Converting to edit friendly video options and differences

Hi all, using my phone as my recording device, 4k60fps, but apparently it is variable refresh rate so I am getting the message to convert. Originally I thought I wanted “best” which is ut video. Now it’s a little faster than “good” which is h.264 but the file size is huge. My source video codec is h.264, am I actually losing anything if I just use that when converting? Other than it taking a little longer? If there will still be some loss in quality how much is there and how much of a difference between good better and best?

Sorry if dumb question, I am pretty knew to video editing, but would really like to upload to youtube, reasonably, the best video quality I can. Do any of the other options matter or just leave them to default? I really can’t find much detail on this.

Also, is there a way to convert multiple files? Dragging them in one by one will be annoying, it took 24 minutes to do a 30 minute video and I have probably 8+ hours of video to go through split up across a dozen videos.

Thanks.

1 Like

No. There was a discussion on this here:

Also, see here for some useful advice:

That doesn’t exactly answer my question but thank you. I think I am going to play around with another editor or maybe a third party app for conversions.

When you say “a little faster” what do you mean? The export time to make them “edit friendly” or how it behaves in the timeline?

No, you won’t. You can actually just stick with the H.264 Convert To Edit Friendly option and you’ll be fine especially since disk space seems like it’s an issue for you. If space is not an issue than use the “better” or “best” options.

Which default are you referring to? The “good” (H.264) option in Convert To Edit Friendly?

If you have more than one file in your playlist then there is a batch export option named Each Playlist Item that becomes available to you in Export > From. There is an H.264 export presets in the lossless category that you can use to batch convert a bunch of video files although you’ll have to tinker with it a bit to get it to be like the H.264 in the Convert to Edit Friendly menu.

@Austin, is the only thing thing that needs to be changed in the H.264 lossless export preset to match the good option in Convert to Edit Friendly the audio codec from AAC to AC3? And raise the audio bitrate to what? Would 512k be enough? Is there anything else?

Thanks for the reply.

When I said “a little faster” I meant how long it took to encode to the edit friendly version.

I’ve learned a bit, or at least have been reading a lot about codecs since I posted this. You said that h.264 VFR source converted to h.264 CFR(aka using the shotcut built in convert to edit friendly ‘good’ option) should be ok, and that is what I am leaning towards, but from what I’ve read I thought the conversion will compress things again causing a reduction in video quality? I was experimenting with shutter encode so that I can set more options that shotcut allows but admittedly it takes quite a while. I am going to experiment with quality in shutter vs shotcut. Also, a lot of people say that editing in h.264 is a bad idea so suggested I actually convert to ProRes, import and edit, then export to h.264 again to upload to youtube. Any thoughts on that? If the only reason h.264 is not recommended is because it is cpu intensive to work with then I’m probably ok as I have a pretty high end system, ryzen 5900x, 32gb ddr4 3600 ram, gen 4 nvme ssd… then again I am doing 4k60fps…

As for options I was referencing the advanced options, use sub-clip, deinterlace, override frame rate, fps, frame rate conversion, and convert to bt.709 colorspace.

Thanks again for the reply.

So I just did a conversion in shotcut. Source is H.264 3.10GB 4k/60fps variable, total bit rate ~72000kbps and length of 6:09. Selected the H.264 option, took 4:25, size 7.25GB with a bitrate of ~173000. That bitrate seem extensive, do you happen to know how Shotcut determines it? In Shutter encode I can manually select the bit rate or doing a constant quality setting where it will determine it automatically. We don’t seem to have that ability with the basic tool that pops up but I’d like to understand this and I think it would tie into your suggestion on how to do batch conversions.

The H.264 that they mean is a regular H.264 final encode which means not with the very high bitrate that the Convert To Edit Friendly option will give you. Also note that the description for the good option for H.264 says “I-frame only” which means it’s very little compression. So the very high bitrate is to give you a file that isn’t going down a generation image-wise and the I-frames only is to give a file that will be much easier to scrub around in the timeline as it’s not compressed like a normal H.264 file would be.

Codecs like ProRes and DNxHR, Ut Video, etc… are fantastic codecs but they will enormous sized files especially for 4K 60fps videos. Those are especially best if you have to take that video and re-encode it again for use in another program and bring it back to an editing program. Like let’s say you want to take a piece of footage from your timeline and do some compositing for advanced special effects then bring it back to your editor to continue editing. That whole trip from the editor to the compositing program and back to the editor would be done with codecs like ProRes and DNxHR, Ut Video, etc because the video quality will not go down in any generations throughout that whole trip. If you have enough space for it then feel free to use them but if you don’t and you’re not going to be taking the footage around in other programs to use again and again before making your final export then the H.264 option in Convert To Edit Friendly will serve you just fine.

