water where I have had problems with the RX10 AVCHD codec previously. The day was clear and sunny but the forest in the shaded valleys had a distinctly bluish cast as it does most often. I set the RX10 to movie mode, 60p, f4 (ND filter), 125th and ISO 125. I also shot the scene with my EX1 knowing these areas of shaded forest would be quite noisy, in order that I had an independent known standard.

When rendered off using the Sony Internet AVCHD Vegas Pro 12 template with no post manipulation the EX1 footage was as expected; noisy and blocky however the AVCHD and the XVC-S footage was not noisy nor was there any visible macro-blocking.

I also rendered off to the Sony Vegas Pro 12 xdcam 35mbps 29.97 frame rate and as expected all three samples were clean and essentially noise free.

I then applied sharpening at the default setting in Vegas Pro 12 and rendered off to xdcam again. The EX1 footage was quite noisy as expected, while the AVCHD and the XAVC-S footage was just fine.

The sharpened footage when rendered off to the Sony Internet 30p template in Vegas Pro 12 the EX1 footage was a mess while the XAVC-S footage was remarkably good and the AVCHD almost as clean but some slight macro-blocking appearing.

I conclude that the XAVC-S is as expected, more robust than the AVCHD when manipulated in post and that when rendered off to XDCAM VBR 35 mbps 29.97p there is no visible difference; they all look just great.

The difference between the XAVC-S and AVCHD footage was not as great as I expected. I wonder if Sony also did something else with the firmware update they are not telling us about. Wouldn't surprise me in the least. In any event I shall continue shooting using the XAVC-S codec, for now.

Nicely done John! Thanks for taking time to perform these comparisons.

Sheraton Wild Horse Pass (Sony RX10 Firmware V2.0 XAVC-S) on Vimeo (https://vimeo.com/105053157)

Sheraton Wild Horse Pass and Rawhide, Phoenix Arizona. First project after V2.0 firmware update.

The RX10 AVCHD codex is exceptional and holds up well in post. The new XAVC-S is equally impressive and may perform better at high ISO. This short project includes several clips (blacksmith shop) at ISO 6400.

Looking forward to seeing what other improvements the RX10 XAVC-S codec provides.

Sony DSC-RX10
Benro A38TB Fluid Monopod & S2 Head
Marumi Polarizer

Hey Dan, Very Nice Video! Actually all of your videos are so well exposed, and what seems the right amount of sharpness, Would you care to share any secrets as to why your videos have that look about them?

Do you shoot with a Polarizer most of the time? I'm new with the RX-10 so I'm learning, what settings do you like on your RX-10? Great stuff!

Hey Dan, Very Nice Video! Actually all of your videos are so well exposed, and what seems the right amount of sharpness, Would you care to share any secrets as to why your videos have that look about them?

Do you shoot with a Polarizer most of the time? I'm new with the RX-10 so I'm learning, what settings do you like on your RX-10? Great stuff!

I'm happy to hear you like the look Leon.

My usual settings are:

Exteriors: 1080/60p, S Mode, Shutter 125, Zebras 100%, Vivid -1 Contrast, +2 Sharpness, -0.3 Exposure, Polarizer.

Interiors: 1080/60p, S Mode, Shutter 125, Zebras 100%, Standard -1 Contrast, +2 Sharpness, 0.0 Exposure, No Polarizer unless shooting through windows.

Dark Interiors: 1080/24p, S Mode, Shutter 50, Zebras 100%, Standard -1 Contrast, +2 Sharpness, 0.0 Exposure, No Polarizer. (24p and Shutter 50 allow better low-light performance).

Auto WB performs well on the RX10 and RX100M3.

With these settings, little if any post editing is required for the look I prefer. Your mileage may vary.

B+H and Marumi polarizers have worked well for me.

Thank you for watching, and the encouraging comments.

Dan, Thank you for sharing your settings on the RX-10! I'm curious, do you have your ISO set to Auto?
Is the S-Mode shutter priority, so you let the camera choose the aperture and ISO?

One of the reasons I like your look is it looks like video, not film! Personally I have always had a difficult time with amateur's trying to get the Hollywood film look all the time. Hollywood does it right, So many freelance videos that I have seen have that soft, and muted colors and IMO, overexposed shadows, Your style of shooting really allows one to see what the cameras sensor and lens can do.
I sometimes don't even mind "Crushing the Blacks" especially in a shot that has lots of rich color, I like doing video landscapes, I'm learning a lot, and having fun, that's the main thing.

I guess I just know what I like, If I had a client that wanted the project to look like film, then I would guess I would have to try an achieve that. Most of the clients I get don't even know what the "Film Look" is. LOL!

Again, Thank you!

Dan, Thank you for sharing your settings on the RX-10! I'm curious, do you have your ISO set to Auto?
Is the S-Mode shutter priority, so you let the camera choose the aperture and ISO?

ISO Auto (6400max).

Camera chooses aperture (ND for shallow DOF).

Good shooting Leon...

Usability of high iso levels was not what I was referring to, my reaction was because you said it may perform better at high iso and it doesn't. I saw another user mentioning he saw much improvement in less noise on high iso and that's why I did a test, from what I see on my camera that is not the case but it's just the opposite, avchd had less visible noise. I"m sure the higher bitrate on the xavc-s codec will not break up as easily but in terms of low light performance and noise levels it's not better then avchd.

Noa - I think what you're probably seeing is actually superior performance by the xavc-s codec.

Given that both codecs are receiving identical noise from the sensor - noise is actually in reality a lot of pixels changing. AVCHD being a less efficient codec would tend to lose detail in this situation and you could mistake this for less noise. OTOH XAVCs at a higher bit rate and using precoding hardware would be more faithfully preserving the noise in the original image. You often see the same thing in raw and jpeg reproduction in stills cameras.

You can probably get a higher detail result from XAVCs and while noise may be more noticeable you can probably use noise reduction software and end up with a superior result to AVCHD. Of course if you are happy with the result from AVCHD you could treat what the codec is doing as automatic noise reduction.

All supposition but generally a codec can't introduce sensor noise. If the noise you are seeing is macro blocking that would be a different matter.