The Cubase 10 generated PDF is full of 0.06pt width lines - Adobe Acrobat DC Pro 2019’s Preflight tool finds 10569 of them on page 1 of the original poster’s second PDF. The rendering of such ‘hairlines’ is renderer implementation dependant - many PDF renderers will thicken these lines. Opening page 1 of the PDF in Adobe Illustrator CC 2019 reveals that each barline is made up of 26 of these hairlines next to each other and each line of the stave is made up of 7 of these hairlines next to each other. In both cases, this really should be a single stroke of non-hairline width. I regard this is a bug in the way Score Editor draws these lines.
I generated a PDF from the Cubase 10.0.20 Score Editor on my Windows 10 system using File → Print and the Adobe PDF printer (part of Acrobat, so arguably the canonical ‘print as PDF’ driver) which resulted in a similar mess - a document full of horizontal and vertical hairlines.
This, with respect, was my point about Cubase’s Score Editor being of limited quality, at least when it comes to generating print and PDF output. Cubase generates output that will render inconsistently, unlike specialist software for notation and engraving.
I’m not a specialist in graphics tech, nor do I know why the SB devs have engineered it this way– but if there is an option in the pdf reader to turn off a feature, which then permits the display to be correct, it doesn’t seem like something that impacts output quality from a layman’s perspective.
I suspect the reason Cubase uses multiple parallel hairlines is some sort of legacy reason - when the code was written, there was some performance or visual rendering issue with drawing a single non-hairline stroke or a single filled rectangle.
It is hard to call what Cubase is doing a bug - the code works, even though the results are visually sub-optimal in many cases as well as being non-deterministic when rendered (in particular because the rendering is heavily resolution dependent). The PDF 1.7 specification strongly recommends against using 0 width lines (this is a stronger statement than in older versions of the specification) which render at 1 pixel no matter what the resolution.
To confirm what Illustrator and Acrobat were telling me, I deflated the original poster’s PDF and checked it manually in a text editor. The horizontal anv vertical lines are 0.06pt wide, which is 0.021mm or 1/1200 inch - exactly 1 pixel wide at 1200dpi and subpixel width at any higher resolution. Such lines behave similarly to a hairline even though they have non-zero width.
I would term this an issue in Cubase. It would be far better if the vertical and horizontal lines were a single stroke of non-hairline width. Such a change will not magically turn the output of Score Editor into something more closely resembling the output of MuseScore or, better still, Dorico, but it will at least provide more deterministic rendering. If such a change were made there is then scope for tweaking the proportions of the lines in a way that addresses the original poster’s complaint.
@David W
Your analyse of the PDF is correct.
I too would have preferred the lines to be made of single strokes.
The fact is that, ideally, the setting to enhance thin lines for display shouldn’t be enabled by default in the first place.
For your information, this is an arbitrary decision made by Adobe to work around an issue with documents created by applications such as those from Office suite where thin lines would disappear when viewing in Acrobat (other pdf readers got the same setting).
In absolute terms it’s not a good choice because the display is not faithful to the original document.
But Office users being the biggest part, as opposed to the other parts such as architects etc, their choice is understandable.
Furthermore, what I notice here is that your first statement is clearly based on prejudice, and you are trying to justify it by analysing the documents afterwards.
Yes but it does much more than only “work”.
I suspect, respectfully, that you have no experience with the Score Editor.
It would probably surprise you but just know that the Score Editor is quite capable of producing the same “quality” than the other dedicated software you mentioned.
I think you have no idea of the capabilities of the Score Editor and you are most probably just relying on what you can read here and there from people who don’t know the software better than you.
That’s true. Single stroke lines would be preferable due to the reasons mentioned above.
But a line made of multiple thinner lines is still a line and should look the same.
Anyway this characteristic does not affect printing on paper.
Thus, and as you quite rightly said, that would not change anything about the look of the produced score from the Score Editor, which, could be as good as that of the other dedicated notation software, depending on the operator and his mastery of the program.
You are taking shortcuts in assessing the quality of the Score Editor.
The problem with hairlines, and the reason why the PDF specification recommends against their use, is that their rendition depends on the resolution that they are rasterised at. A line that is always 1 pixel wide is going to be of very different thickness if rasterised at 2400 dpi than at 150 dpi. Cubase generates lines that have a small finite width - but a 0.06 point (1/1200 inch) line behaves just like a hairline at resolutions of 1200 dpi or lower.
Thickening hairlines is double edged, as you note - it is arguably a loss of fidelity, but in most cases such a narrow line is not what the user really wants. A 1/1200 inch line is all but invisible.
There is no reason for Cubase to use such hairlines - and their use is sub-optimal.