One thing also to keep in mind is that if you use codecs like ProRes and DNxHR, Ut Video, etc… for 4K video you might not be able to even play it on most computers. The bitrate might reach such high levels combined with the high resolution that it might stutter and lag just trying to play it. That’s happened to me. So in those cases the “good” option in Convert To Edit Friendly would be preferable because while the bitrate will be very high, it will still be playable even at 4K 60fps on most computers.

In your specific case, those options don’t seem to be relevant.

  • Sub-clip is if you want to only convert one section or several separate sections of a video file and not the whole thing. Good option if you know you aren’t going to be using the whole file for export and you want to save some space.

  • Deinterlace is if you happen to have a file that is interlaced and want to produce a video file to correct that.

  • Override frame rate is if you want to convert a video file to another frame rate.

  • Convert to bt.709 colorspace is for video files that are brought into Shotcut that are HDR since Shotcut at the moment does not support HDR. It’s in the plans though.

Ok lot to unpack here, but quick side note first. I was actually googling GOP and found a post from you on here from 2018 where you didn’t get much of an answer, but it is obvious you have leveled up a LOT since then.

I’m someone who really likes to understand things so was looking at the export options. There are several H.264 stock options but then under lossless there is a version as well that has GOP frames set to 25 so I am not understanding how that would be I-frame only? When I google it there is usually two values, m and n, and if I understand then both should be 1 to get I-frame only? The other reason I ask about this is because you mentioned in an earlier response that I could batch export these to get my constant frame rate but I’m not sure what would need to be modified to get it exactly like the convert to edit friendly version and suspect it has something to do with the GOP so it all seems to be same topic.

Thanks again, got a lot of notes to write.

Thanks. :grin: Although my knowledge is still nothing compared to some other members here regarding issues like codecs.

Like I said before:

The H.264 in the lossless category is not the same as the one in Convert To Edit Friendly. That’s why in the next paragraph I tagged @Austin who is very knowledgeable about issues like codecs to ask what to modify in that preset to get it to be like the one in Convert To Edit Friendly.

The answer was linked to above. Here you go again one last time

GOP frames set to 25 so I am not understanding how that would be I-frame only?

It is not. Lossless does not always mean I-frame only. Lossless H.264 with GOP=0 or 1 is going to be heavier than what the Convert dialog does, which is x264 crf15 = 70% Quality.

When I google it there is usually two values, m and n

That is a legacy convention/notation from MPEG-2. Basically, N is GOP, and M is the number of B frames + 1. When GOP is 0 or 1, the B frames option is not applicable.

1 Like

There it is indeed. :slightly_smiling_face: I should’ve clicked on it myself. I just took @Kaje68’s word that his question wasn’t answered.

Sorry for bothering you with a needless tag, @Austin. :sweat_smile:

1 Like

Thanks for response.

I know it is frustrating to deal with newcomers but I didn’t follow that thread before because he began with “no” so assumed it was just discussing why the option didn’t exist yet. Thankfully I expanded your comment this time and realized the error of my ways.

Also keep in mind that sometimes a more rudimentary understanding(things someone at your level possible take for granted) needs to be had to understand some of the responses linked to. It doesn’t help that I am someone who doesn’t just want to know how to get the result I want, but why it works that way and your explanation of GOP and B frames combined with the GOP entry on wiki was extremely helpful. As was your explanation on lossless != i-frame only.

So I think my understanding of how to get to i-frame only is where it needs to be now, and how I can “batch” converting to edit friendly by using the export each playlist option and setting a custom codec with GOP and B frame to 0 or 1. My only follow up questions are the “fixed” option. I can’t find anything on that on how it relates to GOP? The other question was the quality percentage/crf. I don’t fully understand how those values directly relate to quality of resulting video but had assumed that 100% was ideal, yet the post above showed it as 70% to match edit friendly conversion options. Is it a diminishing returns kind of thing? And 70% of what exactly? I know the lower the CRF the higher then quality which I assume means higher bitrate and larger file size but is that only difference?

Thanks again, sorry it took me so long to get to where you wanted me lol.

My fault, but since he started with “no” I had assumed it was just a discussion as to why. I think I have the basics of what I need. Though I have a few other things I’d like to understand.

