Color Management: What input space to set non-normal spaces?

What should I do when doing RCM if I have a clip that is shot in a flat profile, but not one of the logs or standard gammas?

Panasonic's Cinelike-D(2), Sony's Cine2, or Fuji's Eterna for example. What input space should I set those to so that they behave as best as possible when working with mixed footage? They seem to be based on modified R709 but I can't tell.

In a perfect world I'd only be having R709, R2020 PQ or HLG and log footage to work with in a timeline to make things simpler but sometimes cinematographers have a preference.

For Panasonic Cinelike I use sRGB input space and make my own custom curve in the Colour Page as the first node. There is no correct input colourspace for Cinelike. Make a single node and save to the Stills gallery for repeated use.

Same here with Sony Cine 1 to 4.

Same here with Sony Cine 1 to 4.
If you need a better solution, have a look at FimConvert CineMatch, you’ll find most of such profiles there.

For Panasonic Cinelike I use sRGB input space and make my own custom curve in the Colour Page as the first node. There is no correct input colourspace for Cinelike. Make a single node and save to the Stills gallery for repeated use.

Interesting. I am curious why sRGB instead of one of the R.709 options? Wouldn't sRGB be identical to R709 Gamma 2.2?

Same here with Sony Cine 1 to 4.
If you need a better solution, have a look at FimConvert CineMatch, you’ll find most of such profiles there.

If I ever get a project with a lot of those sorts of clips, I will definitely take a look, thanks!

Does the manufacturer offer a lut to convert to rec709?

For DJI D-cinilike and Gopro Flat, I use Rec709 as input color space and adjust with primaries wheels. The luts I have found for these, I haven't liked.

Same here, DJI’s Cinelike is easy to grade with the tools in DR.

Does the manufacturer offer a lut to convert to rec709?

For DJI D-cinilike and Gopro Flat, I use Rec709 as input color space and adjust with primaries wheels. The luts I have found for these, I haven't liked.

Not for the Non-Log profiles. HLG and Log profiles have easy conversions but the cine profiles usually do not. I have found the same so far but just wanted to make sure what I have been doing has been alright or if there was a better way. Thanks

For Panasonic Cinelike I use sRGB input space and make my own custom curve in the Colour Page as the first node. There is no correct input colourspace for Cinelike. Make a single node and save to the Stills gallery for repeated use.

Interesting. I am curious why sRGB instead of one of the R.709 options? Wouldn't sRGB be identical to R709 Gamma 2.2?

My Panasonic camera uses sRGB colour space, and no, it's not the same as rec709.

For Panasonic Cinelike I use sRGB input space and make my own custom curve in the Colour Page as the first node. There is no correct input colourspace for Cinelike. Make a single node and save to the Stills gallery for repeated use.

Interesting. I am curious why sRGB instead of one of the R.709 options? Wouldn't sRGB be identical to R709 Gamma 2.2?

My Panasonic camera uses sRGB colour space, and no, it's not the same as rec709.

It doesn't for video, though, does it? The sRGB/Adobe RGB color switch in the camera I thought only effects JPGs.

The primaries and white point for Rec.709 and sRGB are identical but the transfer function is not. sRGB/AdobeRGB are indeed only for stills. For non defined 'quasi-log' video you should tag your input color spaces as Rec.709 if that's what you recorded to and grade from there.

Panasonic documentation specifically states sRGB. My tests confirm it. Using rec709 decoding in Resolve will crush the very low black detail as the rec709 colour space gamma transfer function has a curve at the low end that sRGB does not.

Switching it to sRGB as the input color space in a color managed timeline seems to just flatten the shadows and doesn't seem to result in any lost black clipping detail being gained, just to lift it. I'm getting conflicting messaging here.

The mov container reports rec709 as the encoding format but my tests indicate sRGB transfer curve. It's such a small difference I don't think it matters which you use to decode. I have used both. On some super high contrast sunny scenes shooting black car wheels with shiny detail, I got better results using sRGB and on more low contrast scenes or average contrast scenes I use rec709, which looks better. I always use whatever gives a more pleasing result.

The mov container reports rec709 as the encoding format but my tests indicate sRGB transfer curve. It's such a small difference I don't think it matters which you use to decode. I have used both. On some super high contrast sunny scenes shooting black car wheels with shiny detail, I got better results using sRGB and on more low contrast scenes or average contrast scenes I use rec709, which looks better. I always use whatever gives a more pleasing result.

