I'm going to be shooting wildlife and bird video in Africa and hope to use my Canon XL H1. Unfortunately, my client does not want the footage recorded in HDV. In fact he is pushing me to use the new Sony HD EX1 camera.

Is there a portable recorder (similar to the Canon FireStore FSC-HD100 Portable DTE Recorder) that will mount on to the XL H1 and allow me to use the HD-SDI port to output uncompressed 1080/60i 4:2:2 HD-SDI or compress to a format like DVCPRO HD?

Hello and i am in the same situation, but i have a workaround. Expensive, but it's the solution that i have. We recently intergrated the AJA I/O HD into our workflow so what we do is shot directly to a G SATA drive. In the field we have a MacBook Pro, G-SATA drive, AJA I/O HD, Convergent Design SI, and a 23" APPLE Cinema Display (the connection for this is DVI to HDMI into the AJA I/O HD). When using the I/O we are not capturing audio as we go HD SDI out of the XL H1 to HD SDI in on the I/O. Since the XL H1 doesn't embed the audio, the workaround for that is the Convergent Design SI, which embeds the audio for you via firewire from the camera, then RS 422 and HD SDI into the I/O. The codec that is used is the APPLE PRO RES 422 HQ 1080i. Honestly, you cannot tell the difference between this codec and uncompressed HD. The file sizes are smaller. Hope that i was able to offer you some sort of a solution.

Thanks, Gregory. Thanks for the details to your workaround. I will incorporate this in the next production shoot I do for a customer that needs more than HDV.

However, this seems to bulky for shooting African wildlife (though most of the time I'll be on a flatbed truck/open SUV.

Any other solutions that are more compact? That is mount right to the camera?

Cineform is working on an HDMI version of what you are looking for - unfortunately still in development. Currently, their users who want better than HDV quality are recording to an Intensity-tethered laptop....

John,
I also forgot to mention that Convergent Design is working on a solution for the Canon XL H1 and according to their preliminary specs here is the price: US $4995, available Q1 08; 16Gbyte CompactFlash cards cost ≈ US $200:: When this arrives i will get this as well.
http://www.convergent-design.com/downloads/Flash%20XDR%20Brochure.pdf
I don't know if you have time to wait for this, but this is the PERFECT solution for the Canon XL H1! IMHO

John,

As far as I know...there are no options available RIGHT NOW.
the Convergent Design will be KILLER when it's available...

The idea of cabling into a Mac...with additional audio issues...is a great workaround...but like you said...not "in the field".

I've used the H1 shooting on-location recording to HDV, DVCProHD and directly into our AVID Adrenaline at 220mbs DNXhd codec. It all looks great.

Yes...HDV is 25 mbs...but it holds up very nicely. If you can wait...the Convergence will be the way to go...but if you can't...the HDV signal that the H1 is recording is one of the best (IMHO). I' ve seen other HDV tapes and have been un-impressed with the quality level. Canon says their signal does meet HDV spec...but because of the quality of lens and some (vodoo) they do in the electronics...their signal looks better in the HDV format.

All I know...is I agree.

Gentlemen, thanks for your input. the Convergent Design solution sounds exciting!

Audio is embedded in the SDI.

Audio is embedded in the SDI.

Not on the XL-H1. It is video only.

Fortunately, Convergent Design has stated that the XL-H1 is the #1 priority camera on their list in terms of Flash XDR. They even built in XLR inputs in the XDR to compensate for the lack of audio in the XL-H1's HD-SDI stream.

Michael;
Do have any information when the Convergent Flash XDR will be shipping?

Michael;
Do have any information when the Convergent Flash XDR will be shipping?

From my understanding, they said it would be shipping by NAB, so around April.

Hello everyone,

Can anyone tell me if there is any visual proof of a better quality picture when recording directly to the computer through the SDI output when compared to HDV capture to tape, then output that recording from the tape to the computer via the SDI output?

I am using Matrox Axio LE HD breakout box with PPro CS3 on a 10 or 8 bit uncompressed timeline. I find that the footage can handle the same amount of cc and effects equally as pure HD-SDI. I then export movie which creates a large Matrox AVI file. When I create a Blu-ray or reg dvd disc in Encore CS3, the quality is visually exactly the same on my 10ft screen 1080p set up. I also find that shooting vanilla in camera and then correct in post using the above mentioned method, I get a much better result and more oprions compared to in-camera presets. I know the H1 can only do 8bit. If I output HD-SDI 8bit 4:2:2 what visually do I get with 4:2:2 over 4:2:0 using the process above? I also want to mention that HDV firewire into the computer on a hdv timeline, then out put to Blu-ray does not tolerate too much cc in my experience. Some banding can occur but never in the process above. Can someone please put sense to this? I'm only going by what I see. Finally I also want to point out that vanilla capture with sharpness level to 2 to me is the same in sharpness as HD-SDI with no added noise. Thank you in advance for your input. Maybe someone can tell me how I can benefit with HD-SDI that would justify the added equipment on the field.

