---
title: '4K/60 FPS Explained - Frame Interpolation & Upscaling Compared (Aiarty Vs. Comfy UI)'
source: 'https://youtube.com/watch?v=rG_65eC1vEw'
video_id: 'rG_65eC1vEw'
date: 2026-06-14
duration_sec: 1317
---

# 4K/60 FPS Explained - Frame Interpolation & Upscaling Compared (Aiarty Vs. Comfy UI)

> Source: [4K/60 FPS Explained - Frame Interpolation & Upscaling Compared (Aiarty Vs. Comfy UI)](https://youtube.com/watch?v=rG_65eC1vEw)

## Summary

This video demonstrates how to upscale a video to 4K and interpolate it to 60 frames per second using both a manual Comfy UI workflow and the paid AI tool Aiarty Video Enhancer. The creator explains the processes of upscaling (increasing resolution) and frame interpolation (adding frames for smoothness), then compares the results and resource usage of the two methods.

### Key Points

- **Definitions of Upscaling and Interpolation** [00:42] — Upscaling increases resolution without quality loss, potentially adding detail via AI. Interpolation adds frames to increase FPS for smoother video.
- **Extracting Audio and Frames with FFmpeg** [02:04] — First, extract audio from the original video. Then use FFmpeg to extract all frames as PNG images. Commands are provided in the description.
- **Tools Mentioned: Lossless Scaling and Aiarty** [03:24] — Lossless Scaling is a Steam app for real-time frame interpolation in games/apps. Aiarty Video Enhancer is a paid AI upscaling tool (not sponsored).
- **Comfy UI Upscaling Workflow** [04:01] — Images are upscaled to 4K in Comfy UI. Using an AI model increases processing time significantly (from ~10 minutes to 1-2 hours). The creator's GTX 1070 with 8GB VRAM handles it, but 16GB RAM is insufficient.
- **Recombining Frames into Video** [07:40] — Use FFmpeg to combine frames back into a video, specifying frame rate and adding an empty audio track to avoid issues when merging with audio later.
- **Frame Interpolation in Comfy UI** [10:44] — A node multiplies frames (e.g., 30 to 60 FPS). The process requires significant RAM; the creator had to split the video into 50-frame segments due to crashes with 16GB RAM.
- **Combining Segments and Adding Audio** [14:46] — Use FFmpeg with a text file to concatenate video segments. Then add the original audio track. The final video is 4K 60 FPS.
- **Aiarty Video Enhancer Demo** [17:28] — Aiarty offers a simpler UI with options to upscale (using models), interpolate frames, adjust color, and denoise. It processes a 10-second video in about 10 minutes with better resource management.

### Conclusion

Both Comfy UI and Aiarty can produce 4K 60 FPS video, but Comfy UI is free and highly customizable yet resource-intensive, while Aiarty is paid, easier to use, and more efficient. The choice depends on your hardware and willingness to tinker.