You are taking my comments out of context. I was commenting only on “engraving quality”. Cubase’s score editor is a remarkably rich environment that is often not given the credit it deserves. Having such powerful scoring facilities built into your DAW at no extra cost (so long as you own Cubase Pro or Nuendo) is a tremendous and often under appreciated part of Cubase. Indeed, score editor supports tablature, which is more than Dorico 2 does!
For the purposes of comparison, I quickly entered lucatron’s three page fragment into Dorico 2 - the attached Zip file contains the PDF score, PDF parts and MusicXML score. I have assumed the material is a string “quartet” and the violas should be playing an octave down from the notated pitch (I’ve resisted moving the material to the alto clef). Of course, you would need three violas to play the viola part as written!
Perhaps the intended players are two violin lines (with at least two players on the second violin part to play the octaves at the end divisi) plus a grand stave instrument (though maybe not piano - it would not have sufficient sustain for the right hand part as written). I’m not too bothered about it - this was only done as an engraving test.
There are no manual engraving tweaks whatsoever in any of these files. I chose the rastral size for the full score layout and left the rest to Dorico. I’ve left the rastral size of the parts at Dorico’s factory default setting. One of Dorico’s great strengths is that it needs less manual tweaking to produce a good looking score wihout any object collision than its competitors.
As well as the cleaner horizontal and vertical lines, the things that jump out to me comparing the Dorico full score to the Cubase 10 PDF are:
better appearance of the ties - Cubase’s ties are rather thin and sometimes collide with the notes (see especially bar 17 at the top of page 2)
better handling of the second violin octaves from bar 27 onwards - Cubase has a collision between the first and second violin staves in bar 28 then unequal spacing between the staves in the next system
better vertical distribution of staves and systems, which improves the ability to follow the music
better horizontal spacing of the dotted rhythm in the third beat of bar 3 of the cello part
I realise that opinions on engraving are somewhat subjective - this is my two pence (as a Brit - two cents in many other countries)! Steinberg forum example.zip (215 KB)
The score posted by lucatron was completely “raw”.
(I don’t think it’s a string quartet, I guess it’s POP music and the instruments are synths and bass. Anyway…)
I also did a quick export from Score Editor by importing your XML (by the way thanks for saving me time).
Almost no manual tweaks. I just added some texts and used auto spacing in the Score Editor.
This has become a lengthy post, which deserves a summary.
It is clear TheMaestro and I agree over the power and overall value of the Cubase Score Editor. It is a hidden gem that anyone with notational needs who uses Cubase Pro or Nuendo would do well to explore. Even if you have and regularly use other notation software there are times when it is helpful to have notation functionality in the DAW, rather than having to round-trip in MusicXML. (In a similar vein, I would encourage people to explore other built-in facilities of their DAW, such as Cubase’s Channel Strip and sampler track, rather than always reaching for third-party plugins).
My comments about the limitations of Score Editor were always intended to be about engraving quality only, though I appreciate I could have made that more explicit.
Whilst I doubt that Score Editor is likely to receive substantial additional features or undergo significant improvement, especially now that Steinberg have a dedicated notation product in Dorico, there is still room for issues to be addressed. I still hope that -steve- will report the multiple parallel near hairlines in Score Editor output as an issue to the developers. Replacing these parallel lines with a single line of non-hairline width would reduce rendering ambiguity, potentially improve rendering performance and allow for optimisation of the width of the stave and bar lines.
Dorico 2.2 produces aesthetically pleasing and best practice compliant engraving of lucatron’s fragment with no manual tweaking. However, as TheMaestro’s reworking in Score Editor of the MusicXML I exported from Dorico shows, a small amount of tweaking from a user familiar with Score Editor can produce good quality results.
As an aside, it is very helpful if those complaining about Score Editor’s output provide a MusicXML file of the problematic music fragment, as this allows reworking in Score Editor as well as import into other notation software.
Textual communication online isn’t always the easiest format to communicate clearly. We were of the same opinion all along about Score Editor - it just didn’t come across as such. Certainly I can appreciate why you understood my previous comments as a general commentary on Score Editor rather than them being limited to engraving quality.
Cubase’s Score Editor is a hidden gem, though it is easily overlooked in favour of dedicated notation software. I suspect a lot of Cubase Pro and Nuendo users don’t even realise it is there. Part of its power is that Score Editor is sitting there in the DAW. There seems to be no near term prospect of Dorico getting any closer integration into Cubase and Nuendo, even though Dorico is a Steinberg product. The most that seems likely in the near term are improvements in MusicXML round trip experience.
