You are right, i beg Florian's pardon.
Just now i try again the phenomenon.
So, it is: if...
Strange, because Florian described it with this words (#2 for your use case):
All right, the phenomenon not happens now. I continue to monitor. So it remains: Mp3tag 3.15, 32 bit, with turned ON Library.
Thank you for your help and time, ohrenkino.
LATER
Well, i already know/remember. The above problem has really disappeared, but with Library turned on, to save changes is always much slower, than without. Even then, i calculated: for me (with my habits) it is not practical to use Library. So i turned it off, again. (But leave 32 bit: i trust, i will notice, if there will be a memory problem. I do not load my entire Music directory very often.)
Is this also when you exclude the process mp3tag.exe from Windows Defender`?
From the documentation:
Windows Security intercepts each of the operations performed, so if you trust Mp3tag, making sure Windows Defender doesn’t scan the library database file on each change can speed up things significantly.
Thank you for your attention, @Florian.
I do not use Windows Defender. Its service is disabled in Services. Instead, i use Comodo Internet Security. The real time virus scan is disabled. Now i temporary disabled its firewall and HIPS (host-based intrusion prevention system, looking for processes, as you mentioned with Defender), too. No difference in Mp3tag's behaviour.
Test method
I turned on the Library again. Opened the whole Music directory, Mp3tag scanned it, then i closed the program. After i opened 1 folder from my directory, with 30 FLAC files. Selected all of them.
First method: started a saved Action Group, that adds field NOISY with value=1. Then i deleted this field. In both direction saving is slow.
Second method: started a saved Action Group, that adds to the end of filenames " (mono)". Then with reverse Action Group i deleted the addition. In both direction saving is slow.
Third method: i opened this folder in Total Commander. Selected all 30 files, and with the Multi Rename Tool added to all of them the above " (mono)". Then reversed, opening again the Multi Rename Tool and clicked "Undo the last rename" (it undoes on all the 30 files). In both direction saving is zippy quick.
When i tried in Mp3tag, i monitored Mp3tag.exe in Windows Task Manager: it shows many readings and writings. The system resources (CPU and RAM) were near to absolute free.
In my eyes, things point to writing the Library. My Library is 962 MB (36 thousand music files).
By the way, i did not try it, but interesting question: is there difference with 64 bit exe? Or it does not matter?
An absolutely small letters question.
Foobar's library with the same D:\\Music folder is only 32 MB, if i interpret right, and its location is c:\Users\Itsme\AppData\Roaming\foobar2000\library\
Does this modify a tag field?
Does this modify the filename?
If the answer is "yes" to both then I think you compare apples and oranges.
The action has probably to rewrite the whole file, the manipulation of the filename modifies a single entry in the file allocation table.
Salut, @ohrenkino!
Yes to both. But: i did the third test to compare with the second. It modify filenames, too (and only).
The second method (adding " (mono)" to the end of filename) modifies FAT, too; but, plus, rewrites (it must, and process monitor shows) the big Library file, too.
Of course it rewrites the library - otherwise the link between the data and the file would be lost.
And yes, it is slower to use MP3tag for such things as adding a string constant to a file name as MP3tag first reads all the tag data plus other properties while your file handler only looks at the FAT.
The test would be a valid one if you worked only with MP3tag, with
Each time yout have to take the time and see what is quickest. That would proove it there is a benefit using the library.
And just a personal note: if you really only want to rename files, then MP3tag is probably not the fastest tool. But if you do serious tagging, then I doubt that the file renaming tool is any good.
I mean, with your suggestion, you want to prove, that using the Library is faster to retrieve data. (Or am i wrong?) Of course, i agree, i even tried it (36,000 files: 25 minutes contra 2 minutes.) But i lose much more time with the Library due to slow savings, than i gain with readings data, when files loading. This was, when i meant: "for my individual habits".
Total Commander: in this case, i only used Total Commander as a test. I make all tags' and filenames' changes ONLY in Mp3tag. With Library turned OFF there is no speed problem with both type changes.