That’s fair. I will do the same. Thanks for your help

Instead of starting a new / separate thread on this, I hope to ask a few more (related) questions.

So when using a "pseudo" LOG curve like Cinelike D or Sony Cine 1-4 or Fuji Eterna, in Resolve 18 Studio, Is it best to set the Color science to DaVinci YRGB Color Managed?

Then should Automatic color Management be checked or unchecked?

Then should color processing mode be set to SDR or HDR? (Or something else if Automatic Color management is left unchecked?)

I am guessing that the original footage has around 10-stops of dynamic range as opposed to the 12+ stops of LOG. If I just use Davinci YRGB (and not DaVinci YRGB Color Managed) highlights are clipped pretty early.

Thanks in advance.

Instead of starting a new / separate thread on this, I hope to ask a few more (related) questions.

So when using a "pseudo" LOG curve like Cinelike D or Sony Cine 1-4 or Fuji Eterna, in Resolve 18 Studio, Is it best to set the Color science to DaVinci YRGB Color Managed?

Then should Automatic color Management be checked or unchecked?

Then should color processing mode be set to SDR or HDR? (Or something else if Automatic Color management is left unchecked?)

I am guessing that the original footage has around 10-stops of dynamic range as opposed to the 12+ stops of LOG. If I just use Davinci YRGB (and not DaVinci YRGB Color Managed) highlights are clipped pretty early.

Thanks in advance.

I only know about Cinelike D, which is just a REC709 colourspace with modified gamma and look to make it look a little flatter for grading. The choice of managed or non-managed resolve setup is entirely up to you as there is no Resolve transfer function that understands Cinelike D. It is just REC709 and up to you how you want to grade it.

Which applies to most of these manufacturer specific 'looks'' or 'profiles'.

If your camera is recording only in 8 bit, choose the one that comes closest to your intended look, define it as Rec709 and the rest is color grading. If you have 10 bit, log can be an alternative for situations with higher contrast.

Instead of starting a new / separate thread on this, I hope to ask a few more (related) questions.

So when using a "pseudo" LOG curve like Cinelike D or Sony Cine 1-4 or Fuji Eterna, in Resolve 18 Studio, Is it best to set the Color science to DaVinci YRGB Color Managed?

Then should Automatic color Management be checked or unchecked?

Then should color processing mode be set to SDR or HDR? (Or something else if Automatic Color management is left unchecked?)

I am guessing that the original footage has around 10-stops of dynamic range as opposed to the 12+ stops of LOG. If I just use Davinci YRGB (and not DaVinci YRGB Color Managed) highlights are clipped pretty early.

Thanks in advance.

I only know about Cinelike D, which is just a REC709 colourspace with modified gamma and look to make it look a little flatter for grading. The choice of managed or non-managed resolve setup is entirely up to you as there is no Resolve transfer function that understands Cinelike D. It is just REC709 and up to you how you want to grade it.

Thank you, Peter, for the reply.

I understand that, in the end, I will have to work it out for myself, but when YOU are using Cinelike D, to you work in Davinci Color Managed and Davinci Wide Gamut? Although i didn't use Cinelike D (I just used a flat profile with contrast and saturation turned down), any tips on how YOU grade Cinelike D might be a good starting point for me.

Thanks in advance for your understanding (and patience with my questions).

Which applies to most of these manufacturer specific 'looks'' or 'profiles'.

If your camera is recording only in 8 bit, choose the one that comes closest to your intended look, define it as Rec709 and the rest is color grading. If you have 10 bit, log can be an alternative for situations with higher contrast.

Thank you for the reply, Uli.

It was a two camera shoot, one shooting in 8-bit, and one shooting in 10-bit, and both of them were in "flat" or "neutral" profiles with saturation and contrast lowered in camera. (Neither of them were in LOG, unfortunately.)

I understnad I will have to experiment to see what works best, but would it be a good idea to try DaVinci YRGB Color Managed and change color processing mode to HDR? I know that the footage it is Rec.709 color gamut but I believe I can regain some of the highlights working in HDR instead of SDR.

My only problem is I ALREADY graded in DaVinci YRGB (not color manadged) , so by switching to DaVinci YRGB Color Managed and HDR, all the colors are all messed up.

