p.12 #1Canon 5DS dynamic range: zero improvement (RAWs analyzed)
alundeb wrote:
So with a photosite dynamic range of 15 stops, and a ADC dynamic range of 12.5 stops, why does the Canon architecture still yield only 11 stops DR? The answer is on the sensor chip, in the ISO amplifiers. Each photosite has its own amplifier, with its own output noise.
Part of it is the speed of the off-chip ADC not being able to capture the whole sensel DR. Sony's technical paper explains it far better than I could Here is a pdf http://www.sony.net/Products/SC-HP/cx_news/vol47/pdf/featuring47.pdf
"To resolve these problems, Sony adopted the column-parallel A/D conversion technique. (See figure 2.) Since each column has its own A/D converter in the column- parallel A/D conversion technique, the analog signals read out from the vertical signal lines can be A/D converted directly.
Since the analog signals are A/D converted directly without first recording and storing, the capacitors that take up chip area as circuit structural components are no longer required. Also, the analog CDS (correlated double sampling ) circuits used for noise cancellation are no longer required. The CDS operation previously performed with analog signal processing is now performed with digital processing. This technique makes it possible to perform high precision noise cancellation that is not dependent on variations in the CDS circuits.
Another advantage of the column-parallel A/D converter technique lies in its conversion speed. Since processing is performed in parallel for each column, the A/D conversion frequency is extremely low, and the high-frequency band noise components can be separated from the signal components. This means that even if the number of pixels or the frame rate are increased, high picture quality digital signals with extremely low noise can be acquired."
The Canon ADC runs very fast and is sensitive to picking up "high frequency band" noise as it goes from sensor to ADC - independent of the setting of the ISO gain for each pixel, presumably.
p.12 #2Canon 5DS dynamic range: zero improvement (RAWs analyzed)
Noise bandwidth and "picking up" are two separate issues.
If we set aside interference, and concentrate on noise bandwidth:
Noise as generated by the amplifier itself is dependent on the bandwidth. It is usually specified as a number of nV/sqrt(Hz). Or to put it differently: Many ADC's mean that the output noise from the amp is filtered over longer time before sampled by the ADC, and that lowers the amp noise.
The ouput noise from the amplifier and the ADC noise are in the same point in the signal path, so we cannot easily determine which is the cause. But I got the information earlier in this thread that the ADC's that Canon use are capable of 12.5 ENOB at the actual sampling frequency.
p.12 #3Canon 5DS dynamic range: zero improvement (RAWs analyzed)
RustyBug wrote:
Thanks Jimmy ... will check out the link.
In the mean time ... +4 EV overexposure & -5 EV underexposure seem like they might be kinda foreign concepts to me for my shooting. I tend to push/pull around +/- 2.5 EV at most by intent. At 5 stops underexposed ... I just don't seem oriented to shoot that way very often.
Hi RustyBug,
Up until recently I was a fan of the notion and practice of ETTR, or exposing for my shadows. But, after listening to previous debates over the DR issue, and then following my own curiosities regarding the same, I'm beginning to think that technology has brought us something new to consider in our exposure choices.
As a Canon shooter since the advent of DSLR's with my 300D>30D>5D2>7D, I always exposed with a mind to protect my shadows or risk losing vital image information. Now I'm thinking, those days are over. Pushing and pulling 2.5EV with only 11-stops to play with has traditionally run the risk of both blown highlights and noisy shadows. Pushing and pulling those same 2.5EV with 14-stops to play with apparently opens some new possibilities.
Here's and interesting article from a wedding photographer who discovered he no longer needed to ETTR to protect his shadows, but now shoots to protect his highlights. Why? Because the sensor in his D750 doesn't destroy or otherwise render useless-with-noise his under-exposed regions. Says he...
<< As photographers, we strive for correct exposure but the ability of modern Nikon cameras to find details in shadows opens up a debate as to what correct exposure actually is. More and more, I find myself technically exposing wrong with post-processing in mind.
>>
Lot's of real-life exposures in this one showing what all this means...
p.12 #4Canon 5DS dynamic range: zero improvement (RAWs analyzed)
jimmy462 wrote: RustyBug wroteThanks Jimmy ... will check out the link.
