When will the Flash XDR ship?
 

Just wondering when these units will be on the market?

It seems we have been reading and posting about the unit like it exists for quite some time.

Is it close to becoming a reality?

Also, will the Nano ship when the main unit ships?

Thanks

Hi Tim-

We are scheduling a small production run for the last week of May. Assuming all goes well, we should deliver a few units in early June followed by full production in late June.

This schedule is still very tentative and based on component delivery and debug of the firmware.

The nanoFlash is still in design and a few more months away. We are re-using 95% of the code from XDR, so we expect nanoFlash will follow quickly.

Thanks for the update

any shipments ...
 

Have any of the units been shipped?

Hi Dean-

To the best of our knowledge, the hardware design is finished and we're working on the final code. All the components are due next week, so we should be able to do the production run the week after next.

So, we are still shooting for initial deliveries this month.

Deliveries
 

Given that you are so close to delivery of product... Two weeks in theory... Why no sample videos yet?

Has anyone determined if this will work with FCP?

Thanks

David Schmerin

Dear David,

There is a logical answer to your question.

Mike Schell wanted the sample videos to be created with the final revision of the circuit boards.

Convergent-Design has a Canon XL H1 and a XH G1 in house and has arranged for a Sony EX1 for the creation of the sample videos.

We have been discussing the type of footage to be included in the sample footage.

Your suggestions would be most welcome.

High color scenes and maybe even some blue/green screen shots. This could be one of the most important uses for the XDR. This will really help to show off how important 4:2:2 color is.

Other shots need to include fast complex motion with lots of detail. Nothing breaks a mpg2 encoder faster then high motion with high detail. Try to shoot some nature shots such as fast running water or blowing trees. Sudden changes in lighting also break a mpg codec pretty fast such as rapid stage lighting. It would really help to show how the XDR will hold up in these situations when compared with HDV.

Dear Thomas,

Thanks for the suggestions.

For the fast motion, such as running water or blowing trees, what shutter speed would you like to see?

We should have a formal announcement on the file formats shortly, so please stay tuned.

Not trying to discount the need for file support in an NLE, just trying to keep in mind that Tapeless workflow is a fairly new innovation and that as an industry we are still working out all the kinks.

With that said, I am quite certain you will be pleased with the upcoming announcement, because we also agree that NLE support from the CF cards w/o transcode or file re-wrap is an important feature.

I think most of us are expecting the concept to work, we just want to see video footage
where the concept is at its maximumwith less than optimum footage taken in less than ideal situation....

some non-compressed footage will never look different than compressed footage as long as the the compression is good and the footage fits that compression...

the objective is to get as good as non compressed footage as you can... and later they
expect to upgrade the XDR to full non compression. It don't get much better than that...

As to NLE, I'm still playing with Premier Pro CS3, some issues I'm seeing, but I'm trying to eliminate the GOP with Cineform in my testing, but After Effects seems to handle the footage without any issue that I can see at this time.. you just need to import the footage and tell AE to interpret the footage for the pull down and the progressive footage looks very good.

Showing my age...
 

What is in question here is will we be able to do the same with files created by your system. I am glad you agree that file support without the need to re-wrap or transcoding is an important feature. Is it important enough to see that it this IS a feature?

David Schmerin

Hi Ray-
Agree with you 100%. As time permits, we will continue to post more footage at various bit-rates, so everyone can evaluate various scenes and lighting conditions.

If you or anyone has some uncompressed footage that your like us to process, then please send me an e-mail (sales@convergent-design.com) and we'll do our best to get it processed. Ideally, we need uncompressed Quicktime files, which we can load into our FCP system and then output HD-SDI into a Flash XDR box. I can try to set up an FTP site for uploads, or you could send a DVD directly to our office.

Just started a new thread on the uncompressed option...

Hi David-

Just wanted to point out that Flash XDR and nanoFlash also have HD-SDI output, which, by the way, is the only way you can ingest HDCAM and D5 material (there is, to my knowledge, no software HDCAM or D5 CODEC for file-based transfers).

Maybe you did not realize this capability (HD-SDI out) was also part of Flash XDR / nanoFlash?