Thanks and I am sorry if my questions are confused. I normally work in Panasonic V-LOG and everything is real straightforward.

With rec709 source material I find no benefit is using anything other than rec709 timeline colourspace. HDR makes no sense unless delivering HDR final output. You can't regain any highlights as that is determined by the camera colourspace, not the working colourspace. People tend to complicate the workflow when it is entirely unnecessary when using rec709 sources and also delivering rec709 output. There are tools available in the Colour Page to avoid highlight clipping. Keep in mind the Colour wheels were always designed to work in rec709 SDR colourspace and can be difficult to work with when using Wide Gamut or HDR timeline colourspace, which is why the HDR wheels were added a few versions ago.

Resolve Colour Management is designed to provide an easy way to deal with multiple camera colourspaces from different sources such as RAW mixing with log and rec709. If it's all rec709 there really is not much benefit.

In your case they are all rec709 no matter what camera settings were applied (or bit depth) and so there is no automatic correction from Resolve. You must do all corrections manually no matter which timeline colourspace or Resolve colour management settings you use.

With rec709 source material I find no benefit is using anything other than rec709 timeline colourspace. HDR makes no sense unless delivering HDR final output. You can't regain any highlights as that is determined by the camera colourspace, not the working colourspace. People tend to complicate the workflow when it is entirely unnecessary when using rec709 sources and also delivering rec709 output. There are tools available in the Colour Page to avoid highlight clipping. Keep in mind the Colour wheels were always designed to work in rec709 SDR colourspace and can be difficult to work with when using Wide Gamut or HDR timeline colourspace, which is why the HDR wheels were added a few versions ago.

Resolve Colour Management is designed to provide an easy way to deal with multiple camera colourspaces from different sources such as RAW mixing with log and rec709. If it's all rec709 there really is not much benefit.

In your case they are all rec709 no matter what camera settings were applied (or bit depth) and so there is no automatic correction from Resolve. You must do all corrections manually no matter which timeline colourspace or Resolve colour management settings you use.

Thank you for the reply and explanation, Patrick.

I will have to go through the footage again this morning as I am still trying to figure out what is causing the clipping though and how to fix it.

With an 8 bit Rec.709 capture device, trying to simulate a "log-ish look" with camera settings will make things worse, not better, in post.

Test some things of course. But set contrast no lower than gets the least notable clipping, and sat no lower than gets sat within the Vectorscope outer limits.

Red is typically the channel that has troubles crashing the sat limits in my experience. Red roses for instance.

You only have 256 levels to start with. If you choose to record only say 195 levels, then try to expand that back to 256 in post, well ... you got problems, right here in River City ...

Sent from my SM-S908U using Tapatalk

With rec709 source material I find no benefit is using anything other than rec709 timeline colourspace. HDR makes no sense unless delivering HDR final output. You can't regain any highlights as that is determined by the camera colourspace, not the working colourspace. People tend to complicate the workflow when it is entirely unnecessary when using rec709 sources and also delivering rec709 output. There are tools available in the Colour Page to avoid highlight clipping. Keep in mind the Colour wheels were always designed to work in rec709 SDR colourspace and can be difficult to work with when using Wide Gamut or HDR timeline colourspace, which is why the HDR wheels were added a few versions ago.

Resolve Colour Management is designed to provide an easy way to deal with multiple camera colourspaces from different sources such as RAW mixing with log and rec709. If it's all rec709 there really is not much benefit.

In your case they are all rec709 no matter what camera settings were applied (or bit depth) and so there is no automatic correction from Resolve. You must do all corrections manually no matter which timeline colourspace or Resolve colour management settings you use.

Honestly if all my stuff was Rec709 already, I’d just not use color management at all and grade directly in a Rec709 Gamma 2.4 timeline. Takes away so much complication of input color spaces and stuff.

With an 8 bit Rec.709 capture device, trying to simulate a "log-ish look" with camera settings will make things worse, not better, in post.

Test some things of course. But set contrast no lower than gets the least notable clipping, and sat no lower than gets sat within the Vectorscope outer limits.

Red is typically the channel that has troubles crashing the sat limits in my experience. Red roses for instance.

You only have 256 levels to start with. If you choose to record only say 195 levels, then try to expand that back to 256 in post, well ... you got problems, right here in River City ...