In the mean time ... +4 EV overexposure & -5 EV underexposure seem like they might be kinda foreign concepts to me for my shooting. I tend to push/pull around +/- 2.5 EV at most by intent. At 5 stops underexposed ... I just don't seem oriented to shoot that way very often.Hi RustyBug,
Up until recently I was a fan of the notion and practice of ETTR, or exposing for my shadows. But, after listening to previous debates over the DR issue, and then following my own curiosities regarding the same, I'm beginning to think that technology has brought us something new to consider in our exposure choices.
As a Canon shooter since the advent of DSLR's with my 300D>30D>5D2>7D, I always exposed with a mind to protect my shadows or risk losing vital image information. Now I'm thinking, those days are over. Pushing and pulling 2.5EV with only 11-stops to play with has traditionally run the risk of both blown highlights and noisy shadows. Pushing and pulling those same 2.5EV with 14-stops to play with apparently opens some new possibilities.
Here's and interesting article from a wedding photographer who discovered he no longer needed to ETTR to protect his shadows, but now shoots to protect his highlights. Why? Because the sensor in his D750 doesn't destroy or otherwise render useless-with-noise his under-exposed regions. Says he...
<< As photographers, we strive for correct exposure but the ability of modern Nikon cameras to find details in shadows opens up a debate as to what correct exposure actually is. More and more, I find myself technically exposing wrong with post-processing in mind.
>>
Lot's of real-life exposures in this one showing what all this means...
Those sensors give shooters the freedom to not worry about ETTR if they choose but like all sensors they still benefit from ETTR because more exposure means less noise across the entire tonal range. That means more PP flexibility for not just shadows but deep contrast adjustments as well. My use of ETTR on the 5Dc is more about that flexibility then about shadows. Here's a recent 5Dc example:
p.12 #5Canon 5DS dynamic range: zero improvement (RAWs analyzed)
Now that we are back on track with DR and Canon camera issues (and off the my camera is better aren't I smart I bought it track), ...
@cgarcia: I see via another thread that LibRaw supports the 5Ds, R, so you might give your analysis a run on the files generated from that converter. It eliminates one set of variables that have to be worked around / compensated for. It would increase confidence in your results ... results that I and others thank you for working to produce even though they are only preliminary.
However, I don't see it on their website, but Ilyiah Borg claims it is supported. http://libraw.org
p.12 #6Canon 5DS dynamic range: zero improvement (RAWs analyzed)
Jimmy,
Thanks for the enthusiasm and info ... but, to my eye those wedding pics that were pulled up look pretty HDR-ish to me. I guess that's fine if you like that look, but not my gig.
Maybe a more deft hand @ PP would have done better, but my general thought on it is that there is no such thing as a Free Lunch. Pulling up from deep underexposure, while possible, still comes with a cost. I get the excitement, but I think it is also provides for some lesser workmanship. In the name of expedience and time = money ... I get that some folks will find that justifiable ... just not my gig, atm. I think a two shot bracket (occuring in a fraction of a second) is a better way to cover extreme DR situations ... that, or add light.
Some folks will prefer to work from the single file, pulling up hard, but I think you can produce better finished work from the two files when your DR is extreme. Even with a 5 EV pull possible, I'd still want to grab two shots and put them together, rather than one file pulled up that much. I get that it is nice to have it in your hip pocket ... but i don't think it yields optimal over well executed fill techniques or good brackets for the different lighting.
There's another aspect that most folks don't consider when dealing with extreme DR variance. It typically comes with a different color of lighting WB in the form of mixed lighting. Having the two shots makes for better color correction of the different color lighting source. Likely, I'm one out of 1,000,000 shooters that feels this is worthy to address ... but that's me. Well executed fill flash can serve to correct for color issues also.
The other aspect besides the color of the light is the contrast level of the light. The shade lighting is vastly lower contrast than the direct or background lighting, As such, it warrants a different set of processing to really bring it to its best. Having the second file to work differently makes it much easier to address the fact that your scene really has two different sets of lighting conditions ... but that's just me. Personally, I think the reliance on pulling up the shadows in lieu of either a bracketed exposure (or adding some light to the scene) is a recipe for a lesser product.
So, while I get the greater DR has usefulness ... it still doesn't resolve some of the other inherent problems that come with such extreme scenes. Atm, the adage of "just because you can, doesn't mean you should" comes to mind. If pulling up DR is the holy grail answer ... I guess we'll someday be able to dispense using flash or reflectors at all.
