VHS capture hardware (again)

[ccooper]

I see that there are several recent threads regarding capture hardware, but none seem to exactly answer my questions.

As with most people, I would like to capture some VHS tapes into a digital format. One thing that seems different about my situation though than some I have read is that I see this as a one-time project that I don't ever want to do again. With that in mind, I would like to make sure the digitized video contains as much information and detail as can practically be extracted from the original tape. (I realize there are numerous elements prior to the capture stage that affect this as well.) Also, while I don't necessarily see myself doing any editing in the short term beyond cutting extra lead-in or lead-out that was on the tape, I wouldn't want to preclude myself or someone else from doing that in the future. Lastly, since I only have a few tapes and storage space is continuing to get cheaper I don't see any benefit to storing the video in a lossy format.

From everything I have read, my impression is that my best bet would be to capture raw uncompressed video and compress it using the Huffyuv or Lagarith codec. This would imply that I need a capture device that can output a raw uncompressed video stream. With regard to the codecs, I figure these are a safe bet since they are both open source and likely to be around for a long time. Plus, since the video will be in a lossless format I expect it will be feasible to recompess it with a different lossless codec if a better one comes along or support for the old ones becomes sketchy, assuming the use of compatible color spaces.

So from all of the guides and threads I have read, here is what I am currently thinking:

- One option is to use an external Canopus capture device, but as far as I can tell these only support capturing video in the lossy DV format. So despite the praises for these devices, this doesn't seem so great.

- Another option is to use a MiniDV camcorder in pass-thru mode. While it seems that some such camcorders have really great TBC capabilities, this seems worse than the Canopus option since it also encodes in the lossy DV format and doubtfully has an S-Video input.

- The third option is to use an internal capture card that can be installed in a desktop computer. From a theoretical standpoint, I have seen suggestions that one downside of such devices is that they are subject to a huge amount of RF noise due to being installed inside the computer case. I'm not sure what the impact of this is in practice though. From a practical standpoint, this isn't my favorite option since I would preferably like to use my laptop to do the capturing as it is my most modern computer. However, if this is the best option I have access to a few different desktops that should suffice.

- The last option is to use an external non-Canopus capture device. All of the specific devices of which I am aware appear to use a USB interface. Apparently there are some limitations to using a USB based device, but it sounds like they can be managed by dedicating the USB bus to the capturing process. Using a FireWire based device would be an option, but all of the ones of which I am aware output video in the lossy DV format again.

With the above stated, have I misunderstood or misinterpreted anything I have read? Are there any other options I should be considering? Assuming I am on the right track, what are the best capture devices for raw uncompressed video?

Chris

[robjv1]

Kind of on a side note -- but whichever method you end up deciding upon, make sure your playback device is up to par. If you want to "make sure the digitized video contains as much information and detail as can practically be extracted from the original tape" then you should start with a good VCR. Most VCRs will play back what is on the tape plus a bunch of noise. Some VCRs will strip all of the clarity out of your picture, and then a few will capture it well. I know just what you mean about not wanting to end up having to do this project again -- I have had to redo mine several times with several different decks in the quest for a more detailed final product.

[edDV]

So from all of the guides and threads I have read, here is what I am currently thinking:

- One option is to use an external Canopus capture device, but as far as I can tell these only support capturing video in the lossy DV format. So despite the praises for these devices, this doesn't seem so great.

DV format has more than adequate sampling bandwidth to capture VHS (3MHz luma, 0.5MHz per chroma component). The 5x DCT in frame compression is not heavy. The problem with Canopus devices is not DV format but lack of optimization for unstable sources like VHS/S-VHS players. They don't have adequate TBC or no TBC. They also capture too much out of band noise (above 3MHz luminance). This is also true for all other consumer capture devices and computer tuners.

- Another option is to use a MiniDV camcorder in pass-thru mode. While it seems that some such camcorders have really great TBC capabilities, this seems worse than the Canopus option since it also encodes in the lossy DV format and doubtfully has an S-Video input.

This varies by brand and generation of camcorder chips.

- The third option is to use an internal capture card that can be installed in a desktop computer. From a theoretical standpoint, I have seen suggestions that one downside of such devices is that they are subject to a huge amount of RF noise due to being installed inside the computer case. I'm not sure what the impact of this is in practice though. From a practical standpoint, this isn't my favorite option since I would preferably like to use my laptop to do the capturing as it is my most modern computer. However, if this is the best option I have access to a few different desktops that should suffice.