Sent from my SM-S908U using Tapatalk

Thank you for the reply.

I will test them more when I get a chance and see what works.

With rec709 source material I find no benefit is using anything other than rec709 timeline colourspace. HDR makes no sense unless delivering HDR final output. You can't regain any highlights as that is determined by the camera colourspace, not the working colourspace. People tend to complicate the workflow when it is entirely unnecessary when using rec709 sources and also delivering rec709 output. There are tools available in the Colour Page to avoid highlight clipping. Keep in mind the Colour wheels were always designed to work in rec709 SDR colourspace and can be difficult to work with when using Wide Gamut or HDR timeline colourspace, which is why the HDR wheels were added a few versions ago.

Resolve Colour Management is designed to provide an easy way to deal with multiple camera colourspaces from different sources such as RAW mixing with log and rec709. If it's all rec709 there really is not much benefit.

In your case they are all rec709 no matter what camera settings were applied (or bit depth) and so there is no automatic correction from Resolve. You must do all corrections manually no matter which timeline colourspace or Resolve colour management settings you use.

Honestly if all my stuff was Rec709 already, I’d just not use color management at all and grade directly in a Rec709 Gamma 2.4 timeline. Takes away so much complication of input color spaces and stuff.

Yes, that is a good point.

Thanks.

Cine 1-4 like Hyper gamma 1-8 and for that matter s-Cinetone are not "pseudo log" curves . They are ordinary 709 source material with gentler contrast curves and especially with much better roll off for the upper exposures than you can get with traditional knees. According to Alister Chapman, Sony designed the cine and Hyper gamma curves to shot at 1 stop under and then raise up in post but frankly I doubt anyone uses them that way . I just shoot them as a 709 WYSIWYG curve and grade the same way . Typically I haven't used any CSTs, LUTs or color management with them , but I understand there are reasons to want to work in DWG and there are other threads on the forum about how to set up for that.

I just shoot them as a 709 WYSIWYG curve and grade the same way.

I think that makes a lot of sense.

Cine 1-4 like Hyper gamma 1-8 and for that matter s-Cinetone are not "pseudo log" curves . They are ordinary 709 source material with gentler contrast curves and especially with much better roll off for the upper exposures than you can get with traditional knees. According to Alister Chapman, Sony designed the cine and Hyper gamma curves to shot at 1 stop under and then raise up in post but frankly I doubt anyone uses them that way . I just shoot them as a 709 WYSIWYG curve and grade the same way . Typically I haven't used any CSTs, LUTs or color management with them , but I understand there are reasons to want to work in DWG and there are other threads on the forum about how to set up for that.

Thanks for the reply, and for sharing your tips about working with that type of footage. I will search out any threads on working in DWG with thsoe gamma curves.

So I am a bit confused by something here.

My understanding (probably wrong) is that rec.709 would be recorded at video data levels, while the Cine Gamas would be recorded at FULL data levels. Hence, when ingesting footage shot in Cine 1-4, or s.cinetone, or Panasonic CinelikeD, one has to set the data levels to Full in the clip attributes.

So I was assuming that there was more dynamic range available in those gammas than in the standard picture profiles.

Is that wrong?

Also, is there really no benefit (in terms of dyanmic range captured by the sensor / codec) when shooting a rec.709 profile (like Natural or Portrait or Flat or whatever) and turning down the contrast and saturation, only to add them back in during post?

I guess reducing saturation in a Rec 709 profile MIGHT help alleviate color clipping in certain specific situations only?

… while the Cine Gamas would be recorded at FULL data levels. Hence, when ingesting footage shot in Cine 1-4, or s.cinetone, or Panasonic CinelikeD, one has to set the data levels to Full in the clip attributes.

Not exactly. Some, but not all of these record 'Superwhites', but respect the lower level of video (no Superblacks). So, these are neither video nor full range, but 'Sony range'. You can pull down those highlights in post and gain about two thirds of a stop.

But shooting with very low contrast and pulling up in post is counterproductive in 8 bit, apart from cases where you want such a scene, like a foggy day.

In nearly all cases, the difference between full and legal is how the same data is encoded, not how many levels are encoded.

That full always has more data levels, is a very common misperception.

As noted above, there are a few cameras that may record a few more levels in superwhites. Most record exactly the same data.