Yes, file based transfer is absolutely the best way to "ingest" your video. But, if your editing system does not support our file formats, then you can always use the HD-SDI route, just like HDCAM.

As I mentioned, we should have a major announcement on more file format support shortly. (We already announced MXF file support).

Mad Magazine
 

As a child I was a great fan of Mad Magazine. I recall they used to (and may still) run a regular series of What They Say Vs What They Mean.

When you say, "Maybe you did not realize this capability (HD-SDI out) was also part of Flash XDR / nanoFlash?"

What I hear is... Be prepared to have a 2nd unit on hand strictly for in house use unless you want to send your field unit to back to the office every time you need to send in footage because you can not count on being able to read the data files into your NLE.

The other thing I hear is if you want to send your raw material directly to another house, they should be prepared to have both a Flash XDR unit AND an HD-SDI capture card. Again because there is no assurance the data files will be able to be read. Perhaps you were not aware but when shooting in HDCAM, at the end of the day, I can simply leave the tapes with the client and walk away. Leaving behind Flash XDR units with clients can get to be very expensive.

When you say, "But, if your editing system does not support our file formats..."
What I hear is be prepared for it not to because we don't know. If we knew we would say XXX works and YYY will not work.

I suppose it is the not knowing that bothers me the most because I really want for this product to be amazing and I want for it to work really well with no issues being caused simply because of a choice in editing platforms. And I want these units in house here ASAP.

Based on everything I have read, this product if it has not actually started shipping yet is so close to shipping you can taste it. C-D obviously have Final Cut Pro in house. C-D should have as most recent a model of the Flash XDR unit as could possibly be obtained...

At this time, do the data files captured by the Flash XDR unit work in Final Cut Pro with out the need for capturing via HD-SDI? __YES __NO (Please check one)

My guess is because you are using the HD-SDI out on the Flash XDR to feed into your Final Cut Pro, the answer to the question is No.

David Schmerin

Hi David-
I was also a fan of Mad magazine and always enjoyed "Spy vs Spy" and some of the X-Ray views, especially the one of the gas station with all the various grades of gasoline being fed from the same tank.

Your questions and concerns are all very valid and reasonable. Please give me a few more days and I should be at liberty to answer them freely. Once you see the response, you will understand the delay in the answer.

Flash XDR / nanoFlash Update
 

Here's a copy of the most recent newsletter, in case you did not receive the e-mail blast:

Flash XDR Progress / Update
First, we wish to apologize for the delays in shipping Flash XDR. We had hoped to ship units in May, but it now looks like early July before our initial production units will be complete. The hardware design, including the printed circuit boards and the cabinet is complete. All the components will be at the contract manufacturing site by tomorrow, awaiting to be scheduled onto the production line.

Software development is also progressing well. We have very solid record and capture to the 32GB Transcend Compact Flash cards (now approx US $135) using the 100 Mbps 4:2:2 1920x1080 CODEC. We can create directories and files on the Compact Flash card and are debugging the audio/video streaming to the CF card with the full file system in place.

We were pleasantly surprised at the quality of the Long-GOP video at the 100Mbps rate, even in high motion situations; it’s virtually identical to uncompressed in our tests! We now believe that the 100 Mbps rate will be the ideal choice for 99% of the applications. Yes, Long-GOP has suffered from bad press, primarily due to the overly-compressed 25Mbps 4:2:0 HDV format (which also resizes the video to 1440x1080) coupled with long render times. However, the 50/100Mbps 4:2:2 full-raster (1920x1080) MPEG2 implementation from Sony (nanoFlash and Flash XDR utilize the Sony MPEG2 CODEC) produces video superior in quality to HDV, DVCProHD and even HDCAM (see CODEC quality chart, available on our website).

Additionally for FCP users, you do not need to render back to MPEG2, but rather only decode the video when editing (and use an I-Frame CODEC, such as ProRes to render the effects), thus eliminating virtually all the render issues associated with Long-GOP editing.

