So it has been over a year since Fraps updated...

#1
Is this the perfect version of Fraps of all time? It seems like that there will not be updates for a while.

Windows 9 is rumored to release April of next year, so maybe there is a chance it will have another update to function on that operating system.

Since it has been so long, have I missed anything important or new about the program or has the development team found no changes to be made and/or went on hiatus?

Thanks!
 

deardevildude19

Sexy Duck
Well-Known Member
Site Contributor
#2
Well we do not know if Beepa are about to release a new version yet, it could be next week. But they are probably having trouble trying to get a better compatibility with Windows 8 currently. Give Beepa time
 

billman

Well-Known Member
Site Contributor
#5
Fraps uses some very old methods, I would not be surprised if it has reached a technical limit and would require a major overhaul to be perfectly stable on an OS that is over 10 years newer than it original implementation. I'm just thank full its still around I have tried many other screen recording apps and always come back to FRAPS because it just works.
Microsoft are making it hard for Beepa as the API it uses in now just a virtual layer inside a much bigger API.

The limited amount of coding I have done opened my eyes to the difficult nature of making a cross platform app work correctly and the drastic changes to Windows 8 it a true testament of Beepa skills that it even works at all.

My prediction is it will be 2 years before Windows 8 and 9 are fully supported.
 

De-M-oN

Well-Known Member
#6
I'm not sort of a person who demands new features in. If it works, I'm more than happy with it.
They should get rid finally of the fps multiple cap. Why limiting ingamefps to a multiple of recording fps? DXtory and other software shows clearly that this isnt necessary to do. Why I cant record in 2 seperated audio tracks for ingamesound and microphone?
Why I cant set myself the audio devices it should use? I want to record the direct WAVE output and not that "speakers" windows mix channel which is worse than using direct access to the wave channel (better compatibility and less latency)
DXTory all does offer this. Fraps needs much improvement.
It is also slow.
I have a RAID 0 which writes 352 mbyte/s.
When I record GZDoom with 2048x1152 I get the following ingameFPS while recording:

Fraps: 210 fps
MSI Afterburner: between 330 to 350 fps (doesnt matter if I use Lagarith YV12 or FFDShow uncompressed RGB24. The difference in fps is nearly unnoticable with my hardware and the game.)
DXTory: 900+ fps. (doesnt matter if I use Lagarith YV12 or FFDShow uncompressed RGB24. The difference in fps is nearly unnoticable with my hardware and the game.)

Fraps why are you so slow?

You haven't missed much, it's just been simple bug fixes and the change to a different AVI encoder to record in one big file (we still recommend to split by 4GB for many reasons)
Why? If game crashes the file is closed - only msi afterburner is an exception - but you can write the header manually to fix an unclosed file. At least with my method: http://www.letsplayforum.de/index.p...-fehlender-Header-index-keyframes-reparieren/

Some frames may be lost though if you use compression.

The 4GB splitting has disadvantage, because the HDD has to close the file and open a new one.. (unnecessary additional work and seeking for the HDD) And 4gb is very fast reached with high complexity and high resolution.
Avisynth operates faster with less video sources as well, especially when it comes to Multithreading (avisynth MT ( http://forum.doom9.org/showthread.php?t=148782 ) ) gets slower and more instable with many sources as well.
Then the next disadvantage of 4gb splitting is: There is always a slight difference at end of file between audio and video track length. This gets additional if you append lots of files. This asynchronous endings can lead to be heared in a longer video at the later part of the video. Our Sagaras Scriptmaker would extract the audio with corrected length though. But it makes sense to make it better directly instead of fixing later.
The difference between audio and videotracklength is a technical thing. A video of 30fps can be closed every 33ms (due to the 33ms framelengths) Audio can be closed directly. Also if the framelength thing wouldnt be - first the video track and then the audio track is closed (or vice versa) they are not closed synchronous. Theyre closed through a specific order. Even a camcorder cant close both tracks perfectly together. It are with a good hardware and good program only some ms, but it is there.

I hope I could explain it somehow understandable xD. German is my native language, not english.
 