Sent from my SM-S908U using Tapatalk

Virtually all broadcast cameras are capable of superwhites, 109IRE=1019 in 10 bit, 100IRE is 940.

It’s less than half a stop in straight 709 but with a knee can be a full stop or more.

Good Luck

… while the Cine Gamas would be recorded at FULL data levels. Hence, when ingesting footage shot in Cine 1-4, or s.cinetone, or Panasonic CinelikeD, one has to set the data levels to Full in the clip attributes.

Not exactly. Some, but not all of these record 'Superwhites', but respect the lower level of video (no Superblacks). So, these are neither video nor full range, but 'Sony range'. You can pull down those highlights in post and gain about two thirds of a stop.

But shooting with very low contrast and pulling up in post is counterproductive in 8 bit, apart from cases where you want such a scene, like a foggy day.

Thank you for the explanation.

So just to confirm, when shooting in those Cine Gammas that allow for superwhites, we should set Data Level in the clip attributes menu to full, right?

Then is there a "best" way to pull those superwhites back down in to range? (For example, is HDR panel better than Primaries or LOG wheels for pulling the superwhites back?)

Virtually all broadcast cameras are capable of superwhites, 109IRE=1019 in 10 bit, 100IRE is 940.

It’s less than half a stop in straight 709 but with a knee can be a full stop or more.

Good Luck

Thank you for the explanation regarding 109 vs 100IRE.

When you say:

...but with a knee can be a full stop or more...

You mean that the knee is applied by the gamma curve in camera, not in post, right?

And then we still have to pull the superwhites down in post, right?

Thanks in advance, and sorry if these are basic questions.

BTW: So many videographers talk about how DIFFICULT it is to use LOG because of all the grading required, but to me it seems like it is a lot EASIER to use LOG because you either slap on a LUT or use a CST and you are in to Rec 709.

In nearly all cases, the difference between full and legal is how the same data is encoded, not how many levels are encoded.

That full always has more data levels, is a very common misperception.

As noted above, there are a few cameras that may record a few more levels in superwhites. Most record exactly the same data.

Sent from my SM-S908U using Tapatalk

Thank you, I will have to experiment I guess and see what I can get a way with.

But I am starting to think that the EASIEST thing to do is just shoot in LOG and use either Resolves transform LUTs, or the camera manufacturers LUT, or a CST to get in to 709 space as simply as possible.

But I am starting to think that the EASIEST thing to do is just shoot in LOG and use either Resolves transform LUTs, or the camera manufacturers LUT, or a CST to get in to 709 space as simply as possible.

When done right it is definitely the best way to shoot and do post. But not many operators, editors or colorists have a full understanding of scene referred workflows or they kinda do but still work display referred for the most part. I recently had to grade social assets from a campaign of which the TVC was done by another house. I was lucky enough that they also used Resolve and where happy to share the TVC timeline. I noticed that on every clip the first node was the ARRI LUT and all grading done on top. It isn't wrong as long as creative intent is met, but I happen to know that they've had a ton of feedback on skies being too washed out and bright. Sure the weather wasn't optimal on some shots but a lot of it is also due to the DRT compressing and desaturating highlights which you can easily combat grading in log where you have easy access to the full captured range.

Transitioning from wysiwyg logic to scene-referred has been going on from the 90s and no end in sight. It started with in-house "gamma gurus" explaining to people when and why nonlinearity is useful. Now custom transfer functuons, different gamuts, tone-mappings and whatnot are commonplace but most users still struggle with the same basic principle that data encoding does not have to equal output encoding does not have to equal display encoding. A painting of a scene is not the scene itself, it is a final derived artistic approximation. Modify the scene, then paint it in the end.

When the "hack" came out to shoot log on the original Canon DSLR with video capabilities, that made a huge splash. "Green lantern" or whatever it was. But it was an actual log, and it did get another half stop or so DR. If exposed perfectly.

And suddenly other DSLRs could do video, and most users tried to totally mimic the *look* of the Canon log out of camera.

Which was done by cranking contrast and sat to the lowest the camera could do, and then lowering exposure "to make sure to get all that extra highlight levels recorded".

"That's what log is, right? I mean, it looks sorta log-ish ... must be better ..."

No. Not hardly.

Because the files were normally under exposed so there was terrible noise. There weren't enough levels recorded to fill out even the Rec.709 image dynamic range without banding and other artifacts. And trying to get full saturation back resulted in distorted color.