Of course ... I'll get beat by a myriad of folks who will argue that without the DR they wouldn't have gotten the shot at all. It's a valid argument, but I just don't know that the diff in DR makes the whole case. I mean ... photographers have been producing great images in high DR scenes going back to film @ either double exposures or with light augmentation. To my eye ... the 5 EV push / pull doesn't necessarily produce a superior product, albeit a somewhat more convenient one.
Would I welcome more DR in the Canon's ... most likely yes. Do I think pulling up 5 EV produces optimal results over other methods ... hmmm, not so sure about that.
p.12 #7Canon 5DS dynamic range: zero improvement (RAWs analyzed)
snapsy wrote:
Those sensors give shooters the freedom to not worry about ETTR if they choose but like all sensors they still benefit from ETTR because more exposure means less noise across the entire tonal range. That means more PP flexibility for not just shadows but deep contrast adjustments as well.
ETTR is also about capturing max tonality (tonal levels). The last stop of light samples half the tonal levels of the camera. The bottom 5 stops only sample 32 levels. The sixth stop samples 32 levels all on its own and so on doubling per stop.
This is a reason why I do not like the idea of pushing super dark exposures into lower mids and midtone. In a perfect sensor, where everything is perfectly clean, the tonality will still suffer compared to a blended bracket where important subject matter is always exposed such that the sampling captures a nice number of tonal level variation.
It is not so much a big deal that linear sensors, binary sampled, have limited tonality at the bottom. The human eye is insensitive to detail at low brightness levels. What sucks is that half the tonal sample range is slammed into the single brightest stop of light where the eye is also detail insensitive. At least recovery can work well here by pulling those levels down into range.
p.12 #8Canon 5DS dynamic range: zero improvement (RAWs analyzed)
NormanPCN wrote: snapsy wroteThose sensors give shooters the freedom to not worry about ETTR if they choose but like all sensors they still benefit from ETTR because more exposure means less noise across the entire tonal range. That means more PP flexibility for not just shadows but deep contrast adjustments as well.ETTR is also about capturing max tonality (tonal levels). The last stop of light samples half the tonal levels of the camera. The bottom 5 stops only sample 32 levels. The sixth stop samples 32 levels all on its own and so on doubling per stop.
This is a reason why I do not like the idea of pushing super dark exposures into lower mids and midtone. In a perfect sensor, where everything is perfectly clean, the tonality will still suffer compared to a blended bracket where important subject matter is always exposed such that the sampling captures a nice number of tonal level variation.
It is not so much a big deal that linear sensors, binary sampled, have limited tonality at the bottom. The human eye is insensitive to detail at low brightness levels. What sucks is that half the tonal sample range is slammed into the single brightest stop of light where the eye is also detail insensitive. At least recovery can work well here by pulling those levels down into range.
The original conception/belief of ETTR was that it improved IQ by encoding image data into higher tonal levels where there are more bits available to encode them. That notion was quickly disabused because it puts the cart before the horse - the reason why the brighter parts of an image are encoded with more bits is because those tones have less noise and thus more bits are useful to represent/distinguish those tones. In other words, it's not more bits that improve IQ but the improved IQ that requires more bits. Using more bits than is necessary (specifically, more bits than are needed to encode signal vs noise) only serves to waste storage space. This is why there are so few bits to represent the shadows - the lower signal levels means few bits are necessary to distinguish signal/tonal levels from noise.
p.12 #9Canon 5DS dynamic range: zero improvement (RAWs analyzed)
RustyBug wrote:
One more thought ... what is the DR of MFD cameras ... and why aren't MFD shooters abandoning MFD in favor of SoNikon's greater DR? If Canon is "locked out" by patent, so is MFD (unless partnering with Sony). Does that mean that MFD will die because of SoNikon's higher DR?
Just a thought.
In our studio we did actually. Our Mamiya DF body was costing a fortune in repairs and when it died for good we ran some tests side by side, our Leaf Aptus II-8 back vs our D800e. The key was using the same lens, the Mamiya glass was far superior to the Nikkor glass. Side by side the sharpness and resolution was practically the same, colour, even after building ICC profiles from our Datacolor SG chart for both for use in Capture One, went to the Leaf but not enough that we were willing to invest the sums needed to keep the medium format system working.