If we think about it, it is not really just my individual habits. The job of an Mp3tag is basically WRITING tags. Plenty of tags to read all the time is basically the job of the foobar. So, i do not feel any problem with not being able to use the Library. (Of course, there may be many people, who find it very useful. We all use the program differently.)
I don't know - how many changes could you perform in the 23 minutes that you saved by using the library.
And these 23 minutes will increase, the bigger your collection becomes and they apply each time you load the files.
The comparison between Total COmmander and its functions to rename files and the functions in MP3tag to rename files will only become a real comparison if you also test Total COmmanders tagging abilities.
OK, you claim that you have not benefit from the library (which I still doubt) and you are satisfied the way you work, so everything is fine, I would say.
I think (or for me), the Library would be really VERY useful, if we used it to search for existing records to make it easier to fill out the fields. We have talked about this several times, i am looking for a link right away.
For example: Tag Panel suggestion
Or: Feature suggestion: Allow user to store singer names - #2 by incifinci
Agree, absolutely. As i wrote the same, too.
My habit, when i get new files: about 20-30 times (yes!) to save changes, before new files will "all right" for my phonoteca. Because i use a lot of tags (30-40 or more), they all have to be searched for, typed in -- and always saved.
You are quite right: this topic has been discussed several times.
And I still wait for the real use case.
With the use of the library that helps to load your whole collection in just 2 minutes, you have all the template files at hand they you may ever need.
And for more specialized cases you have the quick actions, action groups and/or converters.
which means that the real time lag comes from the search and the time it takes to write the tag data to the files can be neglected.
I tried it thoroughly, when the Library came out. I do searches in foobar, writes in Mp3tag. So this is the fastest way for me, i tested it.
I wrote above:
About once a month.
For me, the Library would be worth, if there were those drop-down lists, based on it. Then it would be faster, than always to jump to the foobar. Which, by the way, is not so common way in my practice. What i need to copy, i open (most of them) in Mp3tag. I use a lot of export + import, for example like i described here: Data in REAL database, instead of tags?
Often, I modify the exported aaa.txt file in EditPad (or Pro) first (also a great program), before importing it into the new files.
This may be important. I saw a very OFTEN reads-writes. Does this mean, that the Library file has been modified, saved on disk so many times? Because if so, it might be more appropriate to save it only once, at the very end. (Sorry for the perhaps amateur post.)
I doubt that.
How many artists does foobar show you? They all would appear in the dropdown list I would assume - or, if it can be maintained by the user, you would have to remove the unused / unnecessary one-hit-wonders from the list.
Or you scroll endlessly through the dropdown list ("Was it 'Beatles' at 'B' or 'The Beatles' at 'T'?").
So I still think that a dedicated set of actions of the type "Format value" is the fastest way.
The library, BTW, was never intended for external access and purposes. Even the shorter loading time is just a welcome by-product but not an intended feature.
Good databases are random access databases where the sequence of records does not match a particular alphabetic order. Data is usually accessed via an index mechanism which helps to find the record position in the database.
So the reads and writes mean that the database handler made the pointer jump to a specific position and read the data from there.
New data can simply be appended at the end without the need to rewrite the whole database file.
Deleted records are simply flagged as deleted and not physically removed from the database. You have to trigger such grooming in File>Options>Library.
Agree totally, again.
Artist: i have actually 4,000. But, imagine the really perfect solution: type "t". For 3 sec delay nothing happens. This delay must be (even adjustable by the user)! I instantly continue with "les." And i get the Beatles, and everything where "tles" (full-text search!) occurs. I only have one thing left: to click on the line, that is right for me in the list. And the task is done.
I know, there is a huge resource requirement. But even more powerful help at work.
You wrote: if it can be maintained by the user, you would have to remove the unused / unnecessary one-hit-wonders from the list -- i am completely opposed to this, manual variant.