And then some users decided "the real fix!" was to use Full on their monitors. Which of course just screwed things up worse for viewing the file on any other system.

So then, some camera makers offered the option to record to full ... not because it could record more stops, as they always recorded all the data the camera produced. But as a sales gimmick. It *sounded* like they got wider dynamic range.

And though that also screwed things up in post, it sold cameras! Yippee!

Rec.709 should normally be recorded in legal/video range, unless it's 12 bit RGB. Which few cameras can shoot. As then the entire system just works without fuss or bother.

Monitors left to "auto levels" for Rec.709 will nearly always *display* both file types as full. That's as designed.

And Resolve, with Rec.709 levels set to Auto, will also correctly display both standard video levels and RGB 12 bit as full.

And as been noted above, the switch to working "scene referred" has (for many) further confused things. The proper way to setup scene referred projects isn't quite what you might think it is.

So the more experienced heads here help the rest of us sort this out

Sent from my SM-S908U using Tapatalk

Virtually all broadcast cameras are capable of superwhites, 109IRE=1019 in 10 bit, 100IRE is 940.

It’s less than half a stop in straight 709 but with a knee can be a full stop or more.

Good Luck

Thank you for the explanation regarding 109 vs 100IRE.

When you say:

...but with a knee can be a full stop or more...

You mean that the knee is applied by the gamma curve in camera, not in post, right?

The knee is applied prior to gamma encoding in the DSP. The FX9 for example, outputs S-Cinetone with superwhites.

It’s also worth mentioning that Resolve’s IRE and mv scales are finally accurate. Prior to 18.5 selecting video level would compress the signal to 64-940 in 10 bit but would do the same for IRE and mv resulting in incorrect scaling , ~10-90 IRE and 70-930mv. It’s now much easier to destermine whether or not the source material contains information from 700-800mv.

Good Luck

"Knee " is a misnomer when applied to CineGammas, Hypergammas and S-Cinetone all of which are Sony specific by the way . CineGammas are just the consumer or industrial version of Hypergammas and are very similar . S-Cinetone is just another flavor and I don't know exactly how it compares.

Knees are the time honored way to compress highlights for Standard 709 curves. The problem with them is that depending on the shot you can get wildly saturated colors in the highlights. Not always but it can bite you. I first noticed this with an ordinary orange and foliage is notorious for an intense yellow green.
The other gammas were Sony's fix. Basically 709 , but with gentler preset curves that don't oversaturate in the highlights. They work quite well for WYSIWYG video and each curve is a little different with Cine 4 and hypergamma 7 being the most aggressive. Some of them only record to 100IRE (Cine 2) and others include superwhites which you can bring down in post if you want to depending on your delivery . S-Cinetone has largely become the WYSIWYG gamma of choice because it has very nice skin tones , but some people aren't crazy about it for general exterior shooting. I shoot with it often as many clients wisely don't want s-log because they don't know what to do with it and have had excellent results with all the others for years.

You do want to be careful though that flesh tones stay out of the compressed range especially with Hypergamma 7 and Cine 4, because that looks awful. This is why Alister Chapman said they were designed for slight underexposure though no more than a stop. I would just say be careful.

They are particularly good for 8 bit video and as mentioned somewhere else on this thread -DO NOT underexpose or reduce saturation on 8 bit video and don't shoot Log on it either.

So just to confirm, when shooting in those Cine Gammas that allow for superwhites, we should set Data Level in the clip attributes menu to full, right?

It is my understanding that no, you shouldn't do that.

When the "hack" came out to shoot log on the original Canon DSLR with video capabilities, that made a huge splash. "Green lantern" or whatever it was. But it was an actual log, and it did get another half stop or so DR. If exposed perfectly.

Magic lantern was the name, nifty piece at the time. It could do pure raw too, but at first only a few seconds until camera’s internal memory buffer was filled, then it had to write to card which was too slow to record to directly. Most awesome piece of magic lantern project was the way someone managed to dump the firmware of cheap “soap box” cameras by simply blinking it all out in binary with the camera led light and reading it with a light detector. Now that is magic.

I still remember the days of hacking my private GH2, while I had access to Red cameras if needed.
With fast cards and some vintage glass that little thing was a beast - and very unobtrusive.