Video Comparison Tests
We are creating a library of video comparison tests on our website. So far, we have posted the following on our website:
1) Full motion blue screen shot: Video Comparison Uncompressed + 100 Mbps; shot with a Sony EX1 (128MB each)
2) Single frame Uncompressed vs 100 Mbps comparison with Photoshop “Difference” (8 MB)…good luck finding the difference between these two images!
3) Single Frame Comparison: Uncompressed vs DVCProHD, HDV and 50 Mbps MPEG2 (0.9 MB); even 50 Mbps looks virtually identical to uncompressed!

We will continue to test and post additional video comparison clips and still images as time permits.

nanoFlash Announced / Early Adopter Program
nanoFlash, the smaller (<20% by volume) and lower-cost version (US $3495) companion to Flash XDR was announced last week. nanoFlash has all the capabilities of Flash XDR, expect embedded audio only (no analog audio), only two Compact Flash slots (instead of four) and only MPEG2 recording (no upgrade to uncompressed). You can also add ASI I/O to nanoFlash (pricing TBD). Brochure and Press Release are available on our website.

nanoFlash is targeted for Sept 08 delivery and is an ideal upgrade for HD-SDI cameras with embedded audio. If you are interested in becoming an early adopter, please send an e-mail to sales@convergent-design.com with the subject “nanoFlash Early Adopter”. You will get early access to the product and an incentive for your feedback.

Uncompressed Support for Flash XDR
The uncompressed support has been announced for Flash XDR and will be available this Fall as a US $995 firmware upgrade. Four high-speed Transcend 300X 16GB CF cards (US $145 to $170 each) will enable 10/8 minutes of uncompressed 4:2:2 1080p24 8/10-bit footage! This Fall, high speed 32GB cards will be available, which will increase these uncompressed record times to 20 minutes (8-bit) and 16 minutes (10-bit).

The video will be striped across 4 CF cards configured in a RAID0 configuration. Initially the only playback scenario is HD-SDI out of the Flash XDR box. We are investigating the possibility of a program to recombine the files, but capturing via HD-SDI will always remain the fastest route into your NLE. The datarate to the Compact Flash cards will be limited to about 130 Mbyte/sec, which will cover 1080p24, 1080p25, 1080i50 and 720p50/30/25/24 at 8/10-bit. 1080i60, 1080p30 and 720p60 will be limited to 8-bit. (Note the theoretical bandwidth to the four CF cards is about 180 Mbytes/sec, so the 130 Mbytes/sec limit is quite conservative).

ASI I/O Option Update
The ASI I/O code (MPEG2 mapped onto an SD-SDI transport) is in debug using BMS microwave equipment. We hope to have this functionality debugged and field tested in July.

Presentation and Archival Applications
Two additional applications for Flash XDR / nanoFlash, which you should consider, are high-quality presentations and project archival. Both XDR and nano offer seamless playback across multiple files and CF cards. Given their portability, these recorder/players can be used in trade show videos, client review, or presentations. The 100 Mbps 4:2:2 high-quality video looks spectacular even on large Plasma/LCD screens or high-end projectors. nanoFlash will also include a 5V power tap for use with HD-SDI to HDMI converters.

High-quality archival of your finished project is also made simple with these innovative recorders. Simply connect the HD-SDI output from your NLE into XDR / nano and record your finished video at the 100 Mbps 4:2:2 high-quality (visually lossless) setting. You can then transfer the file(s) to a Blu-ray disk for long-term storage (a dual-layer 50GB Blu-ray disk (US $35) holds one-hour of 100 Mbps video).

Future Enhancement to Flash XDR / nanoFlash
Here’s a list of some of the future product enhancements we have on the roadmap (no dates yet):
1) Standard Definition Video (SD-SDI) support
2) Additional file format support (on the Compact Flash cards)
3) Remote control protocol via RS485 twisted pair (status, start/stop, metadata update, configuration).

XDR Audio
 

Mike, Can you give us some information about the audio portion of the XDR???

We know what the XDR can do for video....

what about the Audio besides XLR inputs....

specs ?? or any info please

Dear Ray,

I may be able to help.

I am certain that you know the basics:

Two Channels of XLR inputs, which can be mic or line levels.