## Transcript

Hello once again audience. Welcome back
to another video. And today we're going
to do something different. I'm going to
show you guys how to upscale a video and
interpolate a video to 60 frames per
second. But first, please subscribe to
the channel if you haven't already. I
would greatly appreciate it. This video
costs a lot of time and effort to make
and I would really appreciate if you
could just subscribe to the channel to
get me to a th00and subscribers. I mean,
we got a bunch of views already and a
bunch of watch time. We just need the
subscriber count. So, I'm going to
explain real quick what upscaling means
and what video interpolation also means.
Upscaling is the process of increasing
the resolution of an image or a video
without reducing the quality or creating
ugly stretches.
And
uh in this case you can also add more
details by adding an AI model to improve
the quality or add more details or more
sharpness whatever you want.
And video interpolation is the process
of adding more frames per second to a
video. So basically generating in
between frames uh for your video so you
can make them buttery smooth.
And uh basically what I tried to do
there was uh run a compui workflow to
upscale a video
uh which is getting split into separate
images and that's the the video right
there which I showed in the intro and
those images are getting upskilled
separately. However, due to my PC being
kind of poopy and not that great uh it
crashed. So, what I'm going to do is I'm
going to manually
um split the images. I'm pretty sure
there is a compi workflow node somewhere
which allows you to do parts of the
batch.
Uh so what I did first I was splitting
the audio from the video because if you
put images uh together you don't have
the audio. So, you need to keep the
audio from the original video uh
somewhere safe, which is what I did
first. And next up, I uh I use a little
command using ffmpeg uh which is the
program I use for video and audio
uh processing, editing,
uh just trying to combine videos and
whatever if I get stuck uh grabbing
images from videos, etc., etc. That's
what FFmpeg It's open source library and
what I'm doing now is I'm running a
command to uh extract all the frames of
the video and create PGs from them. And
uh you can see it takes a little bit of
time and it should be around 300 images
because it's 300 frames obviously.
Uh, don't worry. I'll put all the
commands in the description because uh,
you're not going to remember everything
I type. And I guess typing it well,
pausing the video is also kind of
annoying. So, don't worry. I put
everything into the description. Um, so
now we have the little uh, images here.
I'm just going to put them in a separate
folder for safekeeping
so I can easily use that folder in a
comfy UI.
And besides Scumview, I also have
Lossless Scaling for frame
interpolation, which is like a Steam
program which you can use for screen uh
applications or games to create 60 fps
without actually having to do any video
editing. Uh, and I'm also going to show
ART video enhancer, which is a paid AI
tool. I got a free trial. I'm going to
show you guys how it works. Uh, I'm not
sponsored, by the way. I have to say I'm
not sponsored. Uh, I hope they sponsor
me. So, if you're watching the ARD, yo,
sponsor your boy, please.
So, what I did here is, uh, basically I
had a little conf workflow and I grabbed
a batch of images and I upskilled them
to, uh, 4K resolution uh, without the
use of the model. I do have a model in
this little preview, but uh actually
using that model kind of tenfolds the
time. This already took over an hour to
do in total. So, I'm just using like
little batches here to to upskill them.
And uh you can see that it takes about a
minute for like 50 I want to say.
uh maybe a bit longer but the process of
like generating the images and doing all
that stuff setting up confiu is such a
drag uh I'm using Pinocchio
and I think it has like 18 gigs of
storage usage just for nodes and other
libraries and whatnot. It's huge and uh
it really eats up my PC.
Uh, I I had a previous version of this
video and there was just static noise
and just the viewing experience wasn't
that great. Honestly, I'm sorry. It's
just cuz um because of my PC was kind of
dying. So, that's why I'm re-recording.
If you saw the other video, that's why
I'm re-recording
for uh for better audio because I I
checked it out. I'm like, dude, this is
awful. I cannot just put this out. And
also I added some some nice new tweaks
during the results section. So I I'm
really happy how that turned out and
might be a bit easier to spot the
differences now.
So uh yeah, you can see I think this
takes like 2 minutes for 70 uh and then
it's sourced to disk. So let's say it
takes about two minutes for 60 and it's
uh 300 frames.
That's about uh uh 10 minutes for
upscaling.
Uh which isn't too bad uh considering
ARD also takes about 10 minutes.
Uh but there is a quality difference
because ARD is using a
AI model to increase the details and I'm
not doing that here because that would
probably make the runtime close to one
hour maybe two hours
uh just for for the image upscaling and
then I'd still have to do 4K and