Troy

H1 output in HD-SDI = 4:2:2 Colorspace AND depending on the file format chosen, much less compression. (Example: ProRes vs. HDV long GOP)

H1 tape out in HDV = 4:1:1 Colorspace AND long GOP compression

But most agree that the HDV out of the H1 with proper lighting, camera movement, apeture/shutter control is beautiful. Only issues I've seen are complex detail with wild motion such as trees/bushes blowing in the wind or water with a lot of motion.

HDV is 4:2:0 o/w agree.

Oops! Pete is correct - sorry about that.

But the comparisons are still accurate.

Opps, I stand corrected.

H1 SDI - no embedded audio.

I had read a few posts that said they where using a XL H1 SDI out w/ embedded audio. May be they where mistaken and where actually using the firewire out. So, I did more research on the subject.

Also, its interesting that Canons pages on the H1 make no mention of it either.

So then what is H1 recording to tape, then out via SDi to an uncompressed timeline? Bypassing firewire, is there less compression going into the computer? I can see better results on a Blu-ray disc when compared to HDV workflow. Can this be explained?
Troy

Please forgive me jumping in but even after reading the first page - with GOP's and 4.2.2's (whatever they are) I am still unsure if the picture taken from an H1 is better from the tape or out of the SDI to a hard drive.

Is there a definitive on this one?


Thanks
Dave

I will give a crack at it. This is my guess. If I take a photo in 8 bit mode and convert it to 16 bit mode before editing and then take it back to 8 bit mode the histogram will look good w/out verticle lines running thoughout. This might be somewhat true regarding video. If the Chroma and Luninance are being subsampled to 4:2:0 to tape (compression) and then sent to the hard drive via the firewire and finally to a compressed codec timeline like pro rez or hdv, then color correcting will become very restrictive on how far one would be able to push cc or added effects. Mind you alot of effects can hide the clipping and so on by implying that it was intentional.

Now in my experience, when comparing this same footage shot on tape was instead sent out to computer via SDI (BNC Cable) to an UNCOMPRESSED 8 OR 10 bit timeline I got better results. It has a much greater tolerance with cc and effects before smooth tonal gradations are compromised. Now what is going on, I don't have a clue. Is the signal still 4:2:0 but a full 1920x1080 or even 4:2:2 1920x1080 compared to 1440x1080 4:2:0 by way of firewire.

Now on the field with a proper set up like Matrox with a hd break-out box, one can be sure its' a full 4:2:2 1920x1080 subsampled signal. Keeping this signal pure by putting it in a HD project on an 8bit Uncompressed timeline takes all the guess work out of it. You would have the best the Canon H1 can do.

Now here is the most important part for me,"final output". Taking that large data file, the Matrox AVI 8bit Uncompressed file and loading it in Encore CS3, I can produce a HD Blu-ray disc that has absolutely no banding or artifacts due to the fact that it was not compressed twice.

I hope this helps as this is only my opinion and experience with my workflow.

Is 24F out of the HD-SDI port imbedded into a 60i stream or is it 23.976 fps?

Thanks!

Here is some facts to save you from searching through several earlier posts.

1. HDSDI out of H1 is 10bit 4:2:2 @ 1920x1080 HDV is 8bit 4:2:0 @ 1440x1080

1a. 24f is output from HDSDI embedded in 1080i 29.97

2. HDSDI gives a much better grading format and effects work clip to work with.

3. HDSDI gives you true 1080p...not 1440x1080 like HDV recorded to tape

4. HDV (that which has been recorded to tape and directly recorded through the firewire port) is heavily compressed long GOP (like a DVD) format. It REALLY SLOW to render on an old computer. PERIOD. It is horribly blocky for greenscreen work. It is pleasing to the eye...but don't try to modify it!

5. Recording to HDV and then outputing HDSDI does not magically make the video better...garbage in - garbage out. Once it is recorded to HDV the show is over.