Two Channels of XLR outputs (via a single 5-pin XLR connector for space reasons). A common pigtail can be used to split this into two separate XLR outputs).

If the external audio inputs are in use, then the signals can be embedded into the HD-SDI output in real-time. The same applies to timecode. This is especially nice for the original Canon XL H1.

The external audio circuitry was intended to provide very high quality audio.

The audio parts were selected, then the circuits were designed, then reviewed by the audio experts of a major chip manufacturer that specializes in audio circuits.

These experts then recommended we change to their "top shelf" parts.

We changed the design to use their recommended parts and then they reviewed the circuits again.

In the design, only the best available parts were used.

But, while we do have the spec's for the individual parts, this does not convey the necessary information.

The important spec's are the end to end (mic or line level input to output) specifications. This takes into consideration the noise created by the power supplies and the interference, if any, from the video circuits. The circuits were designed with special care to minimize interference.

We have used a Tektronics audio signal generator and an oscilloscope to check on the audio quality on an end to end basis. We were very pleased with the results.

The microphone level gain is variable, and can be 0 dB gain or 10 to 65 dB of gain.

The line level input is sophisticated in that the line level audio does not get padded down to mic level and then amplified as if it was a mic level signal originally.

This is important to ensure quality of sound.

We intend to have the audio capabilities checked by a respected audio expert as soon as possible.

Dear Ray,

To answer your question in another way:

The audio circuits of most cameras are not in the same league as top notch separate audio recorders. Most cameras, but not all, are limited to 16 bit audio recording, usually 16 bit / 48K. A few cameras are capable of recording 24 bit audio.

We have attempted to provide high quality audio circuits. This includes the ability to record at 16 bit / 48K or 24 bit / 48K.

If you wish to record a scratch track on the camera, first feed the signal in the Flash XDR, then route the outputs of the XLR into the camera's inputs. This way, the audio will not be compromised by the camera's audio circuits.

Will limiters be implemented in the audio ??

Dear Ray,

Audio limiters did not make it into the current design of the Flash XDR.

We will consider this for inclusion in the next revision of the Flash XDR.

While we would like to have audio limiters, since we fully support 24 bit
audio, we recommend that you record using 24 bit audio and set the levels to
leave enough headroom.

With 24 bit audio, it is easy to normalize the audio in post without
problems. The extra resolution of 24 bit audio makes this possible and very
effective.


Please note that the lack of audio limiters only affects the external audio
inputs.

The embedded audio from the camera is already digital. As such, all limiting
must be done in the camera or in a mixer prior to reaching the camera. So
the lack of audio limiters has no effect whatsoever on audio first feed into
the camera and then sent to the Flash XDR via the HD-SDI.


Also, if one is using a mixer with good audio limiters, any audio limiters
in the Flash XDR would be redundant, if proper techniques are used. Thus, if
one sets the gain on the Flash XDR to match the mixer output levels, the
mixer will catch and limit any signal overloads and it will be impossible
for the Flash XDR to overload.

For example, one could easily set a mixer to output tone at 0 dbfs, then set
the gain on the Flash XDR to around -20 or -18 dBfs (or your personal
preference), then if the limiters are engaged on the mixer, then Flash XDR
should not overload.

If one is not using a mixer, just set the gain on the Flash XDR to leave
headroom, use 24 bits, and you should be ok.

Flash XDR Update
 

Here's a quick update on the Flash XDR Developments:
1) Production boards are due in our facility next Wednesday (July 16).
2) Cabinets are due on Tuesday (July 15).
3) Code development is going well, we are debugging the most complex part of the design, writing data to the CF card while crossing a file boundary. We're close, hope to make the final jump tomorrow.
4) Owner's manual is in review now and expected this weekend.

Stay tuned we are really close.

Mike would it be possible to add a sata controller and some sort of sata conector to the Flash XDR? My interest would be to use a 2.5in drive for recording long interviews or conferences and not to have to deal with changing flash cards. For most things I am looking forward to purchasing Nano but also having the ability to record loooong things would be an asset. The connector may need to be a lemo or fischer connector and the same on the other end for the drive case, it also needs power.