Consumer capture/tuner cards have the same issues of no TBC and no low pass filtering VHS luma and are worse than DV in that they super sample high frequency noise out to 54Ms/s vs. 13.5Ms/s.


- The last option is to use an external non-Canopus capture device. All of the specific devices of which I am aware appear to use a USB interface. Apparently there are some limitations to using a USB based device, but it sounds like they can be managed by dedicating the USB bus to the capturing process. Using a FireWire based device would be an option, but all of the ones of which I am aware output video in the lossy DV format again.[/quote]

Most all at the consumer level encode to Mpeg2 or MPeg4. This introduces intraframe (motion based) compressing in addition to in frame DCT compression. MPeg encoders usually compress more intraframe than DV. Also these devices lack TBC or low pass luma filtering.

With the above stated, have I misunderstood or misinterpreted anything I have read? Are there any other options I should be considering? Assuming I am on the right track, what are the best capture devices for raw uncompressed video?

So none of the above are optimal for VHS. You can add upstream TBC, proc amp and bandpass filtering and then use DV or an uncompressed tuner. The tuner will still add some internal noise.

If you just have a few tapes why not get them transferred on a pro level frame sync and pro level capture device? Have them cap directly to uncompressed SDI (270Mb/s). You can then compress to huffyuv on your own time.

[ccooper]

edDV, any recommendations on where I can get this done by someone at the "pro level" as you suggest? I have considered this and talked about it with a few local people offering capture services in the Austin, TX area. Unfortunately I haven't gotten the feeling that these people are necessarily experts or have equipment beyond what I could acquire for a few hundred dollars.

[edDV]

Here are some places to call. Ask a TV station or post house who they use.
http://www.mojopages.com/video-duplication/austin/tx

[lordsmurf]

Unfortunately I haven't gotten the feeling that these people are necessarily experts or have equipment beyond what I could acquire for a few hundred dollars.

That's sadly how many of the strip-mall types locations are. They buy the same stuff you have easy access to, at the local Best Buy or whatever. It's consumer crap. If you're lucky, you'll find a few pro pieces. A lot of us operate out of non-retail areas, private locations, be it a converted warehouse or what-have-you. People "dropping in" is a major pain in the ass, we're doing difficult video tasks, not making burgers.

When requested, I can convert VHS filter-less, and to an uncompressed format. The customer must either supply hard drives, or provide funds for them -- these files ARE NOT small.

Be warned ---- cold-calling from a phone list is going to get you all kinds of answers, and you never know what you'll get in return. Early this year, I contacted about 50 different services, with a request for one of two basic tasks, out of curiosity. I was shocked out how many of them gave ridiculous answers, operations simply run by clueless idiots that should be doing something else. Be very wary.

If you're able to actually get a one of the production/technical folks from a local television station to assist you, then that might be a good place to seek help. But honestly, in my experience, a lot of the TV people don't know much about VHS, that's never been their working format.

[ccooper]

Thanks for the advice! When making inquiries about getting such work done, is it reasonable to expect the other party to describe their setup (e.g. hardware, cabling) and workflow? When asking people about their setup and workflow, I haven't gotten very straight answers. On one hand, I can somewhat respect this may be the secret sauce to their business. On the other hand, I want to know whether the end result is going to meet my expectations and objectives. Or, maybe the problem is that I just haven't been talking to serious professionals.

[ccooper]

Re: the SDI suggestion, I have been doing some reading and am confused about two points. As far as I can tell SDI is basically a wire protocol (i.e. electrical signaling and data format definition) for digital video. As far as I know there aren't any VHS players that output directly to SDI format. However, I do see that there are some high-end (?) capture/media devices (e.g. from Blackmagic Design) that have SDI and analog support. Hence, is the suggestion that by using one of these cards it would likely result in a near optimal capture from an analog source and internally the card may even be converting the data to an SDI data stream? Also, wouldn't the capture be written to disk in a format that is independent of SDI itself (e.g. YUY2 or its 10-bit equivalent)? In other words, I'm unclear on the significance of SDI for this application other than it suggesting the use of professional level equipment.