Thalmor Wizard

Moderator
Staff member
Site Contributor
#7
They should get rid finally of the fps multiple cap.
No they shouldn't.
DXtory and other software shows clearly that this isnt necessary to do.
Except that it puts more strain on the system to calculate the true FPS.
Fraps needs much improvement.
No it doesn't, I like Fraps because it's simple and not flooded with needless crap like your precious DxTory >_>
It is also slow.
No it isn't.
I have a RAID 0 which writes 352 mbyte/s.
Good for you (sarcasm). Meanwhile, the majority of people do not use RAID setups.
Because, if a file is forcibly terminated by a program crash or a power cut, the file becomes unreadable. This is the way all software works.
The 4GB splitting has disadvantage, because the HDD has to close the file and open a new one
No it doesn't.
There is always a slight difference at end of file between audio and video track length. This gets additional if you append lots of files. This asynchronous endings can lead to be heared in a longer video at the later part of the video.
Thanks for showing that you don't pay attention to Fraps releases, as well as showing that you didn't search the forum for many fixes to solve the audio/video sync issues.[
 

De-M-oN

Well-Known Member
#8
Except that it puts more strain on the system to calculate the true FPS.
Yes this takes so much performance - wow my poor cpu.

No it isn't.
Then explain me the fps comparison pls. It clearly shows that fraps was the slowest one of the 3 programs.

Good for you (sarcasm). Meanwhile, the majority of people do not use RAID setups.
The RAID is needed to see the performance difference - to ensure that not the HDD breaks down the fps. And I used a game which isnt that cpu intensive and gpu intensive (gzdoom) - this ensured that the hardware doesnt lead to wrong results.

So you could really easy see how good the programs perform. If I wouldnt use a RAID 0 I would have different fps which lagarith yv12 compared to rgb24 uncompressed.
So yes - the RAID 0 is very important indeed to see the performances of the softwares.
If you have an extremely slow HDD, but for example 2 HDDs, then you even could record on these several HDDs to distribute the HDD usage. (DXTory's rawcap option)
But most drives can handle at least lagarith which uses awesome lossless compression.
Also especially with slower drives the fps multiple is even more annoying.
If you want to record in 30fps and you get 50 ingamefps, that fraps forces you to play at 30fps then.
terrible. And I had a slower HDD before without RAID 0.
Damn I loved dxtory for that so much, that it doesnt forced me to play at 30fps.

Because, if a file is forcibly terminated by a program crash or a power cut, the file becomes unreadable. This is the way all software works.
I already told you that it is fixable with my tutorial http://www.letsplayforum.de/index.php?page=Thread&threadID=88428

No it doesn't.
Of course the HDD has to close the active file and open a new one. you arent capture in raw video data and raw audio data. fraps uses file headers as well and keyframe flags and an index and is needed to get them playbacked later. Or what you think why you cant open it, if you dont close the file (power cut how you mentioned)

Thanks for showing that you don't pay attention to Fraps releases, as well as showing that you didn't search the forum for many fixes to solve the audio/video sync issues.
Thanks for not understanding my text at all.
thank you so much.
 

De-M-oN

Well-Known Member
#9
wow you're really getting likes for your post now? wow I gave you already a table of fps comparison and so on. Should I record it via a camcorder, that you believe me that fraps is slower? damn ..

And thats not gzdoom and RAID 0 exclusive lol - how already edited in.

Next example would be Descent 2 with older PC and an external USB 3.0 drive.

Fraps gave me there 90 fps ingameFPS.
With DXTory I had 120 to 150 fps.
With said older pc I had at trackmania United only 30fps with fraps
With DXTory I got 40 to 50fps But even 50 arent enough for fraps to get more than 30 if you record with 30. Because of his multiple thing.

But that pooooor cpu that has to count the fps lol! ->
Except that it puts more strain on the system to calculate the true FPS.
<-

Do you mean that really as a serious pro argument for this? :D Thats ridiculous.

That broke down my cpu so hard, that I could play the TM United game at 50fps instead of 30. So annoying was it to my cpu to count the fps. Because it is a more complex thing than encoding an x264 video with preset placebo, you know? Never let the cpu count your fps rate !!

Damn that argument made my day. sorry.
 

RobiePAX

Well-Known Member
Site Contributor
#10
I've been using only Dxtory for the past week, I am still learning it so I can't yet properly state my opinion of it. Yet I can tell this thus far... if I'd try to compare Fraps vs Dxtory and try to bring up an argument why should a user buy Fraps instead of Dxtory , so far I have only 1 argument which is Fraps simplicity in both operating it and learning it, which is good for newcomers. But that's not big enough argument.

In favour of Dxtory however I find significantly more points to make, and that's only for so far what I learned about it features. Same quality, same price, more built-in features(and actually useful ones such as benchmarking computer Hard Drive speed), codec selection tool(able to select Fraps codec too), MUCH better performance, hooks up to the same number of games without issues, customizable, etc. At this moment I freely can say this, Fraps is perfect and works as intended. However if Beepa won't consider to even the features and performance with their rivals then Fraps will die off quite fast. Right now Fraps lives solely on its rightfully gained reputation of being the first big boy when it came to video capture software. Yet stop aiming for perfection doesn't mean rivals will stop as well.