Would this be possible or is it too much of a PITA. Being able to record from the EX-1/3 in 50-100Mb would be awesome for long form stuff.

Cheers
Robert C. Fisher
Cinematographer

Hi Robert-
Good idea! But, we simply don't have room on the nanoFlash to add another port, it's really crammed! We are looking into several possible solutions for longer recording options. Stay tuned, we should have some announcements soon.

Dear Robert,

With the Flash XDR, recording 1080i60 at 50 Mb, using four 32 GB CompactFlash cards, you have approximately 284 minutes of recording time.

That is 4 hours and 42 minutes without changing a CompactFlash card using today's widely available, low cost Transcend 32 GB cards. We expect higher capacity CompactFlash cards this year. 64 GB CompactFlash cards are predicted.

Until we do have higher capacity cards, there are a few interesting options:

1. Record at a higher compression rate (not the perfect solution).

2. Mount the Flash XDR off the camera (this allows a card to be swapped without disturbing the camera). Then "Hot Swap" the CompactFlash cards that are full, sometime before all four cards are full.

When 64 GB cards become available, then approximately 568 minutes of recording time, without a card swap, will be available at 50 Mb. 568 minutes = 9 hours and 48 minutes.

I feel that there is usually a break, even during long conferences.

Convergent-Design recognizes the need for even longer recording times.

As Mike says: "Stay Tuned".

Flash XDR Update - Production Boards Arrived
 

We just received our first batch of Flash XDR boards today. They are in debug now with the hopes of shipping units early next week.

More info to follow soon.

Does 100Mb half the record time or is my math off? I'm thinking at a 100Mbs:

4x 32GB cards = approx. 2 hours and 20 minutes
4x 64GB cards = approx. 4 hours and 45 minutes

Dear Michael,

Yes, compared to 50 Mb, 100 Mb, will give you approximately half the recording time.

For discussion:

When using a "Constant Bit Rate" MPEG2 compression scheme, it is easy to predicte that the recording time will be cut in half if you double the constant bit rate.

However, with variable maximum bit rate MPEG2 compression scheme you may get more than half the recording time. (I have not yet tested this.)

It would appear to me, that depending on the complexity of the scene, using a 100 Mb variable maximum bit rate compression scheme, then you will not always need the full 100 Mb of data. Thus, your record times may be longer than expected, in some cases.

As Mike reported, the first batch of production boards have arrived, as have the production cases, the rubber overmold, and all other parts.

Previous, pre-production, revisions of the boards were fully tested.

New features were added to these production boards to improve power consumption.

For example, the power supplies for the external audio circuitry, and the external audio circuitry itself, will be powered down when not needed.

With every change to the circuitry, extensive testing must be accomplished, once the boards are built.

As of last night, 10 pm Eastern Time, the testing was proceeding nicely.

The power supplies were tested and the boards were powered using their internal power supplies.

HD-SDI Video was supplied to the boards, and the unit successfully recorded the video, creating files on the CompactFlash card.

Long Form video was tested, in other words, we tested recording past the 4 GB maximum file size for one file. The unit automatically closed one file and opened another, as planned and previously tested.

Then testing of Long Form video was performed with the files crossing over from one CompactFlash card to the next. This was successful.

Playback of these recorded files was also successful.

As of 10 pm ET last night, other testing remains to be performed:

Testing of the External Audio Inputs (Fully tested in previous revisions of the boards)
End-to-End Testing of the Audio Circuits (Fully tested in previous revisions of the boards)
Testing of the Headphone circuitry (I do not have information on previous testing of this circuitry)

And other testing; (I do not have an exhaustive list of the remaining tests to be performed).

All testing, to the best of my knowledge, so far has gone well.

The development of the Flash XDR was done in parallel, with the hardware and firmware being developed at the same time.

Of course, some firmware development could not be performed, or tested, until the appropriate hardware was built.

Some firmware (software) remains to be finished.

The team is working very hard to deliver the first units as soon as possible.

Thanks for your update.

Best of luck with the process.

Is there a way to know when the XDR or the Nano will be available to somebody who orders today?

Meaning how long it will take to fill the pre-orders and start filling new orders?