[edDV]

If you're able to actually get a one of the production/technical folks from a local television station to assist you, then that might be a good place to seek help. But honestly, in my experience, a lot of the TV people don't know much about VHS, that's never been their working format.

Better to ask them about what they do with U-Matic and Betacam SP. That will separate out the better dub houses. Post houses may do it themselves. You may get a reasonable deal if you aren't in a hurry.

[minidv2dvd]

ccooper - to bring the discussion to earth level reality, vhs has at best 240 lines of resolution, or a digital size of 320x480 not a lot to work with or worry about. the hard part is a decent vhs deck. imo the late model jvc s-vhs with 19 micron heads or the panasonic model 1980 are your best bets, if you can find one. capturing to mpeg-2 more than covers the source output, and if dvd is the target destination, it saves an extra re-encoding step that would cause further degradation.

the longer you wait the more magnetic information is lost from the tape.

[edDV]

He has indicated he wants archive quality of important tapes and DV is too compressed in his opinion. True VHS resolution is ~240x480 "lines of resolution*" which takes ~640x480 to reproduce with accuracy but there is much more to video quality than resolution for archival needs.

Time base correction and noise control are key if you want to compress later. If you don't intend to motion compress just use uncorrected DV. It will be 13GB/hr and play ok. Huffyuv will be more like 40-60GB per hour. Done.

MPeg encoders don't like non-timebase corrected or noisy video. Their motion detectors aren't up to the challenge. Actually, there is no way they can restore noise or timebase errors.


* 240 lines of horizontal resolution means 240 alternating white and black lines can be resolved without blending to gray observed with the human eye on a high resolution monitor.

[edDV]

Re: the SDI suggestion, I have been doing some reading and am confused about two points. As far as I can tell SDI is basically a wire protocol (i.e. electrical signaling and data format definition) for digital video. As far as I know there aren't any VHS players that output directly to SDI format. However, I do see that there are some high-end (?) capture/media devices (e.g. from Blackmagic Design) that have SDI and analog support. Hence, is the suggestion that by using one of these cards it would likely result in a near optimal capture from an analog source and internally the card may even be converting the data to an SDI data stream? Also, wouldn't the capture be written to disk in a format that is independent of SDI itself (e.g. YUY2 or its 10-bit equivalent)? In other words, I'm unclear on the significance of SDI for this application other than it suggesting the use of professional level equipment.

The significance is like you say, pro A/D devices* output SDI or analog component. SDI is 8 or 10 bit commercial YCbCr uncompressed video that can be dumped to Quicktime or AVI wrappers in a file format. Best to give them a hard drive or they will charge near list for one. Then you can convert that to what you want.


* Their pro SD converters usually low pass luma to 4.2MHz or a higher bandwidth cutoff. Ideally you would ask for 3MHz for VHS to minimize higher frequency noise capture.

[SatStorm]

IMO, capturing and video processing is not possible at the same time. The correct capture, allows you to store and process your files later, in the best possible way. The big problem is, how to capture correct your VHS source.

In short terms, a full frame TBC is needed, a nice VCR deck and the correct capture device.

The "correct" capture device can be a PC capture card, a DV-based solution (canopus, etc) or a DVD Standalone Recorder.

- With the capture card, you can do many things and cheap. But it is complicated. And, the video process needs time. Huge amounts of time.

The DV Solutions makes things less complicated but some may say that those solutions don't capture the 100% of what your tape has to offer for processing. That can be true (especially for NTSC), but if you gonna filter your source later, this "loss" doesn't really matter. It's only matters if you don't filter your source, which IMHO, it is stupid. Filtering is a must. It enhances and prepares your analogue source for a digital conversion.

The third solution, a DVD standalone Recorder, creates "huge" files, because only on the highest settings we can produce almost macroblock-free results. The good Standalones can also automatically balance the White Colour errors, do some basic image processing to stabilize the picture and in some cases can even eliminate basic colour bleed or do some colour smoothing. And the overall process is really easy and fast: just hit "rec" on the remote. And later, you can rip your discs to your PC and do some more filtering.
A source from a good DVD Standalone Recorder is not a bad source, it is simply inferior compared the other methods. With the post capture filtering, that difference minimize a lot. But the truth is, you never gonna have a source as good as what a capture card or a DV device can offer you. It is easy to do your job with a standalone, but it is not the 100% of what you can do.