6. Is HDSDI better? YES.

7. Is HDSDI more economical? NO.

8. Should I shoot HDSDI in the field? Not unless you have a good sized crew and do extensive testing with a "foolproof" system. You can even record HDV and do a second system capture to uncompressed or ProRes and just slate each shot if you want better quality and safety at the same time.

9. Should I consider a different camera/format if 10bit 4:2:2 is what I want? YES.

hope that helps

*Applause* Thank you very much! I aware of all of that, except that it applies pulldown.

BTW, one thing that might not be clear for G1 and XL-H1 users is those cameras will output 8-bit color packaged in 10-bits. It's not 10-bit color, even though it's in an HD-SDI 10-bit stream.

*Applause* Thank you very much! I aware of all of that, except that it applies pulldown.

Feel free to repost and correct....I am exhausted from typing all that.

Good Morning Will,

Recording to tape is exactly what?
I know output from the tape via firewire is a compressed 1440x1080 4:2:0 signal. In fact in ppro cs3 the file is stated as just that.

Now, again when that same recording from tape is sent to the computer using the SDI then in fact the signal is stated as 1920x1080 and the file is much much larger, better quality and impossible to visually see the difference when compared to pure HD-SDI recording to computer before and after editing the same for both. The key in the workflow is keeping it in an uncompressed timeline.

I know all this to be true because I have tested this again and again.

My question is does the chroma and luma stay subsampled to 4:2:0 or does it in fact output at 4:2:2 via the SDI while at 1920x1080?

Pro rez is better than HDV in my opinion but not better than 1920x1080 from tape to computer via SDI (to an uncompressed timeline) Matrox AVI

Bottom line is that the higher resolusion from tape using the SDI instead of the firewire does give a much better file to edit with and produces a much much better Blu-ray disc.

Thanks

Recording to tape is exactly what?
I know output from the tape via firewire is a compressed 1440x1080 4:2:0 signal. In fact in ppro cs3 the file is stated as just that.Thanks
Correct.


Now, again when that same recording from tape is sent to the computer using the SDI then in fact the signal is stated as 1920x1080 and the file is much much larger, better quality and impossible to visually see the difference when compared to pure HD-SDI recording to computer before and after editing the same for both. The key in the workflow is keeping it in an uncompressed timeline.
You've left out the fact that the camera takes the 1440x1080 and uprezes it
to 1920x1080 when coming from tape, unlike live capture of HDSDI which
yields more resolution, better dynamic range(lots more info in the shadows),
and less compression artifacts.


My question is does the chroma and luma stay subsampled to 4:2:0 or does it in fact output at 4:2:2 via the SDI while at 1920x1080?

The luma is 4:2:2 but unfortunately it is "made up" by the the conversion.
It's kinda like dropping a GIF into a TIFF or PSD file in photoshop.
The file says 24bit color and no compression, but the GIF obviously wasn't!!

Pro rez is better than HDV in my opinion but not better than 1920x1080 from tape to computer via SDI (to an uncompressed timeline) Matrox AVI
Uncompressed is always better than a compressed format like Prores, but
like I mentioned before, you have already saved the footage as HDV and
are just converting from that to uncompressed and not actually shooting
uncompressed.

Bottom line is that the higher resolusion from tape using the SDI instead of the firewire does give a much better file to edit with and produces a much much better Blu-ray disc.
What you are seeing is uprezing of long GOP HDV 1440x1080 to intraframe
1920x1080. It could look marginally better and is obviously easier to edit
and deal with (other than the file size). I'm not arguing that. I'm just saying
that a live ProResHQ or Uncompressed capture from HDSDI during the shoot
looks MUCH better than capturing from tape in any format.

Thanks Will,

The clouds are starting to roll back.

Just one more thing, are you absolutely sure that the mini tape in the Canon cannot receive the full 1920x1080, if not then how is the camera capable of uprezing the file? I though what's on the tape is analog until it is digitized out either by SDI or Firewire. Also isn't the sensor a full 1920x1080, wouldn,t that be sent to tape?
I really appreciate this Will. Is there anything from Canon to substantiate this?

Regards,
troy

are you absolutely sure that the mini tape in the Canon cannot receive the full 1920x1080, if not then how is the camera capable of uprezing the file?
HDV spec is 1440x1080. if it captured as 1920x1080 then it wouoldn't need to uprez it.
I though what's on the tape is analog until it is digitized out either by SDI or Firewire.
there haven't been any analog cameras since BetaSP was king.
Also isn't the sensor a full 1920x1080, wouldn,t that be sent to tape?
not if they want to save it as HDV format.