Just to give you an idea those sums were: a) a new DF with a new shutter and mirror box each year plus a spare for using during repair time = $15K every two years. b) An upgrade to a Credo back (same megapixels but allows live view for precise focusing) for use with an Hcam B1 body which has user interchangeable shutters = $13K plus $2000 for shutters every two years. c) An upgrade to Credo and a tech cam system with digital shutter lenses which have a million actuation life = if you have to ask you can't afford. Or option d) a D800e with a Mamiya 80mm Macro lens that I got from ebay for $200 shipped from Japan which gives practically as good resolution, is sharper in the corners than the 80mm Schneider lens we had been using until now, gives a 200K actuation shutter with fast and cheap replacements, has excellent (comparitavely) live view and is a world more versatile in every way.
My dealer called up Phase One told them that they were about to lose a large studio to Nikon. They offered a 10% reduction on the upgrade to the Credo. Sorry medium format world but you have to do a bit better than that. A fashion photographer who bought into Leaf at the same time as us deserted to the D800 even earlier than we did. The new CMOS MFDB's are good no doubt about it however until they have a camera worthy of the system more and more pro's are going to ask whether the dwindling IQ differences are worth the massive difference in price. To be honest Phase One (the big kid on the block) couldn't care less. They have reinvented themselves with the rich amatuer crowd and are doing very well indeed out of them. Personally we are ready and waiting for the Nikon 50 megapixel offering and will buy two immediately it is released. I was interested to learn of the Canon offering but without the comparable DR, not going to waste our time. We need sensors whose output is extremely malleable in post and which can handle anything our clients can throw at us now or in the future with the minimum of time and effort, both of which are worth very real money.
p.12 #11Canon 5DS dynamic range: zero improvement (RAWs analyzed)
snapsy wrote:
The original conception/belief of ETTR was that it improved IQ by encoding image data into higher tonal levels where there are more bits available to encode them. That notion was quickly disabused because it puts the cart before the horse - the reason why the brighter parts of an image are encoded with more bits is because those tones have less noise and thus more bits are useful to represent/distinguish those tones. In other words, it's not more bits that improve IQ but the improved IQ that requires more bits. Using more bits than is necessary (specifically, more bits than are needed to encode signal vs noise) only serves to waste storage space. This is why there are so few bits to represent the shadows - the lower signal levels means few bits are necessary to distinguish signal/tonal levels from noise....Show more →
I don't quite follow your reasoning but would like to understand it. Do you have a link with discussion where it "was quickly disabused"? All the links I find are supportive of EttR, for example http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html
p.12 #12Canon 5DS dynamic range: zero improvement (RAWs analyzed)
RustyBug wrote:
Thanks for the enthusiasm and info ... but, to my eye those wedding pics that were pulled up look pretty HDR-ish to me. I guess that's fine if you like that look, but not my gig.
Would I welcome more DR in the Canon's ... most likely yes. Do I think pulling up 5 EV produces optimal results over other methods ... hmmm, not so sure about that.
This is my view. Too many of these 4-5 stop shadow adjustments look too HDR-ish especially when dealing with people. I would like greater DR as well (at any ISO not just base) but it has to be effective not just something that is possible. In a previous thread someone gave a link to a couples shoot and was trying to show how the exmor sensor could handle backlit situations. The skins tones/colors looked awful and I can't imagine that couple wanting to pay for those images.
p.12 #13Canon 5DS dynamic range: zero improvement (RAWs analyzed)
From Imaging Resource, 5Ds, 5DsR pre-production full-res files, no longer available...
UPDATE 3/25/15: Canon has requested that we remove links pointing to the full-resolution camera images, so we've done so, as of about 1pm US EDT. They've said that screen-res images with 1:1 crops are OK, though, and we'll be adding a few more examples of that sort, hopefully later today. Apologies for the disappointment; we'll shoot and post more samples as soon as we receive a production-level version of the camera.
p.12 #16Canon 5DS dynamic range: zero improvement (RAWs analyzed)
ggreene wrote: RustyBug wroteThanks for the enthusiasm and info ... but, to my eye those wedding pics that were pulled up look pretty HDR-ish to me. I guess that's fine if you like that look, but not my gig.