I follow the DV root personally, but I gather the hardware over the years. Those things are expensive you know.

[courtney16161]

Trying to find the "right capture device" is nerve racking. But I saw the Blackmagic products came up, when I checked those out it seemed you could only get one conversion say analog to SDI per box and would have top shell out another $470 for another box to go SDI to HDMI (hypothetically) but I also found a Matrox mxo2mini that seems to be what will get the job done borderline simplistically and flawlessly (again hypothetically) but for u I would also recommend the Canopus ADVC-55. Works for the laptop and has a 99% "pass rate". It has upper medium level of "fancy-smancyness" Using a capture card seems like too much hassle not enough result, but I hope maybe one of these products helps

[ccooper]

courtney16161, I think you may have been looking at the wrong Blackmagic product. I noticed as well that they have some conversion boxes like you describe but as far as I can see they wouldn't help get the captured video into the computer to be written to disk, unless you have a separate SDI card or similar. However, it looks like they have several media cards that can be installed into a desktop computer ranging from consumer to professional levels and in price from approximately $200 to $1000. For the purposes I originally described above, it looks like I would either need the $200 or $700 card with the difference being 8-bit color vs. 10-bit color. These prices don't seem insane, but in my case I would have to buy a new desktop computer to support the PCI Express interface, which is not currently in my budget.

Re: the color precision (i.e. 8-bit vs. 10-bit) from what I have read on these forums it sounds like 8-bit isn't really enough if you plan to do any clean-up/correction in software later without losing quality. It has been suggested that it would make more sense to use a proc-amp device to make adjustments before performing the capture. In my case, I don't feel like I know enough to get a proc-amp configured optimally so I would probably be more comfortable capturing with 10-bit color precision where hopefully it would be more reasonable to experiment later and make software based adjustments. One thing that confuses me a bit though is that it looks like many TBCs/proc-amps only use 8-bit color internally so I'm not certain why using software to make adjustments on an 8-bit capture would produce different results. Maybe 10-bit or better TBCs/proc-amps are are common than I think and hence this isn't a problem.

Personally, I was hoping to experiment with a Panasonic AG-7750 VCR as my playback device. I haven't received it yet, but I was able to buy one on eBay for less than $100 with shipping. Despite a few limitations discussed in other threads, it apparently has good pciture quality and a built-in TBC. According to the schematics in the service manual though, the TBC is only 8-bit.

[edDV]

The No. 1 issuse with VHS is time base correction if you intend to process or compress later. Video process algorithms assume pixels line up vertically. When they don't, horizontal pixel jitter is interpreted as motion.

Bit depth just means the number of gray levels captured.

8 bit 256 levels -- black at 16, nominal white at 235
10bit 1024 levels -- black at 64, nominal white at 960

Normal capture process is to levels correct, timebase correct and noise reduce before 8 bit A/D. Pro analog recorders do all this inside the VTR before analog output. Ideally, one would avoid analog output and export 8 to 10 bit digital from the TBC/DNR. Some pro level TBC/frame syncs can output digital.

Here are some pro level TBC/Frame Syncs designed for VHS input. These may be found in a quality dub facility. Most output 10 bit SDI but a dub house can deliver an uncompressed YCbCr file on a hard disk.

http://www.ensembledesigns.com/products/brighteye/be03
http://www.for-a.com/products/fa147/fa147.html
http://www.for-a.com/products/ufm147dfs/ufm147dfs.html
http://www.for-a.com/products/fa145/fa145.html
http://www.for-a.com/products/fa370p/fa370p.html
http://www.bhphotovideo.com/c/product/378720-REG/Hotronic_AP41_SPD_YUV_AP_41SPDYUV_Tim...Corrector.html
http://www.bhphotovideo.com/c/product/307726-REG/Prime_Image_50III_50III_Time_Base_Corrector.html

Prices range from $1200 to $6000

[lordsmurf]

Don't get hung up on bit depth.

[2Bdecided]

Don't get hung up on a lot of the things in this thread!

It's VHS. You need a decent deck, correct levels, decent TBC, decent NR, simple edit, decent encoding, some authoring, and DVD burning. These might be in one or several boxes; hardware or hardware+software.

The rest is icing.

10-bits for VHS is angels on a pin head!

Cheers,
David.