MuseScore is remarkably good, it is free of charge and is making rapid progress now that there are full-time developers working on it. In my experience it is the most commonly recommended of the free notation options and it seems many music students find it adequate for their needs unless they reach a level of composing, arranging, orchestration and/or engraving where a commercial product provides sufficient time savings and/or allow for necessary additional functionality.
Dorico is outstanding for notation driven workflows and it typically gets much further with complex engraving tasks before you have to start tweaking manually than its competitors. Dorico’s two main drawbacks are price and the gaps that still exist in its feature set (it is, after all, still a young product). Notably, these gaps currently include guitar tablature - I expect this will be addressed in Dorico 3. One big advantage of Dorico is that it can likely produce your desired result quicker than competing products if it is capable of doing what you want, though to get to this level of efficiency you have to learn to do things the Dorico way using your keyboard and, optionally, a MIDI keyboard. You will find Dorico a struggle if you try to carry over workflows from competitor products, which are often more mouse centric.
Finale and Sibelius have their fans, though both are expensive mature products that seem unlikely to overcome their shortcomings without a significant amount of disruptive change.
If you have Cubase Pro or Nuendo and have any requirement for notation, I think we would both recommend exploring the Score Editor at least sufficently to have a good idea of its strengths and weaknesses for your needs. “For your needs” is important - we are all different, and what works for one person will not necessarily work for another. If Score Editor does all you need, there is really no need to look anywhere else. Even if it cannot meet all your needs, sometimes integration into the DAW will win over other considerations. I doubt that Score Editor will gain additional functionality or undergo significant improvement now that Steinberg have launched Dorico.
MuseScore is always worth experimenting with if you think it has anything at all to offer you. It is free, so any time you spend becoming familiar with it will not be wasted on expiry of a trial period. On Windows 10 you can install it very quickly from the Microsoft Store.
If you can live within the limitations of Dorico Elements and can afford its modest cost, it is perhaps worth a 30 day trial. Dorico Pro is very expensive - approximately the same price as Cubase Pro - and is far from cheap even if you qualify for a crossgrade and/or educational pricing. Unusually for Steinberg there is an education discounted crossgrade to Dorico Pro; with other Steinberg products educational pricing is only for full licences. I believe everything I did to enter and engrave lucatron’s fragment is possible in Dorico Elements. Unfortunately I cannot run Dorico Elements with my Dorico Pro licence as Dorico works like Cubase Pro and Artist - there is a single installer for both editions, with the edition that runs being determined at launch time by the licence(s) that are available.
Agreed - I don’t think it’s a string quartet either. Dorico requires you to assign one or more instruments to a player before you have a staff for that player to enter anything on to. Using the string quartet preset and changing the clef of the “viola” was a quick way to get something into which I could enter the music in Dorico. If I knew what the intended instrumentation was, I could easily move things to the correct instruments in Dorico. As a cellist, I know the bass clef part would make sense on the cello, though doubt it is intended to be a cello part.
Including the MusicXML was deliberate, as I hoped it would allow you to experiment without having to re-enter the music.
Your rendition is greatly improved over lucatron’s original PDF. It addresses the second and third of my bullet points, which are the biggest flaws in the original PDF to my mind. I prefer to have some bar numbers - I presume they came across from the MusicXML.
I still find Cubase’s rendition of the ties to be somewhat messy - especially the “viola” part in bar 17. However, it is perfectly easy to understand and I appreciate that part of my objection is likely because I am used to Dorico’s output.
I prefer Dorico’s rendition of the last two beats of the “cello” part of bar 3 to Cubase’s. Cubase spaces the noteheads of the third beat equally, though the notes are a semiquaver and dotted quaver. Dorico spaces the noteheads in such a way as to put more space after the dotted quaver than the semiquaver, which reflects the relative length of the two notes. It is a small thing, but it aids understandability to my mind. Again, part of my objection might be my familiarity with Dorico’s output.
Having mentioned MusicXML as an interchange format, it is important to note that MusicXML is imperfect, not least because there is not a single unambiguous way of expressing some concepts. The W3C Audio Working Group are exploring where to go from here - whether that will be a new version of MusicXML, the MNX textual format that is being worked on, or both is unclear; certainly as of Musikmesse 2019 back in April both options were being pursued in parallel. I can see MNX-Common eventually displacing MusicXML as the premier interchange format of choice for western style music, but that is likely to be several years off. MusicXML as it exists today does an excellent job of moving the key elements of scores between applications, though you might land up having to add dynamics and other similar markings back manually.
These limitations of MusicXML underline the value of having notation facilities integrated with your DAW. Sometimes, avoiding round trips to other applications is worth putting up with the shortcomings of the integrated facilities.