And that's really sad to say for me, a person who used Fraps for over 6 years. Even if I would still keep using Fraps I no longer would be able to claim that Fraps is the superior software in comparison to others without proof to support such a claim. There are many programs that came and went who tried to rival with Fraps but they nearly always had their few advantages and much more disadvantages. Dxtory is one of the cases that has only advantages... reminds me of song "Everything you can do I can do better". Most likely I will soon completely stop using Fraps unless I find seriously it disadvantages, though I will still lend support to people who are experiencing issues with it.
 
Last edited:

Thalmor Wizard

Moderator
Staff member
Site Contributor
#11
However if Beepa won't consider to even the features and performance with their rivals then Fraps will die off quite fast
Not really, Beepa should not just copy what everyone else is doing, the fact that Fraps is just simple and doesn't have all the complicated options that Dxtory has is a selling point: most people just want to get a program and use it, not spend time trying to work out what every setting does and which settings are used to start recording, which is what I get the impression from Dxtory. It just seems like the software which you need to be pretty advanced to understand, as I've heard reports of it being worse on the system because of all the settings it has.

For example, the ability to record at arbitrary resolutions would almost certainly cause high overhead, since the CPU is having to load game data, record and arbitrarily resize the incoming GPU stream on-the-fly: I don't want to know what kind of system you'd need to be able to accomplish that, certainly not a cheap system that any new YouTube user would be able to afford. Frankly, if you have a decent system that can record at 1920x1080 @ 60fps, there is no reason for you to be needing to use arbitrary resolutions in Dxtory to waste system resources (which is what I think a lot of people use the software for)
 

RobiePAX

Well-Known Member
Site Contributor
#12
Not really, Beepa should not just copy what everyone else is doing, the fact that Fraps is just simple and doesn't have all the complicated options that Dxtory has is a selling point: most people just want to get a program and use it, not spend time trying to work out what every setting does and which settings are used to start recording, which is what I get the impression from Dxtory. It just seems like the software which you need to be pretty advanced to understand, as I've heard reports of it being worse on the system because of all the settings it has (like arbitrary recording sizes, that is surely going to put a high overhead on your system, more so than Fraps)
I can personally vouch that performance difference alone is sufficient enough reason for any Fraps user at least to "try" Dxtory using the Lagarith Lossless Codec under YV12 mode(I am very surprised to the point where I find it stupid that this free codec is not being included automatically). Yes there are a lot of features that I personally don't use and at first they were scaring me as the beginner, all that Clipping, Cropping, Marging, etc. But generally it reminded of Handbrake interface since they are all optional.

I also agree that Fraps simplicity is its selling point, however as mentioned before there's not enough other reasons to still justify Fraps recommendation. Performance hit is way too great and other developers have proven it's not impossible to improve it and leave quality the same. Dxtory did not get their reputation by "copying" Fraps, but by making it do the same thing better and with more possibilities. It is not a Photoshop, within half an hour the user will be familiarised with the categories layout, and that half of the "scary looking features" are 100% optional. I find it welcoming that FRAFS AVI Info is working on Dxtory files, I feared it won't work as it didn't work with ShadowPlay, useful for reviewing numbers.

I don't say that Fraps has to change, developers made it pretty clear where Fraps stands right now. However I fear Fraps will be just another ICQ that will eventually be replaced with Skype. Hopefully people will understand that reference.
For example, the ability to record at arbitrary resolutions would almost certainly cause high overhead
Exactly, an optional feature that does not force you to use it. If you don't wish so then simply do not change any values and let it to record at the same resolution as the game is being played at. I honestly wouldn't call Dxtory very advanced, I found Handbrake more tricky to learn than Dxtory.
 
Last edited:

billman

Well-Known Member
Site Contributor
#13
I am one of the few that built my PC around the requirements for FRAPS, I cant deny there are more efficient capture methods than FRAPS however I like a stable FPS when recording.

When I was recording JC2MP with the beta version of afterburner it had a frame limiting feature in the RivaTuner Statistics server this allowed me to have a perfect 60 FPS during recordings. This worked much better than VSYNC as it kept the game from changing FPS without the input lag associated from VSYNC.
However this was removed so I gave up on Afterburner.

So form a game play standpoint FRAPS is the only screen cap software that enables me to have a flawless FPS while recording.
Not all games suffer like JC2 when the FPS changes but I personally prefer the overhead if it means my game controls are consistent.

During the beta testing of JC2MP I was able to win soo many races because I was able to keep my system at constant load.
I had more problems driving when I was not recording as my FPS would spike and drop depending on the scene.