[lordsmurf]

This reminds me of the people who insist on 48-bit scans, saved as TIFF over JPEG --- and then print their images on a consumer inkjet at 8x10.

Unless you're planning to submit this VHS to the Smithsonian or LOC, or want it to be integrated into a film shown on a 100" theatre screen, then this is supreme overkill to go SDI, 10-bit archived, etc.

It's just a VHS tape -- a low-end consumer format that can only be made marginally better beyond the typical "restoration" work you see done here by myself and some other members (TBC, level, contrast, chroma+grain suppresion, color de-casting, audio fixes, etc). The equipment any of us has already does internal 12- to 10-bit processing to 8-bit for output. You can really store larger than 8-bit or even work with it unless you want to blow a huge wad of cash on hardware and storage.

Keep some perspective, don't get OCD. It's easy to go OCD, trust me here. It's why some of my projects STILL on my hobby desk have been "in progress" since 2003. (At least my OCD is purely about color corrections and edits -- every clip is like 2 minutes long, and I have about 2000 minutes worth, each needing different attention!!).

[edDV]

I agree that 10 vs. 8 bits is unlikely to yield much advantage for VHS which has less than -45dB signal to noise but I'd bet direct digital out of the TBC/Frame Sync will yield better results than a D/A followed by A/D with a consumer capture device. 10 bit depth SDI is typical at this next grade of TBC so there is no financial advantage to specifying less.

Ccooper has been asking for an uncompromised approach to VHS digital capture supposedly for unique source preservation. The two main areas I see for process improvement would be 3MHz low pass filtering of analog luminance* to reduce out of band noise capture and direct digital out of the TBC. The resulting time base corrected uncompressed YCbCr capture file would be a better candidate for software digital filtering including lossless compression for archive. The raw capture file would be ~ 130GB per hour.


* SVHS would use a 4.5MHz low pass filter.

[2Bdecided]

The two main areas I see for process improvement would be 3MHz low pass filtering of analog luminance* to reduce out of band noise capture and direct digital out of the TBC.

Surely by the time you get to this exotic level, the capture device isn't going to keep anything beyond it's own bandwidth (6.5MHz for D1), and removing 3MHz-6.5MHz is a trivial software filter.

Or hardware filter, as you've said.

Why is this a big deal?

Cheers,
David.

[lordsmurf]

uncompressed YCbCr capture file

This brings up a point of question...

Would YUY2 or YV12 be okay, too?
YUY2 should be a YCbCr codec, but asking anyway, at risk of sounding stupid (a verification).

UYVY doesn't seem to be as common, though I've not looked at it recently.

I've generally used YUY2 from ATI All In Wonder AGP cards, with its 12-bit A/D converter. The higher-end cards just cost more than was able to be recouped later, customers were fine (even my own OCD needs were met on quality). I have access to some, however, if needed.

[edDV]

The two main areas I see for process improvement would be 3MHz low pass filtering of analog luminance* to reduce out of band noise capture and direct digital out of the TBC.

Surely by the time you get to this exotic level, the capture device isn't going to keep anything beyond it's own bandwidth (6.5MHz for D1), and removing 3MHz-6.5MHz is a trivial software filter.

Or hardware filter, as you've said.

Why is this a big deal?

Cheers,
David.

A hardware filter before A/D may be more expensive but it gets the job done. Post A/D digital filtering may work for a full component path. Trouble is most consumer level TBCs are limited to composite I/O which introduces luma high frequency errors into chroma and chroma errors into luma.

The advantage of these pro level TBC/Frame Syncs is a component path and a single A/D. If the digital low pass filtering is done before any image processing, it may be as effective.

[edDV]

uncompressed YCbCr capture file

This brings up a point of question...

Would YUY2 or YV12 be okay, too?
YUY2 should be a YCbCr codec, but asking anyway, at risk of sounding stupid (a verification).

UYVY doesn't seem to be as common, though I've not looked at it recently.

I've generally used YUY2 from ATI All In Wonder AGP cards, with its 12-bit A/D converter. The higher-end cards just cost more than was able to be recouped later, customers were fine (even my own OCD needs were met on quality). I have access to some, however, if needed.

SD SDI is component 4:2:2 YCbCr at 13.5Ms/s (SMPTE-259M). It can be collected into a YUY2 AVI or Quicktime file for import to editors or image processors as uncompressed 4:2:2.