Would I welcome more DR in the Canon's ... most likely yes. Do I think pulling up 5 EV produces optimal results over other methods ... hmmm, not so sure about that. This is my view. Too many of these 4-5 stop shadow adjustments look too HDR-ish especially when dealing with people. I would like greater DR as well (at any ISO not just base) but it has to be effective not just something that is possible. In a previous thread someone gave a link to a couples shoot and was trying to show how the exmor sensor could handle backlit situations. The skins tones/colors looked awful and I can't imagine that couple wanting to pay for those images.
Totally agree, pull out over 2 stops of shadow will lead to very strange tone and color and these can never be corrected (can be minimize only). So I love to have more DR for more easy shooting but not the priority when choosing gear.
p.12 #17Canon 5DS dynamic range: zero improvement (RAWs analyzed)
snapsy wrote:
use of ETTR on the 5Dc is more about that flexibility
+1 ... just not too far TTR (as I used to do, till I learned better).
KKFung wrote:
So I love to have more DR for more easy shooting but not the priority when choosing gear.
+1 ... I think this ^ pretty much sums it up.
Dinosaur that I am ... the concept of 5 EV underexposure just doesn't jive with me as optimal. It does jive with me as technologically cool. It can be a saving factor in some instances, and it can make for some expediency in certain regards, but those are different things from optimal image making.
This 5 EV push / pull concept kinda reminds me of the diff in shooting positives vs. negatives. When shooting chromes ... your exposure latitude was very tight. When shooting uber-wide latitude negatives, your exposure could vary much more greatly. However, the difference in final quality of a well exposed chrome vs. a very underexposed negative was rather different in the end. Given the choice of +/- 5EV exposure pp vs. other methods @ capture (i.e. bracket, or light), I tend to believe optimal capture yields optimal final product (combined with excellent pp).
Some folks lived for the latitude that the negatives afforded them. Others put forth the effort to contend with the lesser latitude of chrome.
I realize that this isn't "exactly" the same as the more DR discussion because less noise is better in most regards. But, the concept of exposure latitude @ 5 EV of adjustment from underexposure is still going to be a "less than optimal" strategy of capturing light and producing best possible images (which is a diff matter from acceptable). Is it a strategy that has its place ... sure, who wouldn't like to have less noise. But, imo 5 EV underexposure (although pretty cool to be able to "acceptably" recover from) still doesn't provide it to be holy grail of optimal image making. Still nice to have, but not my first choice for optimal. More than one way to skin the cat, and just because you can ....
But, that's just me ... I didn't shoot wide latitude negatives (although, I would use VPS 160 for weddings instead of Fujichrome 50 for the film profile diff) for my optimal work. If I underexposed by 5 EV ... it was likely because I wanted a silhouette.
p.12 #18Canon 5DS dynamic range: zero improvement (RAWs analyzed)
Monito wrote: snapsy wroteThe original conception/belief of ETTR was that it improved IQ by encoding image data into higher tonal levels where there are more bits available to encode them. That notion was quickly disabused because it puts the cart before the horse - the reason why the brighter parts of an image are encoded with more bits is because those tones have less noise and thus more bits are useful to represent/distinguish those tones. In other words, it's not more bits that improve IQ but the improved IQ that requires more bits. Using more bits than is necessary (specifically, more bits than are needed to encode signal vs noise) only serves to waste storage space. This is why there are so few bits to represent the shadows - the lower signal levels means few bits are necessary to distinguish signal/tonal levels from noise.I don't quite follow your reasoning but would like to understand it. Do you have a link with discussion where it "was quickly disabused"? All the links I find are supportive of EttR, for example http://theory.uchicago.edu/~ejm/pix/20d/tests/noise/noise-p3.html
I was describing the misplaced belief that ETTR improves IQ from the higher number of bits representing tones toward the right edge of the histogram. ETTR improves IQ by increasing exposure, which reduces noise - the less noise, the more bits that are needed to distinguish tonality. Summary: It's not more bits that improve IQ but the improved IQ that requires more bits.
p.12 #19Canon 5DS dynamic range: zero improvement (RAWs analyzed)
So in short, The Canon 5DSR sensor will score about 80 points on the DXo scale and have 12.4 stops DR max. Humm, not much improvement in sensor tech for about 8 years other than the removal of the banding issue. At least it's something. Of course, all this is pending more testing but the preliminary results seem pretty clear.