Thanks

Dear Tim,

That's a tough question to answer.

Mike Schell knows the number of preorders.

What is not finalized, is how many can be manufactured per month.

The initial production run has to be fully tested to ensure that the circuitry works flawlessly before a high volume order can be placed. As I reported this morning, last night's testing went well. I do not have the results of today's testing at this time.

If we detect a problem in one of the circuit boards it may be that all of that production would have to be scrapped, at significant cost.

If everything goes well, there is a possibility that production can be ramped up.

Hi Tim-
We have the initial XDR productions boards in house now undergoing testing and debug. If all goes well, our first XDR untis will go out the door this week.

Orders placed on the XDR today, would ship at the end of September, nanoFlash orders would ship mid-October.

Thanks for the update.

Nothing like talking.
 

I'm sure you guys are - and it's really great to be kept in the loop with how things are going. I remain excited at the prospect of trying something new and innovative like this and have taken to buying myself pelican-like, windowed, water and shock-proof cases for my recently (substantially) expanded CF collection to keep my self happy - and to stave off the anticipation while you get ready. Looking forward to getting mine and staying with the development of the line.

Justin.

Dear Justin,

As you build your collection of CompactFlash card, please be aware that the Flash XDR and nanoFlash require top-notch performance from the cards in order to pull off the "magic".

Currently, Convergent-Design has qualified two cards:

Transcend 32GB 133x and
Transcend 16GB 300x

Transcend cards seem to meet their specifications, which is a very good thing.

There are a lot of cards that are rated at 133x, but for many the crucial write speeds are lacking, and are no way near the 133x speed.

For up to 100Mb Long-GOP recording, the Transcend 32GB 133x card works well, at least so far in our testing and development.

For higher speeds, and for the future "uncompressed" recording, the Transcend 16GB 300x card is required.

For bit stream rates of under 100Mb, lower speed cards may be successfully used, but, at this time, have not been qualified by Convergent-Design.

There are two main types of Flash Memory, "Single Level Cell" (SLC) and "Multiple Level Cell" (MLC).

The SLC cards have always been used in "Performance" Flash Memory as they are faster, while the MLC cards have been used in "High Capacity" Flash Memory, but they have much slower write speeds.

The write speeds are what is critical.

Other promising cards, as they become available, may be qualified in the future.

We have word that a 32GB 300x card, from a trusted manufacturer, will be available by the end of the year.

We are expecting a very fast (300x) 64GB CompactFlash card, from another manufacturer, to be available this year.

We do not have pricing on these new cards.

You said it. Actually Mike did.
 

When speaking to Mike, he was very particular about this point. I currently have twelve Transcend 16GB 300x UDMA cards and, for my D3 and D300 (and as a back up), six of their 8GB 300x UDMA cards. I haven't found any of the Transcend 32GB variety yet but will look for those, or the 64GB versions which are sure to follow, sometime in the fall or as they emerge onto the internet market.

My EX3 shipped yesterday so I'm nearly set.

Jus.

Dear Justin,

I was worried that you were collecting a diverse group of CompactFlash cards for use with the Flash XDR or nanoFlash.

I am very happy that you choose to purchase the cards that work!

The technology that allows us to record to a 32 GB 133x card at 50 Mb or 100 Mb is pretty amazing. A lot has to happen just right, including some very sophisticated programming to allow this to work.

But, it does, with the right cards.

How is the testing of the first premiere units going?

Dear John,

Testing is going well with the first run of the production boards for the Flash XDR.

There was one part, an inductor, on the first few boards that was rotated 90 degrees.
This caused an problem with the phantom power supply. So the problem had to be diagnosed and corrected.

The circuit boards are built using an automatic "pick and place" machine. So, this has been corrected and should not happen again.

The above problem slowed down the testing of the audio circuits. Yesterday, with the problem solved, all of the power supplies were tested and are working well.

Initial testing of the audio circuits (on these production boards) is scheduled for today.

The recording and playback is working well. Recording and playback across the 4 GB boundary, as well as across multiple CompactFlash cards, is working.

In parallel to the testing, others are developing the firmware for the Flash XDR.