I combining it back into a video. So,
and for the sake of the video, I'm not
using a model, but you could use a model
if you have more RAM. Uh, I have a GTX
1070
and it's doing fine. It's got 8 gigs of
VRAM and it's it's holding up fine. The
problem is,
uh, mostly the RAM. I'm running out of
RAM. I have 16 gigs of RAM, which is
fine like like
uh like eight years ago when I got the
PC, but nowadays it's not. And and Comy
UI just eats everything up and
uh like I'm not sponsored by by ARD, but
ARD has better um system resource
management basically. So I'm just
showing off the um the image there. it.
It's upscaled to 4K now. It looks maybe
a little bit sharper. I don't know.
We're going to find out at the results
section later in the video. And what I'm
doing now is I'm uh combining
the
uh frames back into a video. And I'm
specifying using the minus r uh
parameter. I'm I'm specifying the frame
rate basically. And I'm also setting up
a empty audio clip. Uh because when you
want to merge like concatenate videos or
if you want to add audio to an existing
video, it's sometimes it kind of
uh like doesn't want to because you
don't have an audio track on the video.
So it's like just just a GIF basically.
It treats it like a GIF and a GIF can't
have sound. So if you have an empty
audio track uh by using the 128k thingy
um you ensure that you are not going to
run into issues later
uh adding audio or combining videos
which do have audio and videos with jon
of audio. So I had like um a banner
project before. I had like a banner and
I tried to merge it uh with like an
actual video but the banner didn't have
any sound. It was like just one frame of
banner. um and um that caused issues. So
that's why I know you you find out stuff
like that. Uh again, don't worry. I'm
just putting everything into the
comments.
And you can see this takes quite a long
time. Uh probably because each of the
images is like five megs big, five
megabytes, over 300 images. That's um do
the math, 1.5 gigs of images.
uh it's it's adding into a video and
honestly the video turns out to be like
55 megabytes or something. It's not
really that big. Uh and ARD basically
has the same thing. You know it it takes
a 2 gig raw video and it converts it
into like 55 or 20 megabytes video 4K.
So it does do some very good
optimization there. uh which I think
it's H 264 265 something.
So
uh now I'm going to be
uh let me check.
I
grab the video and I grab the audio.
Yeah. And I combine them back to one. So
now we are finished with our 4K video.
We we've upscaled it perfectly and we're
going to combine it with the other. This
is a very fast process cuz basically
you're just adding the track to to it.
And this is how it looks like now. And
at the end of the video, I'll be making
comparisons so you can compare for
yourself. And let me know in the
comments if you think it looks any
different. Uh and if so, which version
do you think looks better? I'm really
excited to hear you guys opinion. So
this again is the um
load video note which is the first one
on the left there. Uh you can load a
video in and then you can kind of
interpolate it using another node. Uh
basically I'm multiplying it to two
meaning it will get uh 60 frames instead
of 30. And then uh like the first node
is actually splitting the video into
images which uh basically I was doing
myself before as well. Uh and then the
the middle node is taking those images
and then adding in between frames. And
then the last node is actually combining
the images back into a video. So
basically
I could work with only getting the in
between frames and then generating the
file myself. But the last node is not
the issue. The first node is the issue
which again crashed my PC because I
don't have enough RAM. So I had to
create smaller bits of video. So I I cut
all the 300 frames into little parts.
and like 50 frames per video, which is
like 1.8 seconds, I guess, on a 30 frame
video, 30 frames per second video. And I
cut them in six parts cuz it crashed.
And uh this this worked. And it's
basically the same command as before.
We're just trying to to create another
video with the um 30 frames per second
parameter, the minus or 30. empty audio
track just cuz. And uh uh yeah, we
create six of them. And I have a very
bad naming schemes, you know, B1, A1,
C1. Uh I just want to like get this done
and I remove all the files afterwards.
So like when I'm doing stuff like this,
I usually have very default commands in
a in a notepad or something
uh with like a b c uh and they are
sequenced. So I know like oh I'm at C
step which is the third step. So I'll
try to do the same in the comments as
well for you guys. So you can easily
copy paste it in sequence for one
project and then uh you have like a a
nice finished end result.
So uh this like 50 frames not too long
takes uh almost 30 seconds. Yay.
And now I put it in into the um load
video again. Uh I do notice that you can
kind of limit the frames. Uh maybe you
can set a different starting frame. Uh
and you can kind of batch it like that.
That might work. But I'm also scared
that it's because it's a larger video,
it's trying to still load some stuff
into memory, which is going to cause it
to die again, and I don't want it to die