But if you want maximum FPS while recording FRAPS is not ideal but for some stability is more important than peek performance.

Also if you realised fraps captures at a lower level than it rivals resulting in a pause until the frame is written as opposed to newer capture methods that poll the screen buffer and do not halt the system if a frame is not written in time, they will just drop the frame. This method was not viable on the P3 and P4 system when FRAPS was first implemented.
So yes FRAPS is archaic and was built around the hardware of the day.
To make fraps comparable to newer methods would require a new implementation and effectively a complete rewrite.

I do not wish FRAPS to do this as there are advantages to having well timed halts in processing.
One example is my video card runs cooler when I am recording at 60 as opposed to the uncapped FPS.

The MAJOR drawback to fraps is that problematic devices like USB or a flaky sound or network controller will cause frames to be dropped so its imperative to ensure you system is operating smoothly.
Things like overclocking and hardware monitoring can also mess with FRAPS.

To overcome these issues I have disabled Speed Step, Hyperboost and Sleep States.
My PC runs at a flat 4 GHz and I do not see the audio sync issues you and other experience.
I have done 3 one hour recordings with my PC jacked up to 4.2 (40x105) audio sync was perfect right to the end.
So I can confirm this is not an issue with FRAPS but most likely constant changing in clock speeds (something did not happen when FRAPS was first implemented) that results in sync problems.
Newer software take this into consideration and expect your clock speed to jump so alternative methods of sync are implemented.

So yes if your PC is not rock solid and not capable of real time work then DXTory, Afterburner and the like will produces much better results. But for me the difference is negligible.
 
Last edited:

De-M-oN

Well-Known Member
#14
I cant deny there are more efficient capture methods than FRAPS however I like a stable FPS when recording.
and? Possible with both other programs as well. Both have Framelimiter.

When I was recording JC2MP with the beta version of afterburner it had a frame limiting feature in the RivaTuner Statistics server this allowed me to have a perfect 60 FPS during recordings. This worked much better than VSYNC as it kept the game from changing FPS without the input lag associated from VSYNC.
However this was removed so I gave up on Afterburner.
The Framelimiter is still there.

So form a game play standpoint FRAPS is the only screen cap software that enables me to have a flawless FPS while recording.
No.

Not all games suffer like JC2 when the FPS changes but I personally prefer the overhead if it means my game controls are consistent.

During the beta testing of JC2MP I was able to win soo many races because I was able to keep my system at constant load.
I had more problems driving when I was not recording as my FPS would spike and drop depending on the scene.

But if you want maximum FPS while recording FRAPS is not ideal but for some stability is more important than peek performance.
But imagine you record with 30fps recording fps and your ingamefps is swapping from 55 to 65 back and forth. Then Fraps constantly changes the lock 30, 60, 30, 60. back and forth. This can lead sometimes to a slight stutter and you have a greater fps swapping than without this. If you want such things, use framelimiter.

Also if you realised fraps captures at a lower level than it rivals resulting in a pause until the frame is written as opposed to newer capture methods that poll the screen buffer and do not halt the system if a frame is not written in time, they will just drop the frame. This method was not viable on the P3 and P4 system when FRAPS was first implemented.
So yes FRAPS is archaic and was built around the hardware of the day.
To make fraps comparable to newer methods would require a new implementation and effectively a complete rewrite.

I do not wish FRAPS to do this as there are advantages to having well timed halts in processing.
Checkmark in DXTorys options Synchronous Surface Lock and the problem is gone. Performance stays the same.

One example is my video card runs cooler when I am recording at 60 as opposed to the uncapped FPS.
But you have to play with 16ms delay (which is also at the controls then). For fast shooters like Quake 3/Quake Live or fast arcade car races you may wish to have a shorter frametime.

My PC runs at a flat 4 GHz and I do not see the audio sync issues you and other experience.
Damn. I dont mean asynchronous record by hardware problems. I meant that a record always has a slight ms difference in tracklength video vs audio.
This is a technical reason. Even a camcorder cant close the file perfect in sync of perfect zero ms difference.

I have done 3 one hour recordings with my PC jacked up to 4.2 (40x105) audio sync was perfect right to the end.
Then the slight different endings are small enough to not get hearable. They're there though. If you didnt use the 4gb splitting - then of course you have no hearable difference, because only one ending.
So I can confirm this is not an issue with FRAPS but most likely constant changing in clock speeds (something did not happen when FRAPS was first implemented) that results in sync problems.
Newer software take this into consideration and expect your clock speed to jump so alternative methods of sync are implemented.
My clock of GPU and CPU are strictly fixed.