Is there anything from Canon to substantiate this?
Yes. Their website and documentation.

Hello Will,

Maybe I am missing something, but it seems to me that Canon is silent on the color space other that HD-SDI out. I cannot find any detailed info on what you have stated. Can you point me in the right direction.

Thank you for your time, I really appreciate this
Troy

If it is 1080/60i HDV it is 4:2:0 in any camera.
That is the HDV standard.

http://news.sel.sony.com/hdvcamcorder/hdv_technology_book.pdf

Thanks Will,

The clouds are starting to roll back.

Just one more thing, are you absolutely sure that the mini tape in the Canon cannot receive the full 1920x1080, if not then how is the camera capable of uprezing the file? I though what's on the tape is analog until it is digitized out either by SDI or Firewire. Also isn't the sensor a full 1920x1080, wouldn,t that be sent to tape?
I really appreciate this Will. Is there anything from Canon to substantiate this?

Regards,
troyTroy,

HDV is always what the XL-H1 records to tape. HDV is always 1440 pixels wide, not 1920. It displays as 1920 during playback because the pixels are stretched to cover the same width that 1920 pixels would.

That is why you'll hear it said that HDV uses rectangular or anamorphic pixels.

What's on the tape is never analog. It's digital. And it is compressed to meet the HDV standard. I would do some searches on HDV. It is not just a vague term, but rather a very specific way of compressing a digital video and audio signal.

What comes out of the HD-SDI port is not compressed.

Finally I believe the sensor is not 1920, but rather 1440. So the camera is "uprezing" the width when the signal comes out of the HD-SDI port.

HTH ;).

P.S. I have no dobut that HDV uprezed by the camera and stored as ten bits instead of eight will give a better result than capturing w/o the uprezing. But I also have no doubt that, like Will said, uncompressed will look much better. So:

HDV captured via Firewire < HDV captured via HD-SDI <<< Uncompressed HD-SDI.

So if the sensor is only 1440x1080, then HD-SDI 4:2:2 8bit colorspace is actually uprezed and not true. I thought the physical size of the sensor was 1920x1080 1/3". Threads that I have read state the H1 as a higher rez (by sensor) than Sony or Pani, not including the EX Sony.
Also, I understood the HDV standard is 1440x1080 4:2:0.

Is this true?:

Canon compresses the signal captured by the sensor (1920x1080 4:2:2) to HDV 1440x1080 4:2:0 to tape. Therefore the signal is resampled to fake that compressed resolution and colorspace to be interpolated to stretch and fill a 1920x1080 4:2:0 colorspace when sent to computer by SDI.
If this is the case would there not be an issue with artifacts and banding? Why is the signal much better when output to Blu-ray? If the resolution is not there, then it cannot be stretched and produce a cleaner and better result. My workflow as mentioned earlier in this thread also seems to have more color info as the signal results in an uncompromised gradation in the colores from dark to light w/o banding. In theory according to what is being said here this should not be so. Only in pure HD-SDI 4:2:2.

Finally, I have read the HDV standard, read lit on Canon and I see no specifications stated anywhere as to what is truely going on.

Thanks to all for patience as I try to get a clear understanding in a theory or factual statements that back up my obvious results.

thx
troy

http://dvinfo.net/canonxlh1/articles/article06.php

just found this article by Chris

Troy, uprezing performed by the camera can make the picture look better. One very concrete reason is that your NLE will recognize the color as 10-bit, even though only 8 bits are significant. But during color corrections, those additional two bits are used when storing the new color values. So while you start out with eight bit color, the various color effects you add are not limited to 8 bits but are rather allowed to expand to ten bits.

HDV uprezing by the camera will not look as good as uncompressed, but it is entirely possible that it looks noticeably better than HDV.

Those who say you can't get more than what you put in are not being entirely accurate. Uprezing in one form or another is beneficially used in video all the time. Examples include pixel shifting and in camera sharpening. They do actually improve what you put in. In fact, many forms of digital signal processing that the camera performs are variants of uprezing.

You discovered a good improvement over just using HDV directly into the NLE ;). And people are noticing similar results with the HV-20 capturing HDV out of the HDMI port instead of the Firewire one. So I have no reason to doubt what you're saying. It makes sense.

Of course, uncompressed will look considerably better.

Cool!

That works for me. Thanks alot, I appreciate it.

Best Regards,
Troy