CRF sets the bitrate according to the source file and the codec being used. A 0 or 100% CRF would be the complete encoding power that the codec can do but that means the bitrate will be at a max which results in full blown lossless file with a huge file. A CRF of 15 or 70% for H.264 is known as “visually lossless” which means you cannot see with the naked eye any difference between the original file and the exported file at H.264 with 15 CRF. But a CRF of 18 or 64% for H.264 is already hitting visually lossless territory which means bringing it to 15 or 70% is just there to give a bit more assurance with the visual state of the file while still not giving you the crazy high file size you would at 0 CRF.

So now that you know what the settings are for the “good” H.264 Convert To Edit Friendly option, load the video files you want to convert in your playlist then go to the Export menu and follow the settings linked to above. Then in From choose Each Playlist Item then set how you want the files to be named in the menu that comes up and you should be good to go. :slightly_smiling_face:

Awesome thanks. The only question left is the “fixed” option. I am exporting a video now with those settings, I want to see if the time is about the same as the edit friendly option and if the files are the same.

A GOP of 1 is inherently fixed. Otherwise when greater than one and not fixed, then a codec typically uses scene detection to decide to include a new key (I) frame, and the GOP value is a maximum.

Ok that makes sense.

As for exporting vs edit friendly conversion, I just did a little experiment.

I used the edit friendly conversion option and my output was a 7.42gb file, 59.94fps, and total bitrate of 172934kbps

I then tried to export the same file using the settings in the screen grabs linked above. On the video tab I set my export to the 4k resolution I started with, set the FPS to 59.94(not sure if I should do this or just go straight to 60) and left the rest to what they were, progressive, YADIF, Bilinear. I did not get the same results. that file is 5.25gb with a total bitrate of 122232kbps. It also took about 50% longer to complete. Did I miss something?

Reading through the rest of the other thread it looks like in the source code for an older version that cf 11 was used, I wonder what it is with 21.10.31

Newer version source code has crf=15 so I’m not sure what the difference is. Guess I’ll keep looking at source.

For clarity, GOP would be 1 and B-frames would be 0.

Use whatever the source files are. Camera sources will likely be 59.94. A screen capture from a computer game will probably be 60. Mixing them up will cause an occasional dropped or duplicated frame to make up the difference.

In the Export panel > Advanced > Other tab, there is a line called preset that is set to medium in the Convert to Edit-Friendly tool. If you used one of the export presets as a base for batch conversion, it may have had preset configured with a different value. This would have a significant effect on file size and encoding time.

I assume this was intended to create smaller, more manageable file sizes for the common source resolutions (meaning HD 720 and up). CRF 11 would be well past visually lossless for anything 720+, meaning extra processing time and larger file sizes with little appreciable benefit. However, for resolutions below 720 lines, CRF 11/12 often becomes a necessity.

As covered elsewhere in this thread, Convert to Edit-Friendly uses CRF 15 and GOP 1. This is a great option if editing directly on the converted file because it will seek quickly. However, @DRM does bring up a great lossless option.

If editing is done on proxies, and if you’re using a custom export preset to batch-convert the source files to edit-friendly, then here’s a cool option… by using H.264 Lossless (quality at 100%) with GOP 25 (or higher) and B-frames 0, and using ALAC for lossless audio, we get a lossless converted file that is potentially much smaller in size than the GOP 1 version. Higher quality and lower size… what’s not to like? Well, it will seek poorly because it is Long GOP instead of GOP 1. However, when using proxies, seeking is done on the proxies rather than the converted file, so the seek-speed penalty doesn’t matter. I often use this kind of workflow, and only turn proxies off when I want to spot-check crucial frames for color grading, and need the original files for maximum color accuracy. Then it’s back to proxy mode for actual editing.

Let’s take this back to the very beginning…

The source footage is coming from a phone. Current phones as a general rule have terrible color fidelity and tonality compared to larger-sensored cameras with higher-quality codecs, meaning well beneath the quality that H.264 was designed to support. (Play a phone video back on an 88-inch screen if you don’t believe me. Anything can look good on a four-inch phone screen, so that’s not a decent test.) The built-in “Good”-level settings of the Convert to Edit-Friendly tool will be well more than sufficient to transcode phone footage without any noticeable loss.

This is a great workflow, but it’s effectiveness kinda depends on your computer hardware. As you mentioned, people recommend against editing with Long GOP H.264 directly because unpacking Long GOP is processor intensive, which makes the preview laggy. GOP 1 of course doesn’t have that problem. However, even though ProRes is GOP 1, the decoder in FFmpeg is not as optimized as some other decoders. So your computer may or may not decode ProRes fast enough (especially at 4K60) to have smooth playback. It depends on your hardware. If ProRes is choppy and laggy, this is why… assuming slow magnetic HDD that can’t keep up with the ProRes bitrate isn’t an issue too.

1 Like