So yes if your PC is not rock solid and not capable of real time work then DXTory, Afterburner and the like will produces much better results. But for me the difference is negligible.
My PC is quite fast, especially with the RAID 0 and you see the difference in speed in my comparison.

@RobiePAX
able to select Fraps codec too
No you cant record with a decompressor. I dont know why beepa made it listed in vfw system. Shouldnt be like this.
Their encoder isnt available from outside though. Same for DXTory. DXTory Codec only usable with DXTory.

I can personally vouch that performance difference alone is sufficient enough reason for any Fraps user at least to "try" Dxtory using the Lagarith Lossless Codec under YV12 mode(I am very surprised to the point where I find it stupid that this free codec is not being included automatically)
The dxtory codec is way faster though, because its somewhat lower compression you get less cpu load and even more fps (if your HDD is not mega slow. Otherwise you may be better with Lagarith)
DXTory Codec is fastest with 2 processing threads @ yuv420 mode and compress checkmarked if no RAID 0 uses. Otherwise you can further improve speed with disabling the compression.
 

billman

Well-Known Member
Site Contributor
#15
Thank you for the correction on the FPS limit in Afterburner, in first version I used it was out the front and now it hidden behind the spanner Icon. I am grateful for this correction the I have experimented with many frame limiters most produce undesirables effects at some point. The afterburner frame limiting was the most effective for me.

I must admit I have not played with the most recent versions of DXTory and by the sounds of it I need to just for experience sake. I found the earlier versions had strange issues with lossless output and seemed only to work reliably with it own codec.
But I will give it another run, most interested in use of system codecs but this requires considerable understand to configure.
For the average person this is overwhelming and become hours of trail and error.

When it comes to game play everyone is different, I see rare dips to 30 FPS when doing something truly over the top, not during typical game play conditions so I can handle the flipping here for the sake of a frame every 16.6 ms apart. I not to upset what FPS i'm at just needs to be consistent. Yes some people have the need for quicker frame times and if you have the horsepower fraps will exceed the recording frame rate, some situations it will allow FPS to float but most times locks to nearest multiple. So fraps can still record at 60 FPS and play at 240 FPS. But really if you want to play at over 60 and capture at the same time I personally recommend PCI-X capture cards in a dedicated capture rig. Just my thoughts.

On the audio sync problem you describe I have never seen it with my FRAPS outputs I guess I must just be too slow to notice something that "have no hearable difference". Also I examined both the video and audio streams of my test videos septate and got the same lengths, I was not splitting at 4 GB. I will repeat this test and split the output when I get time.

So how about we leave it at FRAPS just works and is great for normal people and DXTory is great for freaks that tweak.
 
#16
well, i suggested to FRAPS privately (since i was kind of rooting for them for being first) that they implement realtime x264 capture, but at this point, whoever can do it probably "wins" (other than nvidia being proprietarily "first")

DXTory is great for freaks that tweak.
nothing wrong with that!
 
#18
i did a quick and dirty test that i've been meaning to do for a while ... there's some odd bottleneck when i try to encode game footage as fast as i can...

on 1 thread with MeGUI, i hit a cap around 27.6 fps encoding 1080p60 with "ultrafast" settings. in this case, it looks like it's the intermediary system, avs4x264mod, that lets MeGUI use 64-bit x264 with 32-bit avisynth.

commandline x264_64, i hit only 15fps with the commandline copied from MeGUI. however, i CAN hit 29.7 fps if i use 2 threads (and another odd anecdote, i hit 18.5 fps using 32-bit x264). still, a far cry from realtime 1080p60 (which Nvidia's Shadowplay can do on GPU (i'm assuming no CPU?))

i'll admit, this is EXCEEDINGLY quick and dirty, as they're probably all different versions of x264. tho the x264_64 IS a very very new build.

edit: looks like MeGUI gets 29.3fps with 2 cores.

i'm thinking one possibility is the FRAPS codec, maybe?
 
#19
well, it looks like FRAPS might be faster than Lagarith ...

some more quick tests, encoding direct from Lagarith 1080p60 file resulted in 8.2fps (yv12 lagarith file) and 4.5fps (rgb lagarith file) in commandline. and slightly slower in MeGUI.

here's some info on the setup:

the source that's being read is a 1080p60 FRAPS capture (or Lagarith convert) on a 4gb ramdisk. commandline encodes are written to NUL. the MeGUI encode is being written to HDD.

and one more update: updated the x264_64 used in previous tests from 7/2012 to 5/2013 (the newest on x264.nl ... i guess it wasn't as new as i thought), and got 2 threaded encodes up to 41.2fps (original FRAPS video)
 
Top