So, the 5DSR poses no threat to the Nikon's/Pentax's/Sony's in the sensor dpt., well, might take away a few sales but that's about it.
p.12 #20Canon 5DS dynamic range: zero improvement (RAWs analyzed)
RustyBug wrote:
Jimmy,
Thanks for the enthusiasm and info ... but, to my eye those wedding pics that were pulled up look pretty HDR-ish to me. I guess that's fine if you like that look, but not my gig.
Maybe a more deft hand @ PP would have done better, but my general thought on it is that there is no such thing as a Free Lunch. Pulling up from deep underexposure, while possible, still comes with a cost. I get the excitement, but I think it is also provides for some lesser workmanship. In the name of expedience and time = money ... I get that some folks will find that justifiable ... just not my gig, atm. I think a two shot bracket (occuring in a fraction of a second) is a better way to cover extreme DR situations ... that, or add light.
Some folks will prefer to work from the single file, pulling up hard, but I think you can produce better finished work from the two files when your DR is extreme. Even with a 5 EV pull possible, I'd still want to grab two shots and put them together, rather than one file pulled up that much. I get that it is nice to have it in your hip pocket ... but i don't think it yields optimal over well executed fill techniques or good brackets for the different lighting.
There's another aspect that most folks don't consider when dealing with extreme DR variance. It typically comes with a different color of lighting WB in the form of mixed lighting. Having the two shots makes for better color correction of the different color lighting source. Likely, I'm one out of 1,000,000 shooters that feels this is worthy to address ... but that's me. Well executed fill flash can serve to correct for color issues also.
The other aspect besides the color of the light is the contrast level of the light. The shade lighting is vastly lower contrast than the direct or background lighting, As such, it warrants a different set of processing to really bring it to its best. Having the second file to work differently makes it much easier to address the fact that your scene really has two different sets of lighting conditions ... but that's just me. Personally, I think the reliance on pulling up the shadows in lieu of either a bracketed exposure (or adding some light to the scene) is a recipe for a lesser product.
So, while I get the greater DR has usefulness ... it still doesn't resolve some of the other inherent problems that come with such extreme scenes. Atm, the adage of "just because you can, doesn't mean you should" comes to mind. If pulling up DR is the holy grail answer ... I guess we'll someday be able to dispense using flash or reflectors at all.
Of course ... I'll get beat by a myriad of folks who will argue that without the DR they wouldn't have gotten the shot at all. It's a valid argument, but I just don't know that the diff in DR makes the whole case. I mean ... photographers have been producing great images in high DR scenes going back to film @ either double exposures or with light augmentation. To my eye ... the 5 EV push / pull doesn't necessarily produce a superior product, albeit a somewhat more convenient one.
Would I welcome more DR in the Canon's ... most likely yes. Do I think pulling up 5 EV produces optimal results over other methods ... hmmm, not so sure about that. ...Show more →
Hi RustyBug,
The purpose behind the my supplied link was to show what was possible with having extended DR available in one's image files. Of course, (and I agree with you) the deftness of one's hand in how they use this extra head- and, er, "foot-" room will be a determinant with one's final results. The point of my response to you was to merely ask you (in a roundabout way) the question, would you rather "push/pull around +/- 2.5 EV", as your stated preference, with 11-stops DR, or do you see merit and the possibilities opened up to you in being able to do so with those extra 3-or-more-stops?
To your comment, "I guess we'll someday be able to dispense using flash or reflectors at all." Well, (and not to be argumentative here) some us already have. I shoot where the "wild things are", miles from a 110-outlet, and bring nothing more to my subjects than a camera and lens and I shoot them in whatever light is there when I find them.
@ "bracketed exposure"...Flapping wings and flittery critters are single-image subject matter for which multiple exposures is not a DR solution.
@ "different color of lighting WB"...This is addressable using the RAW file. And, I'm getting fairly good at finding a happy balance between the "invasive" dappled sunlight and the heavy green light cast found deep-in under the canopy!
Me enthusiastic?! You bet! I cheered on higher ISO's in my 5D2, I was ecstatic over the 8fps in my 7D, and now I want to play with some extra DR! These are exciting times! Send those tools my way!
But, the OP's test results tell me Canon still isn't offering that particular creative capacity in their line-up. Good to know.