Generally it is better to keep the edit session 4:2:2 and encode to 4:2:0 as a last step.

I agree this may be overkill for VHS source but this is the normal work flow. A digital intermediate conversion would eliminate the need for a RAID.

[vhelp]

lordsmurf, UYVY was primarily seen and used in the ATI line of capture cards--they included the encoder/codec, though tied to their MMC suite. It is the same as the YUY2 codec..

..same sampling numbers (units) as YUY2, using the 4:2:2 sampling format, only placed differently in the container, much like YV12 vs. i420, both are the same sample numbers only different (chroma) placement.

If capturing from a VCR, the best route for maximum picture restoration would be the 4:2:2 format..YUY2 or UYVY container. All the consumer level of capture cards I've seen and used by default are 4:2:2, so it is best to stay within the 4:2:2 (YUY2 or UYVY) sampling, at the capture level, editing level, before finallizing to 4:2:0 (ie, mpeg1/ 2/ divx/ xvid/ h264) containers, though for storage or archival purposes, you could archive to mpeg-2 in 4:2:2 sampling but at the cost of larger filesize.

It's all pretty funny how we all go through the laborous routine of capturing an already compressed source (ie, vhs) and then process it (usually) in lower specs, and then process it further, and then finalizing it to a lower format, ultimately in the 4:2:0 sampling. But then again, we are (or were at one time) tied to limited playback and storage issues.

example of various processing scenarios:

vhs -> capture_card[422] -> codec[422] -> edit[422] -> dest_container[422] <-- MAXIMUM BEST, but not practicle
vhs -> capture_card[422] -> codec[422] -> edit[422] -> dest_container[420] <-- BEST, practicle in todays work
vhs -> capture_card[422] -> codec[420] -> edit[422] -> dest_container[420] <-- good
vhs -> capture_card[422] -> codec[420] -> edit[420] -> dest_container[420] <-- least

-vhelp 5146

[2Bdecided]

Trouble is most consumer level TBCs are limited to composite I/O which introduces luma high frequency errors into chroma and chroma errors into luma.

A Panasonic or JVC S-VHS deck has a composite TBC?

I can believe it, but it seems strange. It's not like the chroma on VHS needs much storage - it's what, about 10% the resolution of the luma?

Cheers,
David.

[edDV]

Trouble is most consumer level TBCs are limited to composite I/O which introduces luma high frequency errors into chroma and chroma errors into luma.

A Panasonic or JVC S-VHS deck has a composite TBC?

I can believe it, but it seems strange. It's not like the chroma on VHS needs much storage - it's what, about 10% the resolution of the luma?

Cheers,
David.

I don't know about the JVC deck. I was referring to external units like the Datavideo TBC-1000.

[edDV]

Here is a Fortel patent for a 12 bit 4:2:2 analog to analog or analog to digital TBC including a list of prior art patents.
http://www.freepatentsonline.com/7046299.html

[lordsmurf]

If you're able to actually get a one of the production/technical folks from a local television station to assist you, then that might be a good place to seek help. But honestly, in my experience, a lot of the TV people don't know much about VHS, that's never been their working format.

Better to ask them about what they do with U-Matic and Betacam SP. That will separate out the better dub houses. Post houses may do it themselves. You may get a reasonable deal if you aren't in a hurry.

I'd have to disagree somewhat. Smaller operations have sold off a lot of equipment, due to significantly slowing demand in that area. However, there are still a quite a few consumers still doing basic consumer format transfers. A lot of smaller professional operations have shifted into streaming. Just last month, I saw a fairly decent-sized operation pretty much liquidating its entire inventory of VTRs, for formats like digital Betacam and U-matic. In fact, some of us have actually expanded more in Betamax, 8mm and other non-VHS home formats.

You go where the market is. Not much demand for U-matic or Betacam right now.

[lordsmurf]

The two main areas I see for process improvement would be 3MHz low pass filtering of analog luminance* to reduce out of band noise capture and direct digital out of the TBC.

Surely by the time you get to this exotic level, the capture device isn't going to keep anything beyond it's own bandwidth (6.5MHz for D1), and removing 3MHz-6.5MHz is a trivial software filter. Or hardware filter, as you've said. Why is this a big deal? Cheers, David.

Many of us are visual. What exactly would noise in 3MHz-6.5MHz or 6.5Mhz+ look like?