again because just sucks so much.
Um like I didn't get a blue screen. My
entire PC just completely froze. But you
can see that the first step was finished
way faster now, which again, less RAM
needed. So, it all fits into my little
16 gigs memory. Uh, if you have more
RAM, I I guess you could do this a lot
easier with bigger videos. Let's look at
the the the usage there. It is already
for 50 frames already almost capping my
my memory. It's taking 30 gigs of my
swap file on my hard disk. like you can
set up a certain section of your hard
disk for extra RAM, which is really
slow, but you can see that's also taking
like 30 gigs, 28 to 33. Um, and that's
that's besides my 15 to 16 gigs actual
RAM usage. So, it's almost 50 gigs of
RAM usage for a 1 second video for
interpolation. So, that's you need like
64 gigs at least to do this stuff.
Uh, so I finished all the uh little
parts there
and I'm combining them in order. You can
have a little text file with ffmpeg
and it will combine those videos in
order and it will just concatenate those
videos uh without any transitions. You
might be able to do a transition with
some fancy stuff maybe. I don't know.
You would have to look it up. But if you
don't have a video editor, but you want
to have like a transition between clips,
uh you could maybe do something like
that. And also
um you can easily combine clips using
ffmpeg. You don't need a fancy video
editor. You don't need Adobe Premiere
for just combining clips. Uh also Adobe
Premiere sometimes just crashes and
doesn't want to combine videos and and
just creates corrupted stuff. Uh, Fmpe
is really stable for video creation, but
if you want to do animations and like
fancy text and whatnot, you might need a
video editor.
So, I combined the 60 fps video now as
well. There we go. It's now 4K 60 fps.
There is no sound again because we do
have to add the sound to the video
again, which as we know from before is a
very fast process. So,
almost done with the 4K 60 fps uh AI
version.
I am also going to create a
nonupskilled 60fps version. See how that
runs and I'll show the comparison later.
That was me. Sorry, I pressed the
recording button. Uh that's not your
Windows doing a little plink sound.
That's just me. I'm sorry. I hate it
when YouTubers make a sound like Discord
notification or whatnot and they don't
mention it. This was me, okay? So, I'm
sorry. Uh, don't don't be scared. So,
again, we're we're combining the audio
with the um with the video. And like I
said before, I was going to try to do a
60 fps version with out upscaling. And
that one took maybe like 20 to 30
minutes. Uh, but it didn't crash my RAM.
It just took a very long time. Like I
was still able to watch some YouTube and
and open Discord. I could I could do
stuff while it was running. So,
it's really good. Uh, that's the wrong
version. Here we go. This is the uh the
sound version again.
60 fps with sound.
It looks It looks decent. I guess I'll
have to compare it with uh with the
original one later.
Uh, but that's basically it. It took us
about 15 minutes and 70 minutes even.
And now I'm going to show you ARD video
enhancer. I'll put the link in the
description. Again, I'm not sponsored.
Uh, but it's very easy. It's very
optimized and it checks your hardware
and everything. You can upscale using a
model. You have three models of choice.
I'm not sure if you can get other models
somehow. You can upscale to 4K. You can
use step mode for higher quality. You
can use turbo for faster processing. Uh
but it is uh uh it is lower quality. So
here you can see like I have a little
slider. You can see the edge of the uh
ridge basically gets a little sharper.
Let's look at Maido for a sec. Um yeah,
he does get sharper. Like clearly
sharper. Let's zoom in a little bit.
There you go.
Uh so you can see he does get a little
sharper. You do have to be careful with
the strength though, especially when you
have a very low res image. Uh sometimes
AI is adding some weird details like
with text. I've had instances where uh
the C turns into a D or a O. And uh as
you can see, you can also increase the
frame rate very easily. Uh you can
change the color. Uh you can use HDR uh
audio D noiseis like whatever you want.
and you just press the the uh export
button. You can even do a batch export
of multiple videos using the same
settings. So, you can see it takes about
10 minutes for a 10-second video. Uh
it's a lot less setup. It's very
much faster. So, I hope you enjoyed this
and I hope you enjoy the results in a
little bit. Uh because yeah, I think
they look pretty cool. So, please
consider subscribing to the channel if
you have not already. I would greatly
appreciate it and I hope to see you all
next time. Let me know in the comments
which versions you like the most and if
you notice any big differences or not.
So, thank you for watching. I hope you
found this useful or interesting at
least. And see you tomorrow with Maple
Story again. I'm sorry. Okay, bye-bye
everyone.
[music]
>> [music]
>> Don't
>> [music]
[music]
>> And now a shout out to our members. On
the extreme tier we have Dustin H. On
the normal tier we have Daniel Chavez.
Thank you all